Criticaldiscourseanalysisa [612075]

Sociological Theorizing and
Quantification P ractices in
Survey -Based R esearch

Habilitation Thesis

Ana-Cosima Rughiniș

University of Bucharest

2012

2
Table of contents

Abstract ………………………….. ………………………….. ………………………….. ………………………….. …… 4
Rezumat ………………………….. ………………………….. ………………………….. ………………………….. ….. 6
1 Introduction ………………………….. ………………………….. ………………………….. ………………….. 8
1.1 Professional journey ………………………….. ………………………….. ………………………….. ….. 8
1.2 Current research orientation ………………………….. ………………………….. …………………. 11
2 Sociological theorizing in survey -based research ………………………….. ………………………. 13
2.1 A theoretical sketch: the “trait” versus “interaction -order” divide ……………………… 13
2.2 Checkpoints on the “trait” vs. “interaction order” divide ………………………….. ………. 18
2.2.1 The error – trait device ………………………….. ………………………….. …………………. 20
2.2.2 Quasi -causes versus method s and resources ………………………….. ……………….. 21
2.2.3 Imposed relevance ………………………….. ………………………….. ……………………….. 22
2.2.4 Resonating typologies ………………………….. ………………………….. …………………… 22
3 Theorizing ethnicity i n surveys of ethnic minorities. The case of research concerning the
Romanian Roma / Țigani ………………………….. ………………………….. ………………………….. ………. 24
3.1 Theoretical views on ethnicity ………………………….. ………………………….. ……………….. 24
3.2 Theoretical bearing of survey -based methodological discussions ……………………….. 26
3.2.1 The error – trait device in survey -based research on ethnicity ……………………. 26
3.2.2 Ethnicity as quasi -cause of action versus resource for action ……………………… 28
3.2.3 Relevance of ethnicity: a blind spot of the survey interview ……………………….. 30
3.2.4 Ethnic typologies of the first and the second order ………………………….. ………. 31
3.2.5 A research heuristic: ethnicity as a classification based on astrological signs .. 31
3.3 Previous research contributions concerning survey -based sociological researc h on
ethnicity ………………………….. ………………………….. ………………………….. …………………………. 32
3.3.1 Measurement of Roma ethnicity: hetero – and self -identification ………………… 32
3.3.2 Model specification and data analysis ………………………….. …………………………. 37
4 Theorizing public knowledge of science in surveys of scientific literacy ……………………. 40
4.1 Theoretical views on scientific literacy ………………………….. ………………………….. ……. 40
4.2 Theoretical bearing of survey -based methodological discussions ……………………….. 42
4.2.1 Scientific literacy as a trait ………………………….. ………………………….. …………….. 42
4.2.2 Relevance and typologies ………………………….. ………………………….. ………………. 44
4.3 Previous research contributions concerning survey -based sociological research on
scientific literacy ………………………….. ………………………….. ………………………….. ……………… 45
4.3.1 The error -trait device: scientific literacy as vocabulary, worldview, or trait ….. 45
4.3.2 Understanding trouble in the quantification of scientific literacy: three possible
distinctions ………………………….. ………………………….. ………………………….. ………………….. 47
5 Overview of research contributions ………………………….. ………………………….. ……………. 55
6 Plans for future development ………………………….. ………………………….. …………………….. 59
6.1 Igel – Sociological imagination and disciplinary orientation in applied social research.
An inquiry into present -day market research in Romania ………………………….. ……………… 60
6.1.1 Scientific context: studies of science in applied and corporate settings. ………. 60
6.1.2 Research focus ………………………….. ………………………….. ………………………….. … 61
6.1.3 Methodology ………………………….. ………………………….. ………………………….. …… 62
6.1.4 Impact, relevance, applications ………………………….. ………………………….. ……… 63
6.2 LiSa – Gaps and bridges. Pursuing individual life satisfaction and happiness in the
public sphere ………………………….. ………………………….. ………………………….. …………………… 63

3
6.2.1 Scientific context ………………………….. ………………………….. ………………………….. 63
6.2.2 Research focus ………………………….. ………………………….. ………………………….. … 63
6.2.3 Methodology ………………………….. ………………………….. ………………………….. …… 65
6.2.4 Impact, relevance, applications ………………………….. ………………………….. ……… 65
7 References ………………………….. ………………………….. ………………………….. ………………….. 66

4
Abstract
This thesis presents an overview of my scientific contributions and current research directions ,
against the background of relevant discipl inary debates . I focus on those areas of my previous
research activity that are of most rele vance for my current interests, in the methodological tradition
of survey -based sociological research, which was the focus of my book on “Sociological Explanation ”
(Rughiniș 2007a) . I present , on one hand, my research on ethnicity, in particular Roma / Gy psy
ethnicity (Rughiniș 2011a; Duminică, Lupu, and Rughiniș 2009; Rug hiniș 2010; Gabor Fleck and
Rughiniș 2008) , and , on the second hand, my research in the Public Understanding of Science
program , concerning the stock of public knowledge of science, or ‘scientific literacy’ (Rughiniș and
Toader 2010; Rughiniș 2011b; Vl ăsceanu, Dușa, and Rughiniș 2010) .
In both fields, my overarching research interest has been to evaluate current survey quantification
practices as regards their theoretical grounding and affinities . In this t hesis I start from my previous
research results and I propose solutions for quanti fication methods that support an interactionally –
oriented theoretical perspective. I therefore present the significance of my research in light of a
broader and more far -reac hing discussion, namely the evaluation of survey -based research with
reference to its basic theoretical assumptions concerning social action and interpersonal interaction
independently of their thematic specificities .
One of the groundbreaking works in thi s current of reflection has been Aaron Cicourel’s “Method
and measurement in sociology” (1964) . Much of the ethnomethodologically -informed work
concerning the theoretical bearing of survey -based inquiries takes standardized interviews as a
research topic, studying it as a particular form of social interaction without aiming to formulate
proposals of practical imp ort. Critics that denounce “the folly of the whole enterprise”, as
Hammersley (2010) puts it in his recent tak e on this topic, do not engage the technicalities of survey –
based research because they do not see the possibility to improve them or to render them (more )
theoretically meaningful. My research is significant insofar it contributes to the fine thread of
discussion attempting to bridge the interactional theoretical orientation with the practical ities of
survey -based research , by selecting several points of contention and proposing solutions.
I have oriented my analyses to what I considered to be the dominant debates in survey -based
research in the fields of Roma / Gypsy ethnicity and scientific literacy. After analyzing the problems
formulated in the se debates and the proposed solutions, inside and outside survey practice, I have
directed my attention towards three broad topics of controversy : 1) the ‘operationalization’ issue, or
how to link concepts to measures, 2) the issue of causality in sociological explanation , and 3) moral
quandaries – which derive from survey analyses ’ reliance on common -reason catego ries and
concerns .
Despite the centrality of measurement in survey methodological discussions, in actual research the
theoretical justification of measurement practices is often very limited . The methodological and
material constraints of the standardized interview instrument selectively support some directions of
theoretical grounding at the expense of others. At the same time, in recent years the practice of
advanced statistical analysis has developed considerably and has become increasingly accessible –
bringing its own theorizing affinities . Its development has benefitted from increasing access to
personal computers, and also from its disciplinary -independence: economists, political scientists,
psychologists, sociologists , epidemiologists, and even speci alists in natural sciences can be trained

5
together in advanced statistical techniques, and can contribute to further refinements of statistical
procedures. The repertoire of statistical analysis affords analysts a rich array of methods that largely
support a de -contextualized view of data , encourage efforts to translate data into stable traits of
individual respondents or aggregates of respondents , and boost a causally -oriented vocabulary in
data analysis.
As regards ethnicity, I find that considerable met hodological energy is put into refining ethnic
categories in questionnaires and eliciting ‘honest’ answers from respondents. The issue is largely
defined as one of truthful classification , one that would match respondents’ ‘identities’ . A secondary
researc h move involves refining the measurement of ethnicity conceived as ‘identity ’, departing
from simple , clear -cut classifications while endorsing the view of ethnicity as a shared individual
internal trait. These concern s marginalize theoretical preoccupatio ns with the boundary -making and
other interactional stakes involved in the use and refinement of ethnic vocabularies, and with their
contextual and pragmatic , rhetorical relevance. As a consequence, the common reason practices of
maintaining the stability, omni -relevance, and causal influence of ethnicity are non -problematically
imported into survey -based sociological research.
As regards scientific literacy, I find that the survey use of the National Science Foundation scale has
supported a narrowing of t he concept into a stable trait located in the individual mind , largely
independent of the contexts of learning and use of scientific constructs . I argue that the interactional
relevance of public knowledge of science can be observed and analyzed when looki ng at the deba te
over ‘knowledge of evolution ’, and I propose a distinction between ‘animated ’ and ‘quiet ’ scientific
constructs to address this problem .
My analysis of survey research on ethnicity and scientific literacy illustrates four areas of disconte nt
which I have selected as critical and also amenable to interactionally -aware solutions: the error -trait
device, the use of a quasi -causal vocabulary, imputation of relevance, and resonating typologies. I
propose several corresponding analytical orientat ions to avoid these points of contention: a)
analyzing and interpreting data as collaborative results of situated interaction, b) re-specifying the
concept of ‘error’ in line with participants’ understanding of what counts for a mistake , c) evaluating
anal ysts’ use of categories by reflecting on their variable contextual relevance in social interaction ,
d) increased theoretical attention to typ ification and second -order typology construction , e) an
avoidance of the quasi -causal statistical jargon, and less focus on statist ical significance in favor of a
preoccupation with substantive troubles of quantification and issues of absolute and relative size,
and f) attention to the reliance on common reason categories and concerns, to our commentaries on
them, as a nalysts, and to how our findings are likely to be taken over in common reason social
knowledge claims .

6
Rezumat
În această teză voi discuta contribuțiile mele științifice și direcțiile mele actuale de cercetare în
contextul dezbaterilor relevante din literatura de specialitate . Mă voi concentra asupra celor mai
semnificative arii ale activității mele de cercetare din pe rspectiva intereselor mele curente; este
vorba mai ales de spre publicațiile din tradiția metodologică a cercetărilor sociologice prin anchete
cantitative (sondaje) , către care m -am orientat și în lucrarea „Explicația sociologică” (Rughiniș
2007a) . În cele ce urmează voi prezenta , în primul rând, cercetările mele privind etnicitate a, cu
precădere etnicitatea romă / ț igănească (Rughiniș 2011a; Duminică, Lupu, and Rughiniș 2009;
Rughiniș 2010; Gabor Fleck and Rughiniș 2008) . Apoi voi trece în revistă contribuțiile pe care le -am
realizat în domeniul cunoașterii publice a științei, în tradiția Public Understanding of Science , cu
referire la conceptul de „alfabetizare științifică ” (Rughiniș and Toader 2010; Rughiniș 2011b;
Vlăsceanu, Dușa, and Rughiniș 2010) .
În ambele domenii interesul meu dominant a constat în evaluarea pr acticilor curente de cuantificare
în anchetele sociologice, evaluând întemeierea lor conceptuală precum și afinitățile lor pentru
anumite teorii ale acțiunii sociale . Pornind de la rezultatele cercetărilor mele trecute, în această teză
propun câteva soluți i privind strategii de cuantificare care sprijină o perspectivă teoretică orientată
interacțional. Prin urmare, î n această teză discut semnificația cercetărilor mele pe fundalul unei
preocupări mai ample și cu mize teoretice mai mari – și anume, studiul presupozițiilor teoretice de
bază ale cercetărilor sociologice pe bază de sondaj privind acțiunea socială ș i interacțiunea
interpersonală .
Una dintre lucrările fundamentale în această zonă de reflecție o constituie cartea lui Aaron Cicourel
“Method and M easurement in Sociology” (1964) . O mare parte a cercetărilor inspirate, cel puțin în
parte , de perspectiva etnom etodologică privind relevanța teoretică a investigațiilor prin anchete
cantitative se apleacă asupra interviurilor standardizate ca proces social specific, studiindu -le ca
formă de interacțiune – fără a urmări însă să formuleze recomandări sau soluții. Cri ticii care denunță
„nebunia întregului demers” („the folly of the whole enterprise”, în formularea sugestivă a lui
Hammersley, 2010) nu se implică și în discutarea și evaluarea tehnicalităților sondajelor , deoarece
nu văd posibilitatea ca acestea să fie ut ilizate cu vreo miză teoretic ă. Cercetările mele sunt astfel
importante în măsura în care contribuie la o zonă destul de restrânsă a dezbaterilor, selectând
anumite aspecte problematice cheie din practica metodologică a sondajelor în ariile tematice
discut ate, pentru care ofer ilustrații și propun soluții.
În munca de până acum mi-am orientat atenția către dezbaterile pe care le -am considerat cele mai
importante și mai intense în studiile cantitative sociologice privind romii / țiganii și cele privind
cunoa șterea publică a științei. Pe măsură ce am reflectat la problemel e formulate în aceste dezbateri
m-am îndreptat căt re trei mari teme controversate : 1) chestiunea operaționalizării, sau cum pot fi
legate conceptele de măsurători; 2) chestiunea cauzalităț ii în explicațiile sociologice și 3) dificultățile
morale – care derivă deseori din întemeierea ne-reflexivă a sondajelor pe categorii și preocupări ale
cunoașterii noastre comune, practi ce.
În ciuda centralității problematicii măsurării în discuțiile metodol ogice din cercetarea prin sondaj ,
justificarea teoretică a modelelor de măsurare în practica de cercetare empirică este adeseori foarte
limitată. Constrângerile metodologice și materiale ale chestionarului ca instrument încurajează
anumite direcții de înte meiere teoretică în defavoarea altora. În paralel, practica analizelor statistice

7
avansa te a devenit tot mai accesibilă , inducând propriile sensibilități teoretice . Dezvoltarea sa se
bazează atât pe avansul tehnologiei, mai ales al cal culatoarelor personal e, dar și p e independența sa
disciplinară: economiștii, specialiștii din științe politice, psihologii, sociologii, epidemiologii și chiar
specialiștii în științe ale naturii se pot întâlni în cursurile de formare în analiză statistică, contribuind
ulterior la rafinarea acestora. Repertoriul analizelor statistice permite analiștilor un evantai larg de
metode care, în linii mari, încurajează o abordare decontextualizată a evidențelor empirice , susțin
interpretarea datel or ca indicând trăsături stabile ale res pondenților individuali sau ale agregărilor
acestora , și dezvoltă un vocabular analitic cu concepte cauzale .
În ceea ce privește etnicitatea, în cercetările mele anterioare am observat că o mare parte a
discuțiilor metodologice se concentrează pe rafinare a categoriilor etnic e utilizate în chestionare și pe
găsirea tehnici lor de încurajare a subiecților să răspundă „onest”. Problematica asumată este cea a
clasificării veridice, a adecvării etichetelor etnice la identitatea reală a respondenților. O dezvolta re
secundară ca amploare constă în elaborarea unor scale și proceduri de măsurare a etnicității ca
„identitate” – depărtându -se deci de utilizarea clasificărilor simpliste, dar susținând, în același timp,
concepția etnicității ca atribut individual , interi or persoanelor, chiar dacă împărtășit socialmente.
Aceste interese de cercetare pun în umbră preocupările teoretice privind procesele de trasare și
menținere a granițelor etnice, precum și sensibilitatea pentru mizele interacționale ale utilizării
vocabula riilor etnice ș i pentru relevanța lor pragmatică, retorică. Prin urmare, practicile de simț
comun prin care susținem în viața de zi cu zi stabilitatea, omnirelevanța și puterea cauzală a
etnicității sunt importate ne -problematic în cercetarea sociologică p e bază de sondaj .
În ceea ce privește alfabetizarea științifică, am observat că utilizarea scalei National Science
Foundation a încurajat o re-specificare a conceptului ca trăsătură stabilă a respondenților, localizată
în mințile lor individuale , și relati v independentă de contextele de învățare și invocare a
constructelor științifice. Am argumentat că relevanța interacțională a cunoașterii publice a științei
poate fi observată atunci când analizăm dezbaterea privind „cunoașterea evoluției” și am propus o
distincție între constructe „animate ” și „inerte ” pentru a ține cont de această problemă.
În analiza sondajelor sociologice privind etnicitatea și alfabetizarea științifică am ilustrat patru
aspecte problematice pe care le -am evaluat ca fiind critice, dar posibil de soluționat într -un sp irit
teoretic interacțional: dispozitivul „eroare – trăsătură individuală”, utilizarea unui vocabular cvasi –
cauzal, imputarea relevanței și apelul la tipologii care rezonează cu cunoașterea comună. În această
teză propu n câte va re -orientări analitice prin care aceste probleme pot fi evitate: a) analizarea și
interpretarea datelor de sondaj ca fiind rezultate colaborative ale interacțiunilor situate; b)
respecificarea conceptului de „eroare” în consonanță cu ceea ce participanț ii la acest demers
identifică, în general, ca fiind „greșeli”; c) problematizarea utilizării analitice a categoriilor prin
reflecția asupra relevanței lor variabile și contextuale în interacțiunea socială; d) o atenție teoretică
sporită la procedurile de tipizare și de construire a tipologiilor de nivelul doi; e) evitarea jargonului
statistic cvasi -cauzal și diminuarea importanței argumentative acordate semnificației statistice
(probabilității de eroare p) în favoarea unei preocupări cu problematica de subs tanță a cuantificării și
a interesului pentru dimensiunile absolute și relative ale procesel or măsurate; nu în ultimul rând f) o
atenție crescută asupra preluării categoriilor și preocupărilor din cunoașterea socială practică de
simț comun și asupra felulu i în care noi le comentăm în ipostaza de analiști și a felului în care
rezultatele investigațiilor noastre sunt preluate și re -integrate în practicile cotidiene de cunoaștere.

8
1 Introduction
I will briefly describe in the following section the development o f my research involvement and
interests after completing my doctoral studies . I aim to sketch my professional journey up to date , in
order to clarify the research questions th at have oriented my recent work , their transformation in
current projects , and th eir public relevance . I am deeply indebted to my colleagues, students, family,
and friends for supporting and giving meaning to my work . In writing this thesis I have received
constant encouragement and counsel from P rofessor Lazăr Vlăsceanu, and I have sought and found
much appreciated advice and theoretical guidance from Puiu Lățea. I am grateful to Professor
Dumitru Sandu, Bogdana Humă , and Ștefania Matei for making inspiring commentaries, critiques ,
and correction s of my arguments.
Overall, my professional trajectory covered in this thesis has included a period of interest in social
exclusion and ethnicity, during my doctoral studies, then a post -doctoral period of increasing
involvement and training in advanced quantitative research, and, most recently, a period of relative
disenchantment with survey -based sociological knowledge and the modes of attention that it
encourages for researchers, at least in my own experience. A combination of interest in
quantification practices, appreciation for the practical value and argumentation power of numbers,
and interest in social theory encouraged me to search for way s to put these powerful and highly –
used instruments in the service of sociological theories of the interaction order (Goffman 1983;
Sacks 1995; Warfield Rawls 2011; Warfield Rawls 1987) . This thesis rep orts on my research
contributions up to date, and discusses their significance for an attempt to re -orient the theoretical
bearing of survey -based sociological research.
1.1 Professional journey
My research career in sociology has started with a first change o f mind: I had promised myself that I
would not do research concerning Roma / Gypsy people, because the field seemed already
overpopulated. My plan notwithstanding, I gradually became involved in several research projects
addressing Roma issues and, later, my growing experience invited even more involvement. If, in the
early beginning, I felt little, if any interest in issues related to Roma people or topics such as poverty
or ethnicity, during my six years of doctoral research I often felt curious, perplexe d, sympathetic,
sorry, embarrassed, illuminated, and in many other ways, while making sense of unexpected
interactions.
In brief, my involvement in research projects on Roma and Gypsy issues has been a period of
escalating puzzles, which I tried to sort o ut and finally to shape in a canonical sequence of theory,
questions, methods, and answers in my doctoral thesis. This effort of structuring questions and
answers left some residual uncertainty as to the overall significance of my research approach, and, i n
particular, my quantitative research on topics related to Roma ethnicity .
After reading Michael Stewart’s inspiring draft paper on “Approaches to Roma and Gypsies from
within social anthropology with particular reference to the Anglo -Saxon traditions” (Stewart 2008) , I
started to work on a review of the quantitative research on Romanian Roma , with which I was
familiar. One question that became increasingly important for me was: what can we learn from the
aggregated body of datasets and reports on Roma? Is there any accumulation of knowledge, be it by
the sheer amassment of data? I could vouch that the biographical research experience was worth

9
the effort – but i s there any inter esting pattern that emerges , for a distant reader, from revisiting
these results?
A secondary concern was whether quantitative research on Romanian Roma is a meaningful
contribution to the wider sociological and anthropological research tapestry. Granted t he
methodological difference, is there a theoretical convergence, or some dialogue that pushes
reflection further?
My article (Rughiniș 2010) is the result of my attempt s to put order and make sense of the rich
empirical data collected in the surveys to which I have participated and the ones I had access to. I
concluded that much of the theoretical -cum -methodological discussion on operationalizing Roma
ethnicity revolves on the issue of self – versus hetero -identification, and I argued for the use of self –
identification, on what I see now to be a combination of moral , practical, and theoretical
considerations. I have discussed the use (or lack thereof) of community level variables, which are
deemed important for analysis and for sampling as well. I observed that the distribution of Roma
across heterogeneous and mixed communities, as well as what I termed the “self-identification
reluctance error ” were blind spots of the survey instrument; therefore , we have to rely on other
information or insights to sort them out. Lastly, while examining an overall chart for the distribution
of education in the surveyed Romanian Roma population, I concluded that, although the results
were by and large consonant, there is a non -intelligible variation that does n ot seem to amount to a
tendency and does to seem to reflect variations in the sample type (which was my initial hypothesis) .
Therefore, it seemed to simply indicate imprecision or, maybe, indecision of another kind – such as a
fundamental ambiguity and oscillation of what it means to be a Romanian Roma . Moreover,
important survey questions were phrased differently, thus discouraging comparison. To put it here
more bluntly than I wrote in the published article, I did not find persuasive evidence for the
possibility to cumulate empirical survey dat a on Romanian Roma , at least up to date .
I hav e next turned to the question of concepts (Rughiniș 2011a) : What is it that we observe by
quantifying Roma ethnicity in survey research? Moreover, what is it that we mean by ethnicity? Is
there a dialogue between current understandings of ethnicity in other strands of social research and
the uses of ethnicity in surveys ? By and large, I concluded that conceptual models of ethnicity in
quantitative research are simplistic even for quantitative standards – for example, in comparison
with measurements of religiosity. Ethnicity is measured and often a lso interpreted according to its
common reason functioning – as a categorical, distinctive, individual deep trait that is causally
relevant for understanding behavior. While part of the trouble lies in its operationalization as a
categorical variable, addi tional perils derive from statistical analysis models. For example , controlling
for additional variables such as ‘education ’ or ‘income ’ when studying differences associated with
ethnicity often makes no theoretical sense at all, a point also underlined be fore by Steinberg and
Fletcher (1998) and G. D. Smith (2000) , but largely ignored in empirical research. Regression
coefficients , an often sought -for result interpreted as the ‘influence of e thnicity ’ when othe r
variables are ‘controlled for’ , indicate in many research designs the amount of residual mystery ,
rather than the amount of causal influence. Moderating effect s, which are often not estimated, may
be highly theoretically relevant .
In these two recent papers I have contributed to a steady debate on measuring and using ethnicity in
survey -based research (see for example Aspinall 1997; Chandra and Wilkinson 2008; Connolly 2011;
Kwan and Liem 2000; Mateos, Singleton, and Longley 2009; Phinney and Ong 2007; Simon 2011;

10
Singh 1997; Burton, Nandi, and Platt 2010; Mays and et. al. 2003; T. W. Smith 2008; American
Sociological Association 2003; Aspinall 2001; Bhopal 2006; Brown and Langer 2010; Stephan and
Stephan 2000; Winker 2006) , and to a dist inctive thread on the practices of surveying Roma / Gypsy
populations (Durnescu, Lazar, and Shaw 2002; Ahmed, Feliciano, and Emigh 2006; Babusik 2004;
Emigh and Szelényi 2001; Gábor Fleck and Rughinis 2008; Gabriel Bădescu and et.al. 2007; Hajioff
and MCKee 2000; S andu 2005; Csepeli and Simon 2004; Ladányi and Szelényi 2001; Szelényi 2001;
Ladányi and Szelényi 2002; Prieto -Flores 2009; Zamfir Catalin and Preda Marian 2002; E. Zamfir and
Zamfir 1993) . My discussion fram e has oscillated between a methodological orien tation ( searching
for practical solutions to problems of data management) and a theoretical one (aiming to
understand ethnicity as a social process ). In this thesis I elaborate in more detail the theoretical
relevance of some methodological debates within survey research on Roma / Țigani, which I
investigate as a particular case of survey research on ethnicity.
Besides the sociological questions pertaining to methodology and theory, during some of the many
discussions with Roma and non -Roma people active i n the Roma rights movement and during
collegial debates I confronted some disquieting questions and comments, which I interpreted as
moral , and may also be thought of as paradigmatic. We may ask ourselves : What are we actually
doing, socially, when we are doing quantitative social research and telling reports and other stories
about it – say, on issues related to the Romanian Roma people? How do we participate , in various
daily interactions, in shaping ethnicity as a resource for social classification and a ttribution
processes ?
The first signs of trouble appeared for me when I met the discontent of people in the Roma
movement that most Roma research, including surveys, portrays Roma people as poor, uneducated,
and generally at a loss in society. Roma people’ s lives are explained away by lack of education and
poverty, occasionally including discrimination. Nowadays I understand this problem by the
observation than q uantitative survey research risks to accomplish , in effect , an Othering of the Roma
people, depi cting them as interactionally incompetent for many practical purposes (Warfield Rawls
and David 2006) . When writing the “ Come Closer! ” book (Gabor Fleck and Rughiniș 2008) , we tried,
as a research team, to mitigate th is alterization damage by sever al means: methodologically, by
appe aling to community case studies; stylistically, by including in the report photos that portray un –
diminished persons; contentwise, by including a discussion, in the Conclusions section, on
“Constructing the “Average” and the “Middle” Roma Subject” (pp. 217 -218). The question remained,
for me, how (and if) it is possible to put to use quantitative methods of research and reporting while
acknowledging the radical heterogeneity of experiences of all the people aggregated in a n average.
In retrospect, a more appropriate formulation would be: whether, and how, quantitative social
research can produce portraits of people as socially ( interactionally ) competent persons.
In my postdoctoral research years I have attended several tra ining courses in advanced statistics,
focusing on latent variable m odels (structural equations, latent class, GLAMM) and other advanced
techniques (multilevel analysis, data imputation for missing values, and, generally, methods for
survey design) . During the same period my research interests gradually shifted towards the sociology
of scientific knowledge, and the STISOC1 project offered me an opportunity to get involved in hands –

1 “Science and Society. Interests and perceptions of the public concerning scientific research and its results” (Știință și
societate. Interese și per cepții ale publicului privind cercetarea științifică și rezultatele cercetării), financed by NASR –
National Authority for Scientific Research (ANCS – Autoritatea Națională pentru Cercetare Științifică), principal investigator
L. Vlăsceanu, Department of S ociology and Social Work, University of Bucharest [grant number 203CPII/10.09.2008].

11
on analysis of quantitative evidence on scientific literacy. I took over to th is new field the persistent
questions: what concepts of public knowledge of science are used in survey -based research? Is there
a dialogue between the quantitative strands of research and the ethnographic research on those
topics , or other qualitative appr oaches?
As I have discussed in Rughiniș (2011b) , processes of translation between theory and standardized
measures are bi-directional, because quantification and subsequent statistical analysis accomplish at
least some implicit theorizing . For example, the very same collection of items may bring about a
different concept in a reflective versus a formative measurement model . My research approach in
this field has been 1) to propose conceptualizations of scientific literacy that are better compatible
with the measures used in survey -based research , 2) to point out the incongruencies between
theoretical concepts and specific practices of measurement , and 3) to propose concept
operationalizations and research strategies that would mit igate the troubles, while broadly
addressing the same research questions.
1.2 Current research orientation
The question which I now address, in my research, is how to reformulate the research questions in
quantitative survey -based investigations , together wit h the methodological and discursive apparatus
employed in data production and in reporting, in order to support an analytical sensitivity towards
the competence of social actors and the interactional accomplishment of social institutions . I start
from the assumption that quantitative survey -based social research is here to stay, for the
foreseeable future , and it affords a powerful and potentially useful public argumentation repertoire
for sociologists . How can it be shaped such as to afford a more theoreti cally relevant and
humanistically oriented understanding of social processes ?
In her pungent critique of sociological research, especially of the quantitative flavor, Dorothy E.
Smith presents a recipe for “representing what people think” with three trick s, adapted from Marx
and Engels’ “The German Ideology” (D. E. Smith 1974) :
Trick 1. Separate what people say they think from the actual circumstances in which it is said,
from the actual empirical conditions of their lives and from the actual individuals who said it.
Trick 2. Havin g detached the ideas, they must now be arranged. Prove then an order among
them which accounts for what is observed.
Marx and Engels describe this as making ‘mystical connections’. [. . .]
Trick 3. The ideas are then changed ‘into a person’, that is they a re constituted as distinct
entities to which agency (or possibly causal efficacy) may be attributed. And they may be re –
attributed to ‘reality’ by attributing them to actors who now represent the ideas (p. 41) .
Quantitative research is easily recognizable, Smith also argues further, with its de -contextualized
survey responses that are statistically processed, in what she sums up as “intervening procedures
which it would be tedious to elaborate on here” . My research contributes to the detailed analysis of
these very procedures, as they are used in survey -based sociological research on Roma ethnicity and
scientific literacy , and to the discussion of their theoretical assumptions concerning social action in
general and interpersonal interaction in particular .
My first research contributions have mostly been in the fields of community development (Pop and
Rughiniș 2000; C. Zamfir and Rughiniș 2000; Larionescu, Rughinis, and Radulescu 1999; Rughiniș
2002a; Cozma, Rughiniș, and Sultănescu 2003; Rughiniș 2004a; Rughiniș 2004b; Rughiniș 2001;

12
Rughiniș 2005) and subjective well -being (Rughiniș 2007b; Rughiniș 2002b) . My recent research has
focused on ethnicity (Duminică, Lupu, and Rughiniș 2009; Rughiniș 2002c; Gabor Fleck and Rughiniș
2008; Rughiniș 2010; Rughiniș 2011a) , public understanding of science (Rughiniș and Toader 2010;
Rughiniș 2011b) , secularity (Rughiniș 2006; Máté -Toth and Rughiniș 2011; Rughiniș and Răuțu 2009) .
In the following sections I will elaborate on the significance of my recent research within a strand of
literature dedicated to understanding quantification in sociological research and its theoretical
implications . My contributions consolidate a relatively under -studied research thread that attempts
to make survey research relevant in light of interactional sociological theories.
In what follows I will first review the literature on sociological surveys as forms o f knowledge, and
the theoretical debates in which surveys are central topics , in order to chart the theoretical
landscape in which I situate my research . I go on to discuss survey -based research on Roma / Gypsy
ethnicity, highlighting the theoret ical persp ectives in this field, the theoretically -relevant
methodological debates , and the results of my analysis . I then move to survey research concerning
public knowledge of science, following the same steps: a review of the theoretical context in which
these surveys operate, then a discussion of several central methodological debates, and my
evaluation of the theoretical implications of alternative solutions. I conclude this section by
highlighting the theorizing affinities of methodological choices in survey -based research, and
proposing a revised survey -based research approach that supports an interactionally -sensitive
theoretical perspective.

13
2 Sociological theorizing in survey -based research
There are two areas of sociological theory for which surveys are of i nterest. On the one hand,
surveys produce data that may serve as evidence in light of a particular theory . For example, surveys
on ethnicity generate evidence for questions informed by sociological theories of ethnicity: how
large are social inequalities created by processes of ethnic differentiation ? What is the intensity of
ethnic differentiation in a given society, at a particular moment in time? What is the relevanc e of
ethnic categorization for a specific type of social action ?
2.1 A theoretical sketch: th e “trait” versus “interaction -order” divide
On the second hand, if we adopt a narrower focus , surveys are in themselves a form of social
interaction. Therefore, in order to use them as investigative instruments, researchers rely on a
theory of the survey i nterview situation – be it implicit or explicit (Cicourel 1964) . From a logical
point of view, the theoretical understanding of the intervie w situation represents a particular
instance of the theoretical understanding of social action and interpersonal interaction in general.
Still, the two need not be perfectly aligned in research practice , as I discuss below. I have drawn a
sketch of the two levels of theoretical orientations in Table 1. I have divided sociological theories of
social action into two types, with a coarse granularity that serves my specific purpose, rather than
offering a critical review of the state o f sociological theory. My purpose is to distinguish two
divergent theoretical approaches that are relevant to the practice of social surveys: a perspective
that focuses on individual traits (marked with [IT] for individual traits ), versus a perspective tha t
focuses on the interaction processes in the interview situation (Cicourel 1964; Maynard and
Schaeffer 2000; Hammersley 2010) (marked with [IO] for interaction order ). The first view is the
most generally used in quantitative sociological research bases on surveys. The second has been
used in methodological discussions of the interview situations, and also in research that studies
surveys per se, as social interaction, without attempting to improve their usefulness as research
tools.
An example of the individual trait approach used in research in a varieties of social research
disciplines, originating in social psychology, is the Theory of Planned Behavior and its extended form,
the Theory of Reasoned Action (Ajzen 1991; Madden, Ellen, and Ajzen 1992) . This perspective has
generated a rich body of survey -based research on di verse behaviors, also allowing for meta –
analyses (Sheppard, Hartwick, and Warshaw 1988) , and it continues to enjoy intense empirical and
theoretical attention (Ajzen 2011; Langdridge, Sheeran, and Connolly 2007; Trafimow 2009) . While
this theory incorporates social interaction as previous events that generat e norms and beliefs, it
aims to predict social behavior across different settings, without taking into account its interactional
organization. In other words, social interaction is always in the past, transmitted into present action
via its influence on actors’ beliefs and subjective norms.
Alternatively, interaction order theories see interaction as very much in t he present tense of actors’
actions: behavior is understood as partaking in a local, contextual order which it c reates and to
which it orients (Goffman 1983; Sacks 1995; Warfield Rawls 2011; Warfield Rawls 1987) .
Ethnomethodology and the current s that it has inspired, such as Conversation Analysis (CA) and its
thread of Membership Categorization Analysis (MCA), or Discursive Psychology (DP), attempt to
understand social interaction as situated cooperative accomplishment. From this perspective it does
not make analytical sense to occasion an interaction, with a given structure of relevances (such as

14
the one imposed by the activity of filling in a questionnaire), in order to understand other
interactions, in completely different settings and with alien structures of relevances (such as voting).
In experimental methodological parlance, the issue of representative design (Hammond 1998)
becomes critical. A second major differen ce between trait -based theories and interaction -based
theories refer to their views of language and communication. While trait -based investigations rely
on language to express inner, subjective states (such as beliefs, desires, opinions, attitudes, values,
preferences and the like), from an interactional perspective language use is understood, following
the late Wittgenstein’s philosophy, as a pragmatically oriented interaction, that invokes and
attributes mental states as elements of intelligible accounts of actions. Words lose any
correspondence with alleged inner individual entities, and they become virtually meaningless when
abstracted from interaction (to the extent that this is ever possible).
One of the groundbreaking works that look s closely at surv ey-based research from an
interactionally -sensitive theoretical perspective is Aaron Cicourel’s “Method and measurement in
sociology” (1964) . Ethnomethodology and its related theoretical threads have also supported a
substantial body of research on survey interactions (Maynard and Schaeffer 2000; Maynard, Freese,
and Schaeffer 2010; Moore 2004; Potter 2003; Roulston 2006; Antaki 2006) . Given their widely
divergent theoretical understanding of social action, this corpus of literature is , as a rule, not
convertible into methodological advice for survey researchers. All the same, it is illuminating as
regards the concr ete, empirical work that is done to achieve “standardization” and to carry on the
work of interviewing successfully, in collaboration with the respondents, in the particular settings of
the interaction – and this understanding may carry practical, although implicit or indirect, advice for
practitioners (Houtkoop -Steenstra 2000) . If any gener al conclusion can be drawn to serve the
practical interests of the survey operator, it is that trouble arises when participants in interaction
depart from regular conversational conventions (which may happen if strict standardization is
enforced) , and such trouble requires elaborate repair work if interaction is to go on.
A theory of survey interaction that conceives of respondents’ answers as reflections of their inner
states is not compatible with a general perspective on social action that privileges the interaction
order (Warfield Rawls 2011; W arfield Rawls 1987) ; reciprocally, such a general perspective cannot
accommodate a theory of the survey interaction that sees language as an expression of private
thoughts. Therefore, two cells in Table 1 are empty. The middle co lumn is the one of interest for me.
The upper cell accommodates research that accepts the interactional co -authorship of survey
answers, but still attends to the problem of measurement accuracy, conceiving of answers as
ultimately describing the respondent . Such research often draws on the insights of Conversation
Analysis, recommending practices of “conversational interviewing”, in which operators clarify
meaning and attend to possible contradictions and ambiguities as they occur in discussion (Conrad
and Schober 2000; Schaeffer and Presser 2003; Schober and Conrad 1997; Suchman and Jordan
1990) . The lower cell is the one I aim to explore in this thesis: a practice of survey -based socio logical
research that accommodates an interactional unders tanding of the survey interview and of social
action and interpersonal interaction in general .
In the following pages I will discuss survey -based sociological research, highlighting its frequent
affinity with individual -trait theories of social action (the IT – IT’ cell in Table 1), while aiming to
explore the possibility of an interactional ly-aware use of this instrument (the IO – IO1 cell in Table 1).
I will present my research contributions in analyzing methodological debates in survey research on
ethnicity and on public knowledge of science, and I will indicate their significance in opening a

15
possibility of a research agenda that employs surveys in a theoretical framework attentive to the
interaction order (Warfield Rawls 2011; Warfield Rawls 1987) .
In Rughiniș (2007b) I have argued for the heuristic value of analogy in common reason and
sociological thinking (pp. 63 -70), and in this thesis I illustrate this point by a graphical representation
(in Figure 1) and an analogy -based heuristic (see section 3.2.5 ). For what is worth, they have
definitely helped me in elaborating this material, and I hope they will also be helpful in rendering the
message as clearly intelligible as possible, for an audience with diverse theoretical orientations and
substantive research interests. In Figure 1 I have sketched the main three theoretical approaches to
survey -based sociological knowledge that I address in the following pages:
1) On the first column, I have represented the individual -trait theory of social action, together
with the individual -trait theory of survey response;
2) On the second column, I have represented the slightly dissonant combination of an
individual -trait theory of social acti on with an interactionally -aware theory of survey
response;
3) On the third column, I have represented the combination that I aim to clarify in the present
analysis: a consistent understanding of survey response and other forms of social action as
fundamental ly shaped in interaction.

16
Table 1. Overview of theoretical frameworks for survey practice in sociology

Theory of survey
interaction
Theory of
social action Respondent’s answers are an expression
of individual traits (“conduit v iew” of
communication: answers describe
respondents)
[IT’] Respondent’s answers are co -authored with the interviewer and other participants, in a concrete
interaction situation ( respondents and interviewers are participants in an interaction order )

Survey events have a trans -situational relevance
Respondents’ answers make intelligible other
situations
[IO1] Survey interviews have a situational
relevance
Respondents’ answers are highly situated
and do not make intelligible other situations
Survey interac tions may be studied as
research topic
[IO2]
Individual traits have a causal influence on
individual actions
– Traits are shared, and acquired in
socialization Main methodological concern:
measurement accuracy (construct
validity, reliability)
Solutions i nclude:
– Questionnaire pretests
– Standardization of interviewer
actions and interview settings
– Statistical procedures for
managing errors and missing
data Main methodological concern: measurement
accuracy (construct validity, reliability)
Solutions include:
– Questionnaire pretests
– Meaningful conversational
interventions of interviewers
– Taking into account the interview
interaction (eg: traits of the
interviewer)
– Statistical procedures for managing
errors and missing data
[IT] Configurations of individual tr aits have a
causal influence on individual actions
– Social action is intelligible by linking
it with types of actors
– Types of actors are defined by the
sociologist (ex: as clusters, latent
classes)
Individual actions are performed and
understood as cat egory -bound
– Social interaction makes use of
publicly available categories of
actors
– Types of actors are available as a
resource to members Main methodological concern: representative
design (relevance of the interview interaction for
the life settings in which the topic under study
takes place)
Solutions include:
– Questionnaire pretests
– Meaningful conversational
interventions of interviewers
– Taking into account the interview
interaction (eg: traits of the
interviewer)
– Adjusting data analysis and
interpretat ion to the actual events that
are available as evidence
Shifts in analytical vocabulary:
– Descriptive rather than causal
– Acknowledges cross -sectionality
– Types are preferred over traits Survey interactions are analyzed as situated
social interactions (in eth nomethodology,
conversation analysis, discursive psychology)

Surveys do not illuminate life in other
settings .
[IO]
Individuals orient their actions to the situated
interaction order
– Individuals appeal to institutional
orders and to accounts, as
resource s for situated interaction
Individuals in interaction:
a) Act as competent members
b) Are oriented to interactional goals
c) Act intelligibly and accountably

17
Figure 1. Sketch of three alternative theoretical views of the interview sit uation in relation to the behavior under study

[IT-IT’] R.’s a nswers describe respon dent’s
traits
Interviewers are factored out [IT-IO1] R’s a nswers are co –
authored
representations of
traits [IO-IO1] R’s a nswers are interactional
achievements
R’s a ction is influenced by these traits R’s a ction is influenced
by these traits R’s a ctions are interactional
achievements

18
2.2 Checkpoints on the “trait” vs. “interaction order” divide
The views on the world afforded by “trait” theories of social action are v ery different from those
afforded by interactionally oriented theories. Their vocabularies are widely divergent, pointing to
irreconcilable ontologies. Survey -based social research is, as a rule, easier to conduct in light of a
trait -based conception of pe rsons and social action. Still, despite the instruments’ strong theoretical
inclination in this direction, I argue that there are forms and styles of survey -based research that
may accommodate an interac tional ly focused approach .
A background question is, why would one conduct a survey -based research, when working with an
interactionally oriented theoretical perspective ? Indeed, a standardized interview applied to a large
sample would most likely not be the research instrument of choice in this framework , unless one
studies the interview situation itself. Still, there are cases in which an analyst may find surveys useful
for the purpos e at hand. Among others, survey -based research can afford:
1) Production of scale2: analysts obtain useful grounds for speaking about large numbers of
people, often across a considerable territory and in distant social settings;
2) Typologies and classification: survey -based research is a powerful instrument for
commenting on common -reason types, and for proposing revised and novel ty pes as being
relevant for understanding. In association with the power of statistical ly representative
samples , surveys allow analysts to classify large numbers of people and to count
membership in each category, and to compar e types on various criteria; g iven the centrality
of typification in common social knowledge, surveys have much to contribute to public
knowledge precisely because they facilitate the production, co unting, and comparison of
types;
3) The production of reasonable certainty and uncertainty that can become basis for action, as
opposed to incapacitating confusion; the quantitative vocabulary is a powerful rhetorical
tool for communicating precision (even when precisely communicating imprecision), and it
creates the possibility of strong, pers uasive knowledge claims in the public arena;
4) Effective communication across the discipline: surveys are particularly useful tools for
presenting evidence to public administration bodies, to various organizations interested in
large -scale social processes, and to the general public; they offer a familiar vocabulary which
gives analysts leverage in often unequally balanced relationships.
A second question then follows: how can a sociologist engage with survey -based research, while
working under the auspices o f an interactionally -oriented theory, in order to take advantage of their
classification facilities and communicative force? Or, to put it in other terms: how can one work
meaningfully with surveys, when attending to the creative and methodical work of pa rticipants in
social interaction?
When looking for the answers already formulated, this inquiry highlights a rather fine zone of the
methodological literature. Many starting points and pieces of advice are to be found in papers that
are usually positioned as critical views of quantitative research – while still engaging with the

2 This point was formulated by Puiu Lățea in an informal discussion on survey -based research.

19
enterprise and its intricacies. Such an undertaking may seem like riding two horses at once. As
Hammersley (2010, p. 410) observes, in a paper that illustrates this type of inquiry ,
While the problems surrounding measurement have often been acknowledged by
quantitative researchers, they have usually been treated as technical in character; in other
words as susceptible to remedy through further refinements in measurement technique (see,
for example, Bulmer 2001). By contrast, critics have often taken these problems to indicate
the fol ly of the whole enterprise. (…) So, the assumptions about the nature of social inquiry
built into much qualitative research and much current social theory are at odds with any
conception of social science that puts rigorous measurement at its core.
The deb ate is indeed complicated because it rests on an ‘assumptions war’, on discussing the
fundamentals that are taken for granted in order to make sense of survey results. When disagreeing
with the basic tenets that justify the application and refinements of a method, it is difficult to engage
in arguments on how to use it meaningfully. I believe that t his is the place in the debate where it is
useful to construct a scaffold that somehow spans the divides. Such a bridging structure is
inescapably selective, foc using on some points of discontent at the expense of others. I propose
several critical issues to orient the design of an interactionally -aware survey -based sociological
research (see Table 2). These issues are linked to various f orms of survey trouble, such as “errors”,
lack of respondent cooperation, and public controversies.
Table 2. Checkpoints on the “trait” versus “interaction -order” divide
Type of survey
trouble Interaction processes
that lead to this type
of trouble Survey -based
solutions Specific devices Stage of survey
research
Contradictory
answers

Inconsistencies Pragmatic orientation
to the specificity of
the interaction
situation

Regular methods of
conversation -making:
approximation,
imprecis ion Framing
inconsistencies as
errors
The error -trait
device :
Statistical techniques
for dealing with error
(aggregation, data
reduction such as
factor analysis)
Questionnaire design:
scale selection or
construction
Statistical analysis
The error -practicality
device: re -focusing
attention towards
practicality (is
measurement
precision reasonable
enough to be used?)

Data analysis, survey
reports, scientific
literature
The sampling error
focus : centering
argumentation on
issues of statistical
representativity
(addressed by
statistical
significance), at the
expense of
measurement error Data analysis, survey
reports, scientific
literature

20
Type of survey
trouble Interaction processes
that lead to this type
of trouble Survey -based
solutions Specific devices Stage of survey
research
Non-response
“Don’t know” Mismatch between
researchers’ and
respondents’
structures of
relevance or
propriety o f
conversation Conversational
interventions to carry
the interview to the
end (accounts)

Probing
Interviewing process
Imputation
techniques Data analysis
Contested typologies Using and
commenting on
common -reason
typologies Mixture of technical,
ethical, theoretical
considerations that
favor a certain
typology Resonating typologies Questionnaire design
(choice of categories)

Data analysis and
reporting
(category
construction)
2.2.1 The error – trait device
As discussed above, many survey questions are designed and interpreted as measurements of
individual stable traits, which are assumed to exist and, also, to be (quasi -)causally effective. The
error – trait – cause triangle consists of the following three activities:
1) Looking for traits as relevant, age ntive qualities of individuals;
2) Finding traits by operating with the error device , as detailed below, to decontextualize
responses and attribute them to respondents;
The focus on errors is very important for the efficacy of this device; errors are a core c oncern in
mainstream methodological texts on survey -based research. To illustrate the connection between
the concept of error and the search for traits, I quote a definition of random measurement error –
one of the basic topics of methodological improvemen ts in survey research (Viswanathan 2010, p.
287) :
In essence, a way to visualize ran dom error conceptually is to examine whether inconsistent
responses are provided across time (or items) when the phenomenon in question has not
changed (such as an enduring trait). Some causes of random error include complex wording
or language, questions requiring estimation, vagueness in questions or response categories
(Churchill, 1979; Nunnally, 1978), the nature of administration through distracting factors
and inconsistent administration procedures (cf. Churchill, 1979), and personal factors such as
mood.
This definition, which is a regular working instrument in a methodological discussion of improving
measurement validity and reliability, opens up a series of questions from an interactional
perspective in which responses are “inconsistent” only insof ar an observer defines them as such,
pragmatically (and, in conversation, rhetorically) by deciding that the ceteris paribus clause applies,
here and now, and that responses should then be the same. The indicated “causes of random error”
are a heterogeneou s list, as regards their treatment under an interactional frame. The question is,
what counts for the “sameness” of the “phenomenon”, and the “consistency” of the “response”? For
example, as regards wording, is the “same question” formulated with “differen t words” actually the
“same question”? It is clear that, for some practical purposes, it is not – while for others it is. The
practical purpose attended by survey analysts often consists of unveiling “enduring traits” beyond

21
“responses” that are taken to b e relatively unstable. That these traits exist, that they are knowable
and measurable (representing the ‘true value’ sought for), and that they are “enduring” is, as a rule,
taken for granted. “Error” works as a rhetorical device that a) decontextualizes the contingent,
situated survey response given by the respondent to the interviewee’s question by comparing it to
other responses, and b) attributes it to the respondent, as a stable attribute, with a byproduct of c)
uncertainty . Overall, the “error” device converts situated, methodical, open social interactions into
stable individuals known to us, the analysts, albeit with a certain degree of uncertainty (as
represented in Figure 3).
Figure 2. The “error -trait” device in mainstream methodological treatment: decomposition of observed responses into
true values and random errors

Figure 3. Working with the “error -trait” device from an interactional perspecti ve: conversion of situated interactional
events into individuals (respondents) described by stable attributes

2.2.2 Quasi -causes versus methods and resources
Traits are usually referred to as causes or quasi -causes for various outcom es (behaviors, other traits,
resources etc); by quasi -causes I refer to the use of a causal vocabulary without a clear commitment
to a causal epistemology. If ethnicity is reported to “influence” behavior or to have some sort of
“effects”, this represents an instance of a quasi -causal approach.
I propose that an interactionally oriented vocabulary may usefully replace such causal terms with a
vocabulary of actions, methods , and resources. From this perspective, ethnicity may occasionally be
invoked as a res ource in interaction, and it may thus facilitate or hinder an interactional outcome.
This move also supports a prudent take on nominalization in social research terminology, as
convincingly advocated by Billig (2008) .

Researcher’s
uncertainty Individuals defined
by stable &
knowable attributes Situa ted
interactional
outcomes
Observed response = True value of
individual attribute + Random
error

22
2.2.3 Imposed r elevance
From an ethnomethodologically informed perspective, researchers should be wary to impose their
structure of relevance on respondents – by imputing them concerns or perspectives. For example, as
regards social classification, any individual may be seen from an unli mited number of perspectives –
by pointing out her gender, age, subcultural style, occupation, social class, beauty, ethnicity, and so
on. In concrete, situated interactions, some of these classifications are brought out by members
while all the others are ignored . Sociologists often re -describe actions by reference to categories
that were not observably attended to by participants: for example, all sorts of interactions may be
analyzed with reference to participants’ gender, even if gender was not an obser vable concern for
participants. There is a consistent debate as to what counts as an observable concern in interaction,
whether some categories may be considered as omni -relevant or not, and what professional
discretion the analyst has in choosing her anal ytical categories (M Billig 1999a; Emanuel A Schegloff
1999; M Billig 1999b; E A Schegloff 1999; Hammersley 2003; Potter 2003a) .
In the context of this debate, standardized interviews cannot but impose a structure of relevance on
the interaction and on analysis, since there is no possibility of observing the spontaneous
orientations of members towards one another in the natural situations of interest. If a survey
investigates the relationship between ethnicity and employment, there is no saying whether and
how eth nic categories are actually invoked by participants in those work -related interactions that
are presumably described , in the aggregate, by a correlation between ethnic affiliation and income
or employment status. It is then the analyst’s responsibility to justify the relevance of a certain
classification – by demonstrating that in actual interaction situations that classification is used, by
members, to orient their actions. If there is empirical evidence that ethnicity is relevant in
recruitment decisions, this may be used to justify estimating a correlation between ethnicity and
employment status, for example.
2.2.4 Resonating typologies
The relationship between members typologies and sociologists’ classifications may be understood by
reference to Schütz ’ distin ction between first -order and second -order constructs (Schütz 1953; Kim
and Berard 2009) . This relationship, although fundamental for understanding soci al research as a
social practice and as a form of knowledge, is relatively under -theorized, especially in survey -based
research. The temptation of looking for clear -cut, comprehensive and mutually exclusive classes in
second -order typologies, afforded by q uantitative data, is at odds with an understanding of how
first-order categories operate in social interaction:
“[C]ultural categories generally seem to operate on the basis of family resemblances, with
people using prototypes or exemplars as a basis for d etermining what counts and does not
count as an instance (…) So, a first problem facing attempts to produce classically rigorous
classifications stems from the fact that, at some level, social research will always depend
upon social scientists’ ordinary cu ltural capabilities; that it may not be possible to explicate
these fully in propositional terms; and that these practices depend upon flexible, context –
sensitive categorization. Furthermore, the meaning of any cultural category is usually
context -sensitiv e: what it includes, and does not include, depends upon the context in which
it is being used, including the purposes it is serving. Not only will people interpret the same
category in somewhat different ways according to circumstances, but also what level of
clarity id required may vary in the same way. (…) The character of people’s everyday
categorizations has consequences for social science in a second way, too. This arises to the
extent that, in order to describe and explain their behavior, we need to i nclude in our

23
accounts some representation of the categories that underlie people’s discriminations among
situations, differentiation across types of other people, identification of strategies available
for use, and so on. And this is surely unavoidable in social science. The crucial point is that, if
everyday categories have a flexible, fuzzy, context -sensitive, character, then we should not
pretend that they can be incorporated into analytic categories that have an Aristotelian
form: to do so would be to introduce distortion” (Hammersley 2010 , pp. 418 -419)
Not only that researcher ’ typologies are grounded in commo n reason distinctions and classifications,
but they are continuously re -appropriated by common reason social thinking, crossing disciplinary
boundaries. This raises an issue of anticipative design of second -order types: as sociologists, how are
we to take into account the expected public reception and use of our constructed typologies? To
what extent we can shape their re -appropriation by theoretically -informed decisions concerning
data analysis and reporting?

24
3 Theorizing ethnicity in surveys of ethnic min orities. The case of
research concerning the Romanian Roma / Țigani

In this section I will present the significance of my research contributions for an attempt to
reposition survey -based sociological research in an interactional theoretical outlook. I start by
presenting two alternative theoretical standpoints concerning empirical research on ethnicity; in
order to clarify their practical consequences, I discuss an analogy that illustrates the interactional
perspective on ethnicity; I then analyze their affinities with specific methodological choices in
survey -based research.
3.1 Theoretical views on ethnicity
In the methodological literature dedicated to the quantification of ethnicity , there is widespread
discontent concerning the links between theory and measurement . A series of r ecent contributions
aim explicitly to address this mismatch (for exampl e Aspinall 2001; C. W. Stephan and Stephan 2000;
Aspinall 2009; R ughiniș 2011a; Rughiniș 2010; Burton, Nandi, and Platt 2010; Brown and Langer
2010; Phinney and Ong 2007) . Solutions depend on the theoretical problems identified by the
authors; still, there is unanimous appreciation that a categorical variable with seve ral categories
which the respondent can check does little service to a complex reality. Categorical measurements
of ethnicity have two serious theoretical drawbacks (Brown and Lan ger 2010) . On the one hand, they
seem to support a primordialist view of the common reason variant that asserts ethnicity as
membership in discrete, stable, essentially different groups with specific cultures. On the second
hand, given that categories are often taken from administrative use, they serve to reinforce the very
inequalities and power relations that they purportedly study.
The authors take three main directions when recommending solutions for a better fit of
measurements with theories . The firs t direction, most often taken in relation to administrative
measures, refers to “improving on categories”: allowing respondents to self -identify, refining
categories to reflect current usage , allowing for multiple affiliations, adding categories on other
dimensions (Aspinall 2001; Stephan and Stephan 2000; Aspinall 2009) . Secondly , there i s the
internalizing direction, which conceptualizes ethnicity as an identity construct and propose
additional dimensions of the subjective experience of ethnicity, to be measured in addition to
categorical affiliations (Rughiniș 2011a; Burton, Nandi, and Platt 2010; Phinney and Ong 2007) . An
illustrative example is provided by the Multigroup Ethnic Identity Measure – Revised scale (MEIM -R)
(Phinney an d Ong 2007) , that includes the following items besides categorical self -identification ,
measuring the dimensions of ‘exploration’ and ‘commitment’ (p. 276) :
1. I have spent time trying to find out more about my ethnic group, such as its history,
tradition s, and customs.
2. I have a strong sense of belonging to my own ethnic group.
3. I understand pretty well what my ethnic group membership means to me.
4. I have often done things that will help me understand my ethnic background better.
5. I have often tal ked to other people in order to learn more about my ethnic group.
6. I feel a strong attachment towards my own ethnic group.

25
Thirdly, there is the externalizing direction, whic h is considerably less studied i n its implications for
measurement . This approac h sees ethnicity as a practice, rather than an individual attribute , and
locates it firmly in interaction rather than within individual minds. Related theoretical development s
have been formulated by Wimmer (2008) , Brubaker (Brubaker 2004; Brubaker, Loveman, and
Stamatov 2004; Brubaker 2002) , Hale (2004) drawing on a substantial tradition that gained
momentum after the seminal work of Barth (1969) , who challenged the definition of ethnic groups
as ‘culture -bearing units’ based on biological self -perpe tuation, focusing on maintenance of social
borders as the key definition of ethnicity (idem, pp. 10 -15).
One key issue brought forward by this conception is that of the variable relevance of ethnic
classifications : under what conditio ns, and how do people appeal to ethnicity to orient their
interactions? From this perspective, e thnicity is only relevant when people make it relevant in
interactions, displaying ethnic affiliations and orienting towards them. The issue of salience becomes
the main empirical c oncern: how is ethnicity done, in social interchange?
The situational production and thus variability of relevance is of course a blind spot for the
researchers who go around inquiring about ethnicity, since they actually bring it about by their very
inve stigative orientation (Moerman 1974) . Empirical research on the situated interactional
constitution of ethnic identity is mostly conducted in the Conversation Analysis (CA) and its
Membership Categorization Analysis (MCA) thread , and Discursive Psychology (DP) traditions that
look at naturally occurring conversations and observes how participants methodically display their
occasioned ethnic identities and orient towards them in action (Hansen 2005; Schilling -Estes 2004;
Wilkinson 2011; Augoustinos and Every 2007; D. Edwards 2003) . Surveys are definitely not the
instrument of choice for inquiring into the salienc e of a particular classification, since they most
often provide classifications for the respondents in closed -ended questions. Still, an attempt to use
surveys to investigate the salience of particular ethnic categories, in the broader frame of
discussions of ethnicity, is presented by Aspinall (2012) , using open ended ethnic self -affiliation to
inquire into ethnic “superdiversity”. As regards aggregate measures of ethnic differentiation, mostly
used within comparative politics and econometrics, a methodo logical proposal that attends to the
issue of salience belongs to Chandra and Wilkinson (2008) who distinguish between measures of
‘ethnic structure’ and ‘ethnic practice’ (p. 523):
Ethnic structure refers to the distribution of descent -based attributes —and, there fore, the
sets of nominal identities —that all individuals in a population possess, whether they identify
with them or not. Ethnic practice refers to the act of using one or more identities embedded
in this structure to guide behavior.
A second important is sue in understanding ethnicity as interactional resource concerns its
occasioned use in troubled interaction . Ethnicity offers participants accounts of group differences, of
divergent interests, conflicting cultures, of essentially different kinds of peopl e – and such accounts
and identit y categories are particularly useful when interaction goes awry (Garfinkel 1945) ,
becoming part of processes of Othering (Warfield Rawls and David 2006) . Of course, not any
interchange gone wrong elicits ethnic identification and a ttribution – raising again the issue of
salience. Still, if this is put aside under the assumption (preferably based on specific empirical
previous knowledge) that salience of ethnicity is high, in a specific social context, it makes sense to
inquire into ethnicity as a correlate of social distance, segregation, and disparity. Such disparities
offer structural opportunities for troubles interactions and, insofar as they are available in public

26
stereotypical accounts, they may become resources for dealing wi th them, often at the expense of
the disadvantaged party.
Taking one step further, Brown and Langer (2010) propose to measure eth nicity (at aggregate levels)
not in itself, but to include measures of ethnic diversity and ethnic disparity as indicators of social
distance. The authors make this move aiming to address the theoretical inadequacy of categorical
measurements in relation t o current theorizing of ethnicity; their bold statement, which is worthy of
detailed analysis in itself, is that:
[I]f we reconceive of ethnicity as an indicator of social diversity and disparity, the fact that
the data we typically employ in the measureme nt of this indicator are neither as fluid nor as
multidimensional as theorists of ethnicity qua ethnicity contend is less problematic because
we take this as an issue of measurement bias and measurement error: bias, because the
categories we use are acknow ledged to be those legitimized by historical state practices;
error, because the exclusive and exhaustive codings employed do not fully capture the
nuance of ethnic identity. It may seem perverse to contend that some of the theoretical
problems with measur ing “ethnicity” are resolved by acknowledging measurement bias and
measurement error, but our contention is that this is less problematic than reconciling
“primordial” data with constructivist concepts. Measurement error and measurement bias
are inherent f eatures of virtually all quantitative variables (…) (p. 417).
As a side comment, Brown and Langer’s position is a telling illustration of a rhetorical error –
practicality device (see Table 2): it literally makes theoretical proble ms fade away. In all fairness, the
authors have also proposed a theoretical solution, basically by subsuming ‘measured ethnicity’ as an
indicator of social distance, and refraining from treating it as an indicator of ‘theoretical ethnicity’.
The error devi ce is brought in to make this solution work: a measurement with error may still be
used, as a reasonable approximation for a task at hand, while an inadequate classification is at best
useless and at worst undermining. Political bias is converted into meas urement bias, while imposed
relevance is converted into error due to nuance -conflation.
An interactional perspective on ethnicity does not preclude an interest in quantitative research on
this topic. Such a theoretical stance keeps open an interest in ethn ic inequalities, which are seen as a
result of repeated social interactions that are oriented towards ethnic classifications; the ASA
statement on The Importance of Collecting Data and Doing Social Scientific Research on Race is a
clear formulation of such a research agenda (American Sociological Association 2003) . The question
arises, then, how can such a theoretical orientation be pursued with survey instruments. In order to
address this question, I look at survey research and I inquire into its theoretical implications.
Specifically, I discuss methodological debates, and technical discussions, and I ask what theoretical
perspectives are (often implicitly) supported by altern ative solutions.
In order to analyze the theoretical bearing of methodological discussions concerning survey
research, I will explore the three critical issues that I have proposed as a scaffold: the error – trait
device , the quasi -causal vocabulary, impos ed relevance, and the use of resonating typologies.
3.2 Theoretical bearing of survey -based methodological discussions
3.2.1 The error – trait device in survey -based research on ethnicity
Ethnicity -related variables , the main unit in survey -based information product ion, are usually
interpreted as ethnicity -related attributes of the respondents (or of higher order units such as
groups, communities, as the case may be). Therefore, survey -based research affords easily an

27
interpretation of data as evidence for difference s among individuals concerning their attributes. The
social interaction that has produced the recorded answers is, as a rule, methodically erased by
standardization guidelines and statistical analysis. Moreover, attributes are often considered to be
stable across situations; that is, they are interpreted as stable traits. The question then arises, how
can survey -based research accommodate an interactional perspective on social phenomena?
In relation to this distinction, we can classify theoretical perspecti ves on ethnicity in two broad
strands:
A. Ethnicity as difference: ethnicity is a specific cultural profile of individuals, produced by
socialization in an ethnic community
o Ethnicity as a differentiation of humankind in essentially different, stable cultures :
 Individual ethnicity as (public) membership in discrete, essentially different
cultures ;
 Ethnic differences in patterns of actions or resources may be causally
attributed to the influence of ethnicity on individual behavior ;
 It is often encountered in com mon -reason accounts of ethnic differences .
o Ethnicity as socially (historically) constructed and ever -changing shared identities
available to individuals in their life course :
 Individual ethnicity as private affiliation to a community of cultural
relevances ;
 Ethnic differences are a result of socially visible patterns of actions of people
who affiliate with ethnic communities (ethnicization of social differences) –
and may become relevant and thus influent in interaction .
B. Ethnicity as discursive resource for differentiation: ethnicity is invoked to make sense of,
demand, and account for differentiated interaction
o Ethnicity is a common -reason discursive resource for making sense of differences
and troubles in interaction; ethnic categories and accounts of ethn ic attributes are
invoked in interaction in order to signal and manage boundaries, incommunicability;
thus, ethnicity is not a cause, but a common -reason theory (with practical relevance)
of social difference and distance.
o Ethnicity represents a work of classification in inter -subjective boundary
maintenance processes :
 Individual ethnicity as outcome of inter -subjective classification processes;
 Empirical concern: when do people invoke ethnic classifications?
 People may orient to ethnic categories in intera ction, or may attend to other
classifications (ethnicity is not necessarily omnirelevant in action although
some ethnic / racial classifications may be very often relevant) ;
 Accounts of ethnic differences are invoked as resource in interaction – and
may be thus reproduced or challenged in interaction and the locally
produced accounts ;

28
 Sociological research on ethnicity contributes to the common –
reason production of accounts of ethnic differences .
3.2.2 Ethnicity as quasi -cause of action versus resource for action
In survey -based research employing ethnicity there is widespread use of a causal vocabulary,
including expressions such as ‘influence of ethnicity ’, ‘consequences of ethnicity ’, or ‘effects of
ethnicity ’. This vocabulary is widespread in epidemiology, an d it is also present in social research
within sociology, social psychology, or social work (see, for example, Wink 1997; Lott and Saxon
2002; Pellebon 2000; Connolly 2011) . For what is worth, epidemiological uses of causal terms have a
different grounding than those related with non -biological processes, since e thnicity is used as a
proxy for shared heritage and thus for shared biological traits (genes or other configurations). The
use of such expressions in sociological research does not necessarily mean that the authors adhere
to an epistemological position tha t favors causal explanations in social theory, or that they see
individual ethnicity as a causal influence. ‘Influence of ethnicity’ (in a different meaning from the
influence of inherited biological traits) may refer, among others, to different processes such as:
– Causal influences of ethnically specific individual traits on behavior or other individual
conditions , under a model that causally links behavior to mental states , including ‘ethnic
identity’ or ethnically specific values, attitudes, beliefs etc ;
– Results of social actions that orient towards ethnic classifications – referring to the social
organization of ethnicity, and not to i ndividual ethnic identification;
– Results of social actions that make use of resources which have been acquired in social
processes that are oriented towards ethnic classifications – and for which, therefore,
ethnicity has a ‘second -hand’ relevance.
The use of causally -loaded vocabulary is favored by particular statistical analyses such as regression
and path models, which oft en rely on concepts as ‘explanation ’ and ‘ effect ’ as part of the statistical
jargon. It is therefore relevant to ask to what extent such technical terms are used with theoretical
implications, and what sort of implications are favored by various styles of analysis.
An alternative vocabulary, supported by an interactional perspective, conceptualizes ethnicity as
resource for action. To make use of an analogy, the distinction between ‘influence on’ and ‘resource
for’ action is easy to make when asking about t he relationship between happiness and money. Does
money bring happiness? This formulation points towards the presence, or absence, of an influence
of money on happiness – somehow irrespective of the person whose money or whose happiness is
at stake. The va st body of conversations, quotes, humor, and other reflections on money and
happiness is partly an indirect take on the agency that can bring about happiness: where is it
located? In material structures? In personal capabilities – be they intellectual, spi ritual, emotional,
cognitive or of other kind, depending on the relevant classification? In interpersonal relationships? In
broader, community and society -shaped processes? The rhetorical contrast of money and “buying”
versus happiness is often used to poi nt to human agency and interactional outcomes, against
impersonal forces. An alternative rhetorical use of the money / happiness pair keep s money as a
happiness -relevant fact but conceptualize s it as a resource, thus re-delegating agency to the actors
who use this resource, and to their skills and orientations. A quote that illustrates this take on the
distinction belongs, maybe ironically, to Henry Ford: “The object of living is work, experience,
happiness. There is joy in work. All that money can do is b uy us someone else's work in exchange for

29
our own. There is no happiness except in the realization that we have accomplished something”
(quoted in Leagans 1964, p. 96) .
We can therefore distinguish two broad perspectives and associated vocabularies concerning the
relevance of ethnicity for understanding social action :
A. Individual ethnicity shapes actions in accordance with ethnic community values : it
has an internally causal influence on behavior . Therefore, appropriate statistical
analysis may include regression models in which effects of ethnicity are estimated by
controlling fo r other influences (education, income, gender, age etc).
B. Individual ethnicity is used by participants in interaction to make sense of their
different orientations and to account for interactional troubles. In this perspective,
ethnicity is not a cause of distinctive behavior or of interactional conflicts – but it is
an interpretive resource used to signal differences, to mark relevant aspects of the
interactional context, and to account for the success or, often, the failure of the
(inter -ethnic) interact ion.
o Individual ethnicity is displayed and used by participants in interaction to
account for their actions; thus, ethnicity is continuously re-defined by
differential treatment and different actions . Ethnicity is relevant for
understanding interactional o utcomes and it correlates with them, but it is
not an internal cause of individual behavior ; its force resides in its
interactional deployment . It follows that ethnicity should not be reported as
a cause of outcomes, but as a correlate of outcomes.
o Ethnici ty is used in interaction to re present types of people, jointly with
other contextual elements (including individual attributes ) that are available
for participants; thus , ethnicity should be investigated in its interaction with
other attributes to define types of persons. Therefore, it often makes no
analytical sense to control for other variables in order to estimate the pure
effect of ethnicity.
Table 3. Survey -based research on ethnicity under alternative theoretical perspectives on ethnicity
Issue [1] Ethnicity as shared attribute with causal
influence on behaviour [2] Ethnicity as shared concern and sense –
making resource in interaction

Ethnicity as
membership in
discrete cultures Ethnicity as identity:
private choice of
affili ation to
communities of shared
cultural relevance Ethnicity as common –
reason theory of
social difference and
social distance, as
sense -making
resource for dealing
with troubled
interactions Ethnicity as
occasioned, inter –
subjective
outcome of
classificatio n in
interaction
Observability Ethnicity is
conspicuous , public,
and may be hetero –
observed Ethnicity is private and
needs to be self –
assessed Ethnicity is an inter -subjective classification
and it is best measured by both self -reports
(ethnic affiliat ion) and hetero -reports
(ethnic classification) in face -to-face
interaction
Differences between self -reports and
hetero -reports are a relevant topic of study
concerning the interactional organization of
ethnicity

30
Issue [1] Ethnicity as shared attribute with causal
influence on behaviour [2] Ethnicity as shared concern and sense –
making resource in interaction
Type of measurement -Categorical variable
-Categorical variable
indicating (fuzzy)
affiliation (with open
and closed answers)

– Labels may be
multiple and
overlapping

-Dimensional variables
indicating salience and
strength of
identification on
various criteria – Categorical variables with open and closed
answers

– Labels may be multiple and overlapping

– Additional variables may include modifiers:
certainty of classification for all those
involved, ethnicity of family members, types
of residential community
General research
questions -What ar e the effects
of ethnicity on
behavior? -How does ethnic
identity orient actions
in a variety of settings? – What are the
patterns of
incommunicability, of
difficulties in
interaction that are
associated with ethnic
differences?
-How is ethnicity
used by
participants to
orient in
interaction and to
account for their
actions?
-How is ethnicity
accomplished in
daily life?
Specific survey -based
research questions – What is the influence
of ethnicity on
behavior, controlling
for other attributes? -How does e thnicity
correlate with various
attributes that are
publicly seen as
relevant for ethnic
distinctions? (ethnic
gap models,
identification of types
of actors)

-How do outcomes of
particular interactions
(where ethnicity plays
an interactional role –
such a s employment,
income, school grades)
vary systematically
with ethnicity
independently of other
attributes (ethnic
discrimination models) – What are the
aggregated disparities
associated with
ethnicity, associated
with specific settings
of interaction? -Wha t types of
persons are
visible when
taking ethnicity
into account,
together with
other relevant
attributes?

-How do types of
persons defined
by ethnicity
(uniquely or
among other
attributes) differ
as regards
attributes of
interest (life
chances,
resource s,
vocabularies of
motive, etc)?

3.2.3 Relevance of ethnicity: a blind spot of the survey interview
Surveys that include ethnicity as a variable and then report it in correlation with other attributes
unavoidably assert the relevance of ethnicity for the socia l processes that have led to the reported
attributes. This affirmation of the relevance of ethnicity represents a move in the very social process
of ethnic differentiation. Survey analysts are therefore not detached observers – but active players
in the so cial processes of ethnicization. In an interactional perspective, the most important attribute
of any social classification is whether it is used or not in actual interactions. In a survey -based
research there is no practical way to observe such an occasio ned relevance. It is then the analyst’s
decision whether to report education, religiosity, voting as being related to ethnicity. It is important

31
to stress here that the issue is not whether voting behavior actually correlates with ethnicity. The
issue is w hether there are empirical reasons to believe that the specific voting behaviors that are
described, in the aggregate, by survey data actually were made in light of first -order ethnic
classifications. Any classification may be brought in the foreground, pu shed in the background or
ignored in a process of social interaction, and this is an empirically observable process. Reporting a
correlation when there is no corresponding social process in which the two attributes were linked
represents, in the parlance o f causal analysis, a “spurious correlation” – an association that does not
reflect patterns of meaningful interaction.
3.2.4 Ethnic typologies of the first and the second order
Ethnicity is a process of first -order classification: occasionally people orient to e thnic distinctions, act
upon them and / or comment upon them. As regards the second -order classifications constructed by
analysts to reflect ethnic identities or processes, they are closely linked with first -order distinctions
in the large body of research that uses categorical measurements of ethnicity. This close coupling
implies a process of selection: for example, researchers have to decide what particular labels to
include and to exclude from the questionnaire. At the same time, these labels are going to be a
subset of the labels people may use in daily life.
Although second -order typologies of ethnic identifications have been built by researchers that have
proposed more elaborated, dimensional measurements of ethnicity conceptualized as identity, as
discussed below, I have not seen instances of such work of re -conceptualization in research on Roma
/ Gypsy ethnicity.
3.2.5 A research heuristic : ethnicity as a classification based on astrological signs
In attempting to understand what difference does it make, in survey -based research of ethnicity, if
we work under a subjective identity theory or an interactional theory of ethnicity, I found the
following analogy to be useful: let us assume that, over the centuries, astrological signs have
become common elements of conversation, allowing for self -presentation and attribution of
personal traits in all sorts of settings, including Censuses, employment interviews, speed dating and
so on. Let us further assume that, as analysts, we do not believe that celestial bodie s and their
configurations have any kind of influence on personality, behavior, or fate. That is, let us assume
that we believe that astrology lack s any descriptive relevance or explanatory power whatsoever. We
find ourselves in a situation in which a soci al classification is widely used in interactions, supporting
accounts of compatible and incompatible personalities, proper and improper course of actions – and
which is real in its consequences. The analogy is not farfetched if we think about biological th eories
of race, instead of ethnicity. The question is, then: what sort of survey -based research, and
interpretation of evidence, is appropriate to investigate this phenomenon? This counterfactual
scenario highlights the problematic use of several research operations:
a) The salience of astrological signs in social interaction would be of central interest: in which
situations do participants make use of astrology to orient their actions and those of their
partners’? When is astrology not invoked?
b) The issue of measurement error in measuring individuals’ astrological signs by self -affiliation
would become irrelevant. In those cases in which individuals would diverge in their self –
affiliation from the hetero -attributed sign estimated by the interviewer, on the ba sis of their

32
date of birth and other observable marks, this divergence would be of primary analytical
interest, and it would not be considered error or conceptual fuzziness.
c) In this scenario, an analyst that would correlate astrological signs with attribu tes that are not
socially seen as relevant for astrology would find little justification for her research question
and resulting coefficients. Also, typologies involving astrological signs would be of interest
only insofar they are used in regular interact ion to make sense of events and to guide
behavior. Analytical relevance of types would be directly related to their social relevance.
Non -related typologies would be considered analytical artifacts, without adding any insight
in the social working of astro logical signs.
d) The rhetoric of data analysis would carefully avoid causally -loaded terms (such as “effect”,
“influence” or “explained variance”) – in order not to inadvertently support theories of the
causal efficacy of astrological signs.
This analogy ill ustrates some of the main points of controversy between individual trait approaches
and interactional approaches. Key issues concern the treatment of anomalies and absences, the
attention to the minute details of actually occurring events, and an explanato ry versus a descriptive
interpretation of empirical associations .
3.3 Previous research contributions concerning survey -based sociological
research on ethnicity
I will present several research choices in survey investigations of ethnic topics, in particular re lated
to Roma nian Roma / Țigani, discussing alternatively available solutions. My argument is that the
technologies of survey design and statistical data analysis more often than not support a specific
theoretic conceptualization of ethnicity – that is, ethnicity as a trait and a quasi -cause of behavior,
functioning as a conduit of cultural difference in individual actions. In this understanding, ethnicity is
a manifestation of structure and culture at work. I discuss how survey -based investigations may
support a view o f ethnicity as participants’ sense -making resource for interaction (including
troublesome interaction) – and the implications of such a theoretical standpoint for research design,
data elicitation, analysis, and reporting.
3.3.1 Measurement of Roma ethnicity : hetero- and self-identification
In virtually all survey research on Roma / Țigani, ethnicity is measured with categorical variables. This
practice has opened two debates, which are shared in other social contexts of categorical ethnic
identification, in rese arch and administrative settings as well: the issue of hetero – versus self –
identification, and the issue of the most adequate categorical labels. Both of them have been
scholarly concerns but also, and even primarily, common reason concerns of Roma and non -Roma
people in all ways of life, especially participants in the larger public arena: state administration,
politicians, non -governmental associations, media, other public figures, including academics, and so
on.
The topic of hetero – and self-identificatio n has been, in my experience and in my retrospective
evaluation, one of the most engaging academic mysteries. More than ten years after the publication
of the first survey -based research on Romanian Roma (E. Zamfir and Zamfir 1993) , it was still
drawing attention and invit ing contributions and novel concerns (see for example the inquiry of
Covrig, 2004 concerning the degree of deliberation in refusing to declare one’s identity) or practical

33
solutions (such as Durnescu, Lazar, and Shaw 2002) . Before outlining my research contributions on
this topic and their sign ificance in the larger inquiry context which I have oriented to in this thesis, I
will highlight the specific tensions of this problem, as they appear in survey research on Roma /
Țigani issues.
The overarching frame of the methodological discussion concer ning the use of hetero – and self –
identification of ethnic categories consists in contrasting the two, and interpreting the resulting
differences as forms of error. The discussion is, overall, one of self -identification versus hetero –
identification – both o f them being conceived as alternative forms of pinpointing the same
phenomenon, namely the individual trait of ethnicity. The debate of hetero – versus self –
identification of Roma ethnicity is a telling illustration of the functioning of the error -trait dev ice.
Empirical differences between the results of ‘self -attribution’ and ‘hetero -attribution’, in different
interactional settings, are clearly observable and significant. By treating them as ‘errors’ the
discussion can then proceed in technical terms, eva luating what procedure or combination of
procedures is better suited for capturing the trait – instead of inquiring into the actual processes
that lead to these different outcomes , and asking what sort of interaction is more similar to the ones
that are of substantive research interest . It is taken for granted t hat the se various types of
interactional outcomes represent measure s of the same phenomenon, that is individual ethnicity,
which is an “enduring trait” of the respondent (Viswanathan 2010, in the quote discussed above) .
From the point of view of the survey resea rcher, addressing this error issue is complicated and
challenging, and a solution appears after juggling with theoretical, technical, common -sense and also
ethical considerations:
a) Under the most frequently espoused theoretical perspective on ethnicity as i dentity, the
actor is the one who has privileged access to her own ethnic identifications ; self-affiliation is
then the method of choice (Stephan and Stephan 2000; Rughiniș 2010) ;
b) At the same time, inter -ethnic encounters may involve conflicting claims and attributions of
ethnicity, and the definition of the more powerful participant in interaction (such a s a doctor
versus a patient, or a state official versus a claimant) may orient more the result than the
subjective actor’s identity; perceived ethnicity (hetero – attributed) is thus a relevant variable
in understanding distribution of wealth, health, educa tion and other resources (Raț 2005) ;
c) As a matter of empirical observation, a vowals of ethnicity are contextually dependent . This
variability is translated into a question of uncertainty in measurement (Mateos, Singleton,
and Longley 2009) . Some researchers take it to reflect the fuzziness and complexity of the
construct itself, especially in social contexts in which ethnicity has low salience (T. W. Smith
2008) , while usually researchers frame it as a problem of measurement error that can be
reduced, at least to some extent, by improvement on the adequacy of categories.
d) Avowals of ethnicity are also strategically oriented. G iven the serious consequences
attached to ethnicity or nationality, when stated in official contexts, people may not choose
an affiliation which i s considered stigmatizing or otherwise risky, such as the Roma / Țigan.
This strategy of ethnic affiliation is translated into measurement bias , which is then treated
as an obstacle to be overcome. The Roma / Țigan ethnicity is thus considered under –
repres ented in Census and , possibly , in surveys – and the resulting technical orientation is
that the techniques that produces more Roma respondents, within a general method (such

34
as self -affiliation), is the better technique. The associated theoretical commitme nt is that
whoever may, in some contexts, declare themselves a Roma without joking, irony, or some
other form of (intentional) lying, is to be considered as a member of the Roma population.
e) From a ethical point of view, given the essential uncertainty of an observer as regards the
subjective self -definition of a person, in terms of ethnicity, religiosity, sexual orientation etc.,
hetero -identification represents an imposition; as Simon (2011) discusses in his review
article, the United Nations Recommendations for Census collection of ethnic data insist tha t
individuals should be free to declare or not their ethnicity, while the Framework Convention
for the Protection of National Minorities3, while acknowledging membership in an ethnic
minority as an “objective ” reality4, also stipulates that “every person b elonging to a national
minority shall have the right freely to choose to be treated nor not to be treated as such”
(Article 3);
f) A second ethical dilemma in ethnic identification concerns the use of racialized, pejorative
categories in measurement. While re searchers and a considerable proportion of the
population may consider a specific appellative to be offensive or d egrading, others may still
identify with it – polemically, politically , or just simply so (Aspinall 200 9). This is also the case
for the ethnic appellative “Țigani”, which has been at the same time heavily contested, and
also used as a self -identification by some people.
g) Last but not least, the problem of hetero – versus self -identification features in publ ic
debates, in relation to concerns of finding the true number of Roma / Țigani, in relation to
complaints about the difficulties of gathering data about them, and in anecdotes of
communities where everybody is known to be a Țigan or Rom but, in the Census , there are
only a few people.
This multifaceted resonance of the problem of hetero – versus self-identification makes it an enticing
research puzzle. There have been two major types of research approaches in the literature on Roma
/ Țigani issues, which I have analyzed in Rughiniș (2010) . One direction has been to study empirically
the phenomenon of self – and hetero -identif ication of Roma / Ț igani people , in Census and in survey
settings . The second approach has been to take stock of difficulties as presented in the literature,
frame them as a methodological challenge of reducing measurement error and bias, and propose a
solution.
As regards Census self -affiliation, in Rughiniș (2010) I have reviewed a substantial amount of case
study and survey evidence to discuss this issue, which I have framed as a “reluctance error”. There is
considerable evidence, especially from qualitative studies, of communities in which Roma i dentity is
occasionally accepted, but whose members have chosen to self -identify as non -Roma in recent
Censuses based on self -identification. In Romania, where Censuses were conducted in 1992 and
2002, mentions of such communities can be found, for example , in Briciu (2007) , Grigoraș (2007) ,
Preda (1993) , Șerban (1998) , Voiculescu (2002) , Salat and Veres (2009) . In some cases, a majority of

3 Available on July 23, 2009 on the Council of Europe site:
http://conventions.coe.int/Treaty/EN/Treaties/Html/157.htm
4 The Explanatory Report for the Convention adds (Para. 35) that “This paragraph does not imply a right for an
individual to choose arbitra rily to belong to any national minority. The individual’s subjective choice is
inseparably linked to objective criteria re levant to the person’s identity. Available on July 23, 2009 on the
Council of Europe site: http://conventions.coe.int/Treaty/EN/Reports/Html/157.htm

35
the community members declared Roma identity in the Census, while in other cases only few of
them did. Quantitative information from the UNDP 2002 dataset indicate that, in Romania, 80% of
those who declare Roma ethnicity in the survey have also declared it in the 1992 Census5. The same
dataset indicates considerable variability at national level, since rates of self -identified Roma range
from 35% in the Czech Republic to 80% in Romania and Bulgaria.
As regards ethnic affiliation in survey settings, Ladányi and Szelényi (2001) , Csepeli and Simon
(2004) , and Ahmed, Feliciano, and Emigh (2006) have analyzed diversity in hetero and auto –
attribution, highlighting the socially contingent processes of ethnic labeling. Ladányi and Szel ényi
explore d self-identification with Roma ethnic labels and hetero -identification by interviewer s, using
an “ethnic oversample” , consisting of respondents who were hetero -classified as Roma by first -stage
interviewers while conducting various samples at national levels. Authors have argued that hetero –
identification is not a solution but an additional source of variability : from those people who were
hetero -identified as Roma by the first -stage interviewer, only 71.7% were also hetero -identified as
Roma b y the second -stage interviewer (p. 86) . The same data indicate little consensus between self –
identification and interviewer hetero -identification: from the 368 respondents selected as Roma by
survey interviewers, only 30.7% self -identified as Roma in the s ubsequent survey.
Hetero -identification by experts raises similar issues . The debate on hetero -identification has also
covered the legitimacy of the use of hetero -attribution by Roma observers or local experts (Prieto –
Flores 2009; Babusik 2004) , and hetero -attribution at community level (Sandu 2005) . Ladányi and
Szelényi (2001) point out that, in comparison to interviewers, experts have detailed knowledge of
“social problem” cases, and therefore if one were to select a sample based on their indications, it is
likely that it would be biased towards the poorer households ; experts may also have more detailed
knowl edge of ancestry, but it is unclear what their criteria are for categorizing people with ethnic
labels based on information about the past.
In the Inclusion 2007 survey (Gabor Fleck and Rughiniș 2008; Rughiniș 2010) operators interviewed
self-identified Roma heads of household but also an additional, randomly selected member of the
household. While in some case the randomly selected member was the same with the head of t he
household, there are 548 random respondents which were different members of the household. 53
out of them, around 10%, self -identified with non -Roma ethnicities (Romanian, Hungarian, other).
Since they lived in a household which had a member who self -identified as Roma, they belong to a
population with a high probability of Roma self -identification – but they do not actually self -identify
as Roma. Still, interviewers hetero -attributed Roma ethnicity to 32 out of the 53 persons, amounting
to a proportion of 60%. Therefore, as we have seen before, interviewers may use contextual
information to override individual self -identification .
There is a range of diverse approaches meant to link ethnic self -identification to hetero –
identification as approximations of the same individual trait – either by designing the interview
interaction, or by conceptual operations:
– In the Inclusion 2007 survey (Gabor Fleck and Rughiniș 2008) we have asked for a primary
and a secondary affiliation, attempting to see what proportion of respondents would declare

5 Unfortunately, since the UNDP survey was conducted in Romania in December 2001, respondents’ answers
refer to the Census in 1992 and not to the 2002 Census, and there fore may be influenced by information
retrieval errors. Information about the UNDP survey period are available in the online rapport annexes at URL:
http://roma.undp.sk/

36
a Romanian ethnicity and also offer a secondary Roma identity. I n the comparative sample,
out of 956 respondents, 4.7% declared a primary Roma ethnicity while only an additional
0.4% (4 respondents) declared a s econdary Roma ethnicity.
– The UNDP survey, assuming a “reluctance error ” in survey self -identification, proposed an
“implicit endorsement identification” strategy. The interviewer approached the potential
respondents, hetero -identified as Roma, with the op ening question “Good morning/day, we
are conducting a survey among the Roma population. Would you mind being interviewed?”
Explicit rejection led to the cancellation of the interview, but acceptance was interpreted to
mean that the respondent is Roma (UNDP – United Nations Development Program 2003) .
This meth od requires a leap of faith, since it is not clear whether acceptance of the interview
really indicates an implicit acceptance of Roma ethnicity, or alternative reasons – such as
politeness or misunderstandings. This dataset also indicates variability betw een national
contexts. The proportion of selected respondents who explicitly rejected the Roma identity
was 5% in Romania but 14% in Bulgaria.
– Durnescu, Lazar, and Shaw (2002) adopt a sequential approach in which they first ask for
various information on the cultural background of the respondents, asking for explicit self –
affiliation only at the end of the qu estionnaire;
– Székeli, Csepeli, and Örkény (2003) develop a c lassification of Roma ethnicity by combining
respondents’ ethnic affiliation with information about their ethnic background ;
– In a community -level survey, Sandu (2005) refers to a population “which probably self –
identifies as Roma” (p. 42), thus interpreting uncertainty as an epistemological attribute of
the investigation, rather than a featur e of the phenomenon of ethnicity .
In my methodological research on Roma -related surveys (Rughiniș 2011a; Rughiniș 2010) I have
analyzed this central topic and its theoretical implications. From a theoretical perspective of
ethnicity as identification with a community, contextual instability in declarations is easily
interpreted as error . The problem becomes one of finding a way of access to true individual
affiliations – which are subjective, but not arbitrary, and, therefore , are intersubjectively available to
people with whom they interact on a daily basis, even if not for the researcher. In theory, access may
involve establishing the trustworthiness of the researcher, deflecting reluctance to self -identify as
Roma, or finding what closely involved partners of interaction know about the respondent’s ethnic
identity, assuming a certain degree of final uncertainty. Hetero -affiliation by socially distant
observers, such as interviewers, remains to be used as indicator of another phenomenon, namely
“perceived ethnicity”.
Alternatively, if ethnicity is understood as an outcome of interaction, the pro blem of measurement
error disappears. The question then becomes, what particular ethnicity is attributed by all
participants based on their interaction, what sort of interaction what that, and what can we learn
from this outcome that is of relevance for un derstanding other interactions. An apparently Roma
person who declares a non -Roma ethnic identity is not, in this approach, a potential source of
measurement error. Her specific answer to a specific question, asked by a specific person in a given
situation , is to be understood as the very manifestation of ethnicity, the concrete instantiation of the
operation of ethnicity in action.

37
From this perspective hetero -identification cannot correct or adjust information from self –
identification; both of them , and their occasional differences, are relevant i f they occur in
interpersonal interaction, in order to understand the dynamics of the mutual classification process.
A secondary implication is that ethnicity that is self -reported under conditions of perfect an onymity,
such as in postal surveys, is of less analytical value than ethnicity that is self -reported under
conditions of face -to-face interaction, provided that we have a model of that interaction. If we are
interested in ethnicity as it happens in daily c onduct, then we are interested in its socially
occasioned expression; self -presentations in postal surveys have little to tell by way of similarity with
other, real life situations of self -presentation.
From an interactionally oriented theoretical perspect ive, a number of over one thousand personal
encounters in which ethnic information is solicited and offered represent a potentially significant
body of evidence. Still, the researcher would not be interested only in the final check on the box /
boxes, but on the process in which this mark has been achieved – including steps such as: reaching
the house of the respondent, negotiating entry, making a first impression as to the probable
ethnicity of the respondent and the other household members that are presen t, asking the question
about ethnicity, clarifying it, reacting to the answers, recording the answers. A detailed body of
evidence on first inter -ethnic encounters in survey situations would offer potentially illuminating
insights on the social efficacy of ethnicity as a classification available to participants in interaction, to
be used for their own purposes.
3.3.2 Model specification and data analysis
Survey -based research on ethnicity and other forms of quantitative data that include ethnicity
indicators affo rd estimates of average differences between categories of people identified by
particular ethnic labels. What is the theoretical relevance of such differences?
In my research (Rughiniș 2011a ) I have differentiated between three different research models
concerning ethnicity. Firstly, ‘ethnic disadvantage models’ measure a given inequality in average
access to resources or in risk incidence. Secondly, ‘discrimination models’ attempt to isolat e the
outcomes of ethnic discrimination from other sources of inequality , and to measure it. Thirdly,
‘ethnic difference’ models pursue the relationships between ethnicity and other social phenomena
which are theoretically linked with the process es of ethn ic differentiation. The three types are not
mutually exclusive, because an ethnic disadvantage or a discrimination model may at the same time
investigate a phenomenon which is relevant for ethnic differentiation. Also, what counts as a
resource or a risk i s a matter of normative choice: any feature assumed to be desirable may be
analyzed from an ‘ethnic gap’ perspective – for example, wealth, education, health, but also
consumption of a specific product, religious belief or participation, or adherence to a given tradition.
– The main focus of ethnic disadvantage and discrimination models is precise measurement,
and a secondary focus may consist in the identification of other relevant predictors besides
ethnicity . Control variables usually include socio -demogra phic features. While most ethnic
disadvantage models do not control for subjective variables, such as values or attitudes,
there is a growing body of research that connects ethnicity with Theory of Planned Behavior
(TPB) models in diverse topics (Blanchard and et. al. 2003; Bryan, Ruiz, and O’Neill 2003)
under the assumption that ethnicity explains shared beliefs, attitudes, or the perceived

38
norms of significant others, which are developed by socialization in social networks and
communities which have a specific ethnic profile.
– Discrimination models using cross -sectional survey data confront a specific array of
challenges . All relevant variables re lated to respondents’ competence s and respondents’
preference s relevant for the outcome under investigation (such as income, employment ,
access to social services ) must be controlled for, in order to identify unequal treatment.
Since discrimination involve s unequal treatment for persons who have the same relevant
qualifications, the normative issue raises, what counts as a relevant qualification and what
counts as an irrelevant, thus discriminatory, criterion. Therefore, as discussed in detail by
Hanquinet and et.al. (2006, pp. 51 -52), the use of multivariate analysis to measure
discrimination is vulnerable to several sources of bias, including the ‘omitted variable bias’,
when releva nt controls are not included, the ‘included variable problem’, when some
controls already capture variation due to discrimination, and the ‘diverting variable bias’,
when controls include variables that should not be controlled, according to the normative
concept of discrimination employed.
– Ethnic difference models range from simple models that trace differences between ethnic
groups in language use or religious denomination, to complex models that explore
influences of ethnic affiliation on intergroup att itudes, parenting styles, religious
participation, conflict management, and so on. While in ethnic gap models ethnicity is
usually measured as a categorical variable, used to chart unequal distributions of resources
or risks across categories , in ethnic di fference models the focus is on processes of ethnic
delineation , also by including more complex measurements of individual ethnic identity .
Estimates of average differences between ethnic categories represent aggregate representations of
quantified social phenomena. The theoretical relevance of such aggregates depends on:
– The criteria used for aggregation: whether it is an individual attribute or a set of attributes;
– The interpretation of the link between the differentiating criteria and the observed
differ ence.
For example, regression models and more complex path or structural equation models afford
estimating and reporting “direct effects” for all variables of interest, including ethnicity. Leaving
aside the causal innuendo of the concept “effect”, what so rt of representation does this estimate
offer? Direct effects of individual variables, such as ethnicity, represent an aggregation of differences
across types of actors and situations of interaction, isolating the respective trait as a focus of
analysis. From an interactional perspective such heavy aggregation makes little sense, as it bears
virtually no link to any concrete situation of social interchange in which empirical instances of
differentiated outcomes occur. If we estimate the direct effect of Rom a ethnicity on school
education, controlling for gender, generation, and type of residence, what type of life trajectory is
this estimate referring to? What types of schooling (types of schools, types of parental support,
types of collegial relationships w ithin the classroom) are portrayed? What seems to be an
analytically purified estimate of the influence (or relevance) of ethnicity becomes, in an interactional
perspective, an analytical method of conflating social processes to the point of unintelligibil ity, or, in
explanatory parlance, an instance of amalgamated models (Rughiniș 2007b, p. 221) .

39
If aggregated quantified representations are to have any relevance for an interactionally -oriented
research, they should be as closely related to a specific type of situation of interaction, a type that is
distinctive and intelligible enough to warrant the assumption that it generates a typical course of
interaction, which leads to the observed difference. To put it simply, ethnic differences constructed
by quantitative aggregation should be interpre table as results of typical interactions in typical
situations (involving typical participants). This links survey -based research directly to the
epistemological problem of typification in social interaction and sociological research , and the
criteria for the constructions of adequate typologies , following the line of inquiry opened by Schütz
(1953) in his discussion of “Common sense and scientific interpretation of human action”. Given the
centrality and ubiquity of typification in common sense -making, the activi ty of sociological
typification is not to be taken lightly: it is a core process of knowledge production, at once relying on
common -sense methods and constructs, while attempting to introduce more clarity and
understanding power (Kim and Berard 2009 , p. 285 ):
Typification is simultaneously a scientific method, a commonsense metho d of perception and
communication, a topic for the social sciences, and a theoretical insight which has since the
sixties become a tried and tested heuristic for a tremendous variety of empirical studies
across a variety of disciplines and literatures.

40
4 Theorizing public knowledge of science in surveys of scientific
literacy

4.1 Theoretical views on scientific literacy

In order to illustrate the specificity of survey -based sociological research on scientific literacy, it is
useful to briefly review the theoret ical development of the concept. Although surveys describe the
general public, and thus speak about public knowledge of science, the concept of ‘scientific literacy’
has originated in education research, being particularly relevant for studies of science t eaching and
learning in schools. Laugksch (2000) discusses the long tradition of the concept of scientific literacy ,
having had a rich development in education sciences. However, it was after Miller's (1983) influential
article in Daedalus that the concept gradually established its position in the sociological inquiry.
Laugksch (2000) distinguishes between three “interest groups” within the scientific community that
use the concept of scientific literacy. Firstly, there is the “science education c ommunity”; then, there
are the “social scientists and public opinion researchers concerned with science and technology
policy issues” (p.75). Apart from these, sociologists of science and science educators, who address
scientific literacy sociologically, a re interested in “how individuals in everyday life interpret and
negotiate scientific knowledge” (idem). These three interest groups point to rather distinctive
methods and audiences: science education studies focus especially on students and teachers, whi le
the other two fields include various segments of the wider public. The Public Understanding of
Science (PUS) field , to which I have referred in my research, relies extensively on surveys and media
analyses with a quantitative methodology, while on the o ther hand there is a considerable body of
in-depth, qualitative scrutiny of the ways in which science interweaves with people’s everyday life.
In what follows I focus on the Public Understanding of Science approach , discussing the theory and
measurement p ractice of the National Science Foundation (NSF) Scientific Literacy scale and its
controversies with reference with the critical issues that I have outlined before: the error -trait
device, the error -practicality device, treatment of scientific literacy as a quasi -cause, its relevance
and its associated typologies .
The Public Understanding of Science research program has relied substantially on the
conceptualization and measurement of “civic scientific literacy” advanced by Miller (1983) . In this
article the author proposed a concept of scientific literacy to b e used in social surveys of the general
public, relying on the distinction introduced by Shen, in 1975, between practical, civic, and cultural
scientific literacy ( apud Laugksch 2000 , p. 77 and Miller 1983 , p. 32). The concept refers to “a level of
understanding of scientific terms a nd constructs sufficient to read a daily newspaper or magazine
and to understand the essence of competing arguments on a given dispute or controversy” ( Miller
1998 , p. 204). More specifically, it represents a “minimal threshold level” of “understanding of
science and technology needed to function as citizens in a modern i ndustrial society” ( Miller 2007b ,
p. 2).
Initially, Miller defined civic scientific literacy using three dimensions: an understanding of the
“norms and methods of science”, “cognitive science knowledge”, and “attitudes towards organized
science” ( Miller 1983 , pp. 32 -34), which he subsequently refined as follows:

41
“(1) a vocabulary o f basic scientific constructs sufficient to read competing views in a
newspaper or magazine, (2) an understanding of the process or nature of scientific inquiry,
and (3) some level of understanding of the impact of science and technology on individuals
and on society” ( Miller 1998 ).
Most often, “scientific literacy” has been employed in quantitative research only with reference to
the first two theoretical dimensions, the latter receiving less attention (Miller 2006a) .
A first commentary on the conceptualization of scientific literacy is that its definition has a pragmatic
orientation: it is a type of knowledge that al lows people to “read competing views in a newspaper or
magazine” and to “function as citizens in a modern industrial society”, as indicated above. Therefore,
it is a concept particularly suited to be conceptualized as a resource. The question is, how can w e
determine what types of knowledge are useful when reading competing views, or when living in a
modern industrial society? And, moreover, are the two types of knowledge overlapping? This
question seems to invite an empirical answer – one that presumably a lso depends on the particular
society under study. Still, the survey -based conceptualization of scientific literacy has skipped these
questions and it has proposed a measurable concept that has no obvious link to either of the two
practical uses to which i t was theoretically assigned. Nor have there been empirical investigations to
discuss, a posteriori, whether the operationalization is adequate in light of its initial definition. The
measurable construct has been implicitly defined as a stable individual trait, with trans -situational
relevance, being also deemed comparable across countries.
In Table 4 I have summarized the terminology employed to refer to the public knowledge of science.
It is generally referred to as a type of “ knowledge”, a cognitive dimension, and, with the exception of
Miller himself, the “vocabulary” concept is not used. We can see here how the initially proposed
definition could accommodate an interactionally oriented operationaliz ation, if scientific litera cy
were defined as the “mastery of a vocabulary of scientific constructs that allows for meaningful
conversation on scientific topics in given context”. A standardized form of such a conversation could
then be included in a questionnaire. The measurement m odel for a vocabulary would then rather be
topic -specific, for example concerning antibiotics, or smoking, or astronomy etc., and would rather
be formative than reflective (see the comparison below in section 4.3.2.1 ), focusing on several
constructs of particular thematic interest .
Table 4. Terminology used in referring to the “knowledge of scientific constructs” dimension
Key concepts Source Term Page
Knowledge about
science (Nick Allum,
Sturgis, et al.
2008) “public knowledge about science
and technology”
“basic “textbook” knowledge
about science”
“knowledge about scientific facts and processes”
“knowledge about scienc e” 35-37
Knowledge
Facts (M. Bauer
2009) “knowledge of basic textbook facts of science”
“factual knowledge” 223
Knowledge
Facts (Godin and
Gingras
2000) “knowledge of S&T facts” 52
Knowledge
Concepts
Scientific knowledge
Knowledge about
science (Hay es and
Tariq 2000)
“scientific knowledge”
“knowledge and understanding of some basic scientific
concepts”
“knowledge of science” 433-434
“correct knowledge of scientific matters” 435
Knowledge
Constructs (Miller 1983)
“Cognitive science knowledge”
“Knowledge of basic scientific constructs” 34

42
Key concepts Source Term Page
Vocabulary
Constructs (Miller
2006b) “basic vocabulary of scientific terms and concepts”
“construct vocabulary” 2-3
Vocabulary
Constructs (Miller 1998) “vocabulary dimension of basic scientific constructs”
206
“construct vocabulary dimension” 209-210
Knowledge about
science
Facts (National
Science
Board 2008) “factual knowledge about science” Chapter 7
Knowledge
Scientific knowledge
Comprehension (Pardo and
Calvo 2004) “cognitive dimension of public perceptions of science”
“scientific knowledge of the ‘know -what’ type”
“appropriation of scientific theories of the world” 203-204
“comprehension of central concepts and propositions about
the natural world” 205

In my work I have focused on the first component in Miller’s definition, the “vocabulary of basic
scientific constructs”, whi ch has also included in its operationalization the construct of evolution , to
which I pay special attention . I refer to this vocabulary whenever I use the concept of “scientific
literacy”. I employ the concise expression “scientific literacy”, rather than “civic scientific literacy”,
while keeping in mind the concept’s definition as a citizen’s skill for participation in public debates
and, implicitly, in conversations , and the focus on the vocabulary dimension.
4.2 Theoretical bearing of survey -based methodol ogical discussions
4.2.1 Scientific literacy as a trait
Resulting from collaboration between Miller in the United States and Thomas and Durant in the
United Kingdom (Miller 1998b, p. 207; Nick Allum, Sturgis, et al. 2008, p. 38) , the scale used to
measure scientific literacy includes several items assessing knowledge of scientific constructs. The
initial version has been reduced for the use of the National Science F oundation (Bann and Schwerin
2004) . Eurobarometer s urveys have employed it as well, for example in the 63.1 / 2005
questionnaire (TNS OPINION & SOCIAL 2005a) as presented in Figure 1.

43
Figure 4. The scientific knowledge quiz included in the Eurobarometer 63.1 / 2005 questionnaire. Source: TNS OPINION &
SOCIAL (2005, p. 14)

In order to measure th e vocabulary dimension of civic scientific literacy, an important question had
to be addressed: which are the scientific constructs to be included in the measurement? This
process of selection is documented in several papers (Durant, Evans, and Thomas 1992; Miller 1998;
Miller 2007b) . Firstly, Miller (2007c) advances the central criterion of durability that distinguished
“basic constructs” considered foundational in understanding current issues, like atomic structure or
the DNA, from “specific terms, such as the fallou t of strontium 90 from atmospheric testing”. In an
account of the elaboration of the Oxford Scientific Knowledge Scale, Durant, Evans, and Thomas
(1992, p. 165) invoke difficulty levels and disciplinary fields as criteria for the selection of items in the
quiz6. Later, the authors defined the scale as measuring the understanding of ‘scientific product’,
through wh ich they meant:
“[T]he elementary theoretical and factual findings of science – for example that light travels
faster than sound, that diamonds are made of carbon, and that sunlight can cause skin
cancer. The 23 items used to measure this dimension were d rawn from a wide range of
natural and medical sciences (…) (Evans and Durant 1995, p. 58) .
In addition to these criteria, the scale had to pass analyses of internal cons istency and dimensionality
(Miller 2007c, p. 4; Durant, Evans, and Thomas 1992, p.179) . For instance, when elaborating the
shorter version of the NSF scale, Bann and Schwerin (2004, pp. 4 -5) examined the distribution of the
items across content areas, the dimensionality of the scale, as well as psychometric properties of

6 “In the domain of knowledge, our aim has been to assess levels of acquaintance with the factual and
theoretical content of science. After careful piloting, we established a suitable level of average
difficulty for items in this area (…). Our knowledge quiz comprises more than 20 simple propositions
cove ring the fields of physics, chemistry, geology, and the bio -medical sciences”.

44
individual items within an Item Response Theory model, focusing on item difficulty and item
discrimination.
To summarize, the constructs included in the Oxford and NSF scales to measure scientific literacy
had been selected from medical and natural sciences, such as to be fundamental in science, and
therefore durable, and to cover an empirically appropriate spectrum of difficulty for respondents.
Items have also been examined in their relation with the overall scale, by examining the Cronbach
Alpha reliability and the dimensionality of the scale, and Item Response Theory parameters.
The selection of fundamental, elementary, “textbook” scientific constructs has bee n accomplished
by scientists with reference to scientific criteria of what “elementary” means. Respondents’
experience with scientific constructs had only been introduced in the selection process through the
evaluation of item “difficulty”, that is a stati stical estimate based on the probabilities of a correct
answer. Because the selected scientific constructs are fundamental from scientists’ perspective, they
are considered to be similar topics of knowledge for lay people. Once introduced in the scale, the
difference between lay representations of various scientific constructs is measured, rather than
investigated. More precisely, it is statistical estimates such as Cronbach alpha reliability, factor
loadings, and IRT difficulty levels and discrimination po wers that are employed to express, evaluate
and control construct heterogeneity.
4.2.2 Relevance and typologies
The Public Understanding of Science research thread, in which I have positioned myself, is highly
reflexive and there has been a rich body of work in theoretical and methodological reviews (such as
Bauer, Allum, and Miller 2007; Bauer 2009; J. D. Miller 2007; Vlăsceanu 2011; N. Allum et al. 2008;
Pardo and Calvo 2004) . The field is also not free from longstanding controversy. Most of the debates
can be traced to issues of relevance and to c ontested typologies. The two are related, as the
argument goes: irrelevant tests of knowledge are said to assist in a classification of people into
worthy scientists and to -be-educated lay people. An illustration of this joint critique is formulated by
Fayard (1992) in his memorably titled article “ Let’s stop persecuting people who don’t think like
Galileo !”:
“Throughout history the tables have constantly been turned, with yesterday’s victims
becoming today’s persecutors. This is why, peering into t he cradle of Public Understanding of
Science, I would like to make the following plea (even if it’s only wishful thinking): let’s stop
persecuting people just because they don’t think like Galileo! We are told, for example, that
many people do not know tha t the Earth goes round the Sun. I confess that I myself have
never woken up in the morning saying ‘the movement of the Earth on its axis is such that the
Sun can be seen in the east’ – in my daily life the Sun moves round the Earth. (…) The question
is: how does a venture in the public communication of science and technology see its public?
As empty vessels to he filled, as warped minds in need of straightening out, as citizens with
whom to enter into dialogue, or as taxpayers to be convinced of the necessit y of funding
research?” (p. 15).
The “ignorant public” has been the target of critiques directed against the “public deficit model” of
inquiry , a style of research that measures public scientific literacy, thus allowing for a classification of
publics int o more or less literate, and implicitly focusing on individual knowledge (of a declarative
kind) as indicator of civic competence. This focus also directs attention to the public, assigning
ignorance as an individual attribute. Following debates, this appr oach has been replaced by a
“science and society” orientation, focused on trust deficit, expert deficit, confidence crisis, and

45
elaborated notions of science’s publics (Bauer, Allum, and Miller 2007, p. 80) . Nevertheless, despite
significant discussion focused precisely on what was understood to be an impoverished
representation of people in their relation with science, the methodolo gical operations involved in
measuring and interpreting scientific literacy in the general public are still (unwittingly) reproducing
the same deficit orientation – insofar that knowledge of scientific constructs is located within
individuals . People are t hus classified as being more or less knowledgeable, in reference to a body of
knowledge that is authoritative and external to them.
The theoretical line of alternative conceptualization of public knowledge of scientific constructs
locates it interpersonall y and institutionally, in the multiple social settings in which people actually
conduct their lives. Scientific literacy is thus conceptualized as “collective praxis” (Roth and Lee
2002) . Ignorance of scientific concepts is no longer construed as an attribute of individuals, but as a
possible outcome of various types of engagement and disengagement with scientific institutions.
Ignorance of radioactivity -related constructs in the setting of a nuclear fuels reprocessing pl ant is
seen as the outcome of the socially and organizationally required trust in the expertise of
organizational scientists; therefore, ignorant people are shown to be competent social actors,
effectively attending to their contextually legitimate concern s:
Having arranged several discussion groups with apprentice fitters, plumbers, electricians and
others, and having prepared a series of questions exploring their understanding of, for
example, the different properties of alpha, beta and gamma radiation an d the different
protection measures they required, we were dumbfounded to find a version of the same
passivity we had found elsewhere. Not only this, but the workers defended their ignorance
vigorously. We eventually realized from their explanations that t hey were intuitively the
competent sociologists, and we had been operating with very insensitive assumptions. In
effect they were saying that as employees in a large and hazardous industrial complex they
had to engage in disciplined work procedures and det ailed operational rules which should
have the best scientific understanding of radiation hazards built into them. Scientists and
engineers in the firm and its surrounding regulatory bodies had designed these rules, and
workers had to trust that they had do ne so competently, just as we have to trust that the
local garage bas serviced our car brakes properly (Wynne 1992 , p. 39 ).
Ignorance of something is therefore investigated as an activity – which may involve orientation to
other structures of relevance, and even concrete and strategic actions of ignoring, obfuscating,
refuting and rejecting knowle dge – as illustrated by Desantis (2003) in his account of ignorance of
the negative health effects of cigar smoking, or by Auyero and Swistun (2008) in their research on
residents’ uncertainty about the reality, causes and effects of local pollution.
4.3 Previous r esearch contributions concerning survey -based sociological
research on scientific literacy
4.3.1 The error -trait device: s cientific literacy as vocabulary , worldview , or trait
The continuous advance and consolidation of the Public Understanding of Science research program
during the last three decades, and the recent public controversies on the use of evolution as an
indicator of scientific literacy (Bhattacharjee 2010) allow us to examine, in retrospect, the implicit
theoretical assumptions that underlie the substantial body of research employing the NSF scientific
literacy scale.

46
Pardo and Calvo (2004 , p. 204 ) identify as the main, foundational presupposition of the PUS research
program the idea that lay resistance to scientific rationality is mainly due to prejudice, reinforced by
traditional constructs, and that this resistance gradually fades away as a result of dissemination of
knowledge by schooling and popularization programs. This idea has been translated into the
hypothesis that scientific literacy , as measured by the Oxford or NSF scales, is positively associated
with attitudes towards science – a hypothesis that has been extensively evaluated (see for example
M. Bauer 2009; Nick Allum, Sturgis, et al. 2008) .
Shifting the viewpoint of assumption analysis, we can identify a similar framing within the process of
operationalization and measurement of scient ific literacy: when it comes to science, to know her is
to love her – or, at least, to believe in her. More specifically, the NSF and Oxford scientific literacy
scales rely on four presuppositions. Firstly, it is assumed that all fundamental scientific con structs are
similar objects of knowledge for lay people . Secondly, there is the idea that a reasonably attainable
level of lay understanding of scientific constructs leads people into believing in the scientific facts
with which these constructs are mutual ly constitutive . That is, familiarity with the vocabulary of
science develops concomitantly, and is intrinsically linked, with a personal appropriation of the
scientific knowledge about the world. The third assumption is that each adult individual has a
relatively stable and trans -situational disposition to understand scientific constructs to a certain
degree . In other words, some people have a higher disposition towards correct understanding, and
other people have a lower disposition. Finally, the fourth p resupposition is that this disposition is a
form of knowledge or ability . The operationalized concept of scientific literacy represents exactly this
ability of an individual, which assumingly underlies one’s beliefs in all textbook scientific facts.
In fact, all these four assumptions represent arguments that can be evaluated empirically: (1) that
there are similar ways in which lay people get acquainted and develop their knowledge of various
textbook scientific constructs; (2) that understanding of scie ntific constructs is strongly conducive to
believing in their existence; (3) that each individual has a stable, trans -situational disposition to
understand (and thus, according to the second assumption, to believe in) elementary scientific
constructs, and (4) that this disposition is a form of ability or knowledge, which may adequately be
termed “literacy”.
At a yet lower level of methodological analysis, there are other assumptions which have remained
under -evaluated , and which I also do not engage . For e xample, incorrect answers to quiz items are
routinely aggregated with “don’t know” answers, although the two represent different respondent
behavior and interaction outcomes (M. Bauer 1996 , p. 43 ). The resulting dichotomous quiz variabl es
are handled on the presupposition that the unobserved variable of scientific literacy is best modeled
as a continuous numerical variable (a dimension), rather than a discrete numerical, an ordered , or a
categorical one.
An examination of the evolution i tem in the NSF scale is useful for understanding the theoretical
relevance and empirical value of the four conceptual presuppositions outlined above. For example,
this inquiry points out that familiarity with scientific concepts often leads to the acceptan ce of
related scientific facts, but sometimes it does not. It is especially the case when those scientific facts
conflict with matters of importance for people’s lives.
Common sense or vernacular knowledge (W. Wagner 2007) of scientific facts is constituted in
diverse life situations, and is put to use in conversations an d other actions that have highly variable
meaning and importance. This practical significance explains the radical discontinuity between

47
scientific knowledge and common -sense knowledge (W. Wagner 2007; Schütz 1953) , and the
divergent lay representations of scientific constructs too. For example, the lay understanding of
those “genes” that deter mines the sex of a baby may not be the same as for the “genes” from a
genetic disorder, or for the “genes” from a genetically modified tomato – and it may only have a
slight family resemblance with the scientific understanding of the “gene” construct. Therefore,
people’s experiences with fundamental scientific constructs are heterogeneous: some constructs
matter more, and in different ways, than others do. In order to account for people’s representations
of a scientific construct it is essential to observe the institutional settings in which they learned of it
(Vlasceanu 2011 , p. 559) and otherwise dealt with it. Furthermore, belief in contested scientific
facts may derive not only from knowledge of science, but als o from other individual dispositions,
such as trust in the cognitive power of science, or appreciation of the effects of science.
Therefore, the first presupposition outline above holds only with reference to a subset of the
scientific constructs included in the scale . This subset consists of those constructs that are of no
practical consequence for the daily decisions of adult respondents, being limited to encounters in
contexts of formalized education, such as natural sciences classes, or through other me dia, possibly
documentaries, museums, and other channels of science popularization. I have refer red to these
constructs as quiet , contrast ing them with the animated constructs, which people encounter,
confront with and rally to in important life situations .
Of course, it is a matter of empirical analysis to identify in a given public the quiet scientific
constructs that are acquired in school -like settings, and to identify situations in which such settings
occur after graduation. Still, it seems that many constructs in the NSF scale (such as the center of the
Earth, oxygen, lasers, electrons or continents) match this acquisition profile for diverse publics, thus
accounting for the overall empirical value of the instrument according to survey -based
methodolo gical criteria .
At the same time, any animated construct that is experienced by adult people in their consequential
everyday life situations has the potential to expose the three assumptions. Firstly, a clear lay
understanding of the construct may not be strongly conducive to a belief in all scientific facts
associated with it. Then, a belief in the construct may not reflect a stable, individual disposition
shared with other scientific constructs; that is, individuals come to know it and believe in it in
situationally specific ways. Lastly, if there is a stable disposition that influences individual belief in
this construct, it is not necessarily knowledge of science, but rather appreciation of science, or trust
in science. Through these animated items, wh ich have a different social life than their quieter
counterparts, it is possible to trace back and better grasp some of the occasional problems with
scale items and its overall functioning.
4.3.2 Understanding t rouble in the quantification of scientific literac y: three possible
distinctions
There are several points of trouble in the literature dedicated to the Public Understanding of Science
research thread: a) the internal coherence of the scale is not very high; b) some constructs in
particular are troublesome , especially evolution – which, in the United States, has a weak correlation
with the general factor extracted from the scale (Miller 2007b , p. 5 ) while, in addition, its inclusion is
also publicly contested (Bhattacharjee 2010) ; c) there is some discussion of the self -assumed
ignorance (“Don’t know” answers) concerning scientific constructs, but this was not followed by an
analysis of these responses , with the exception of M. Bauer (1996) .

48
My analysis of these debates indicates that they can be understood by introducing two main
distinctions:
a) Between reflective and formative measurement models: scientific literacy is usually considered to
be measured under a reflective model; still, a formative model may be more adequate if there is a
particular theoretical interest in specific constructs that are weakly correlated;
b) B etween quiet and animated scientific constructs: a reflective model may be used only with
constructs that have similar learning and use contexts; this is not the case with the animated
constructs, such as evolution, which are embedded in heated interactional situations.
4.3.2.1 Distinguis hing measurement models: reflective and formative
The theoretical affinity of measurement models
A measurement model presents the relationships between an operationalized concept and the
observed values of measured indicators (Billiet 2010) . For example, the measurement model allows
us to d erive values for the unobserved operationalized concept, also called the latent variable, from
the observed values of a questionnaire quiz. At the same time, the operationalized concept is not
valuable in itself, but only in relation to the theoretical con cept that it claims to represent (Saris and
Gallhofer 2007 , pp. 15 -29).
In survey -based research of public knowledge of scientific constructs , scientific literacy is a
theoretical concept, defined as the capacity to understand c ompeting scientific arguments as
presented in newspapers and political debates. It has been operationalized as the respondents’
ability to answer correctly survey questions on fundamental scientific constructs; this individual
ability is not directly obser vable, and therefore it represents a latent variable. The quiz items in the
NSF scale are its indicators, and respondent answers to these items are directly observable.
Methodologically, reflective measurement models differ from formative ones in terms of the
relationship between the unobserved operationalized variable, and the observed indicators. In
reflective models, the observed indicators are considered to be effects of the latent operationalized
variable, while in formative models the relationship is somehow reversed: the operationalized
variable is understood as resulting from the observed indicators (Bollen and Lennox 1991) .
Interpretation of items is different by design between the two models. In a reflective model, the
correlations between the true val ues of the indicators are theoretically assumed to derive from their
shared causation by the latent variable. Unless additional, external variables exercise some other
shared influence on the indicators, the items are not expected to covary when the latent variable is
controlled for. Also, in a reflective model items are understood as essentially replaceable by other
similar items: the scale includes a sample of the theoretically expected effects of the latent variable.
Practically, indicators in a reflecti ve model are interpreted as alternative measures of the underlying
latent variable.
On the contrary, in a formative model the observed variables may, or may not correlate when the
latent variable is controlled for, since they are its components or its caus es, not its effects. Each and
every observed item is considered to be an essential and irreplaceable part of the model, and the
meaning of the latent variable changes if an item is added, removed or replaced. The overall
satisfaction with a product, measur ed in market research, is a typical example of a concept that may

49
be operationalized in a formative model, when defined as the result of multiple judgments of
specific product features. Another concept which is conveniently measured with a formative
operat ionalization is the socio -economic status (SES) defined as “a combination of education,
income, occupation and residence”, as exemplified by Diama ntopoulos and Winklhofer (2001 , p.
270): “[i]f any one of these measures increases, SAS would increase (even if the other indicators did
not change); conversely, if a person’s SES increases, this would not necessarily be accompanied by
an increase in all four measures” (idem). Formative unobserved constructs consist or derive from
configurations of heterogeneous and possibly uncorrelated attributes, which in turn may function as
indicators for the presence of the construct. With regard to the public engage ment with science, a
person’s adherence to a scientific worldview may be modeled as an outcome of her judgments
about a series of significant scientific constructs.
J. R. Edwards and Bagozzi (2000) and Borsboom, Mellenbergh, and van Heerden (2003) present
edifying discussions of the relationships between operationaliz ed variables and their indicators
under the two measurement models. The choice of a measurement model is theoretical, depending
on the correspondence assumed to exist between observed and unobserved variables of interest.
Generally, reflective models are much more frequent than formative ones. A possible reason may be
that it is easier to think of unobservable causes of observed effects, than of unobservable effects or
configurations of observed phenomena. There are also statistical reasons, such as the fa ct that the
formative model is always under -identified and needs to be inserted into a larger model in which the
operationalized variable acts as a predictor for a dependent variable. For example, one may explain
attitudes on genetically modified foods by adherence to a scientifically informed worldview about
life, which is measured formatively. The formative model parameters are then estimated
simultaneously with the causal influences of the latent worldview on the dependent attitude. This
dependence on a predicted variable raises theoretical questions about the meaning of the formative
latent variable. Heise points out that a formative variable “is the composite that best predicts the
dependent variable in the analysis .[…] Thus, the meaning of the latent construct is as much a
function of the dependent variable as it is a function of its indicators” (1972, p. 160, quoted in J. R.
Edwards and Bagozzi 2000 , p. 159). Although it may seem t o introduce an uncomfortable instability
in the structure of the operationalized concept, this dependence on the predicted variable actually
reflects the pragmatic value of judgments of all kinds, which are often formulated for a task at hand,
and not in abstracto .
The underlying reflective model of scientific literac y
Research practice generally inquires scientific literacy within reflective models, without usually
addressing the choice of measurement models, which remains implicit. The operationalized
knowledge of scientific constructs is understood as the respondents’ ability to provide correct
answers to miscellaneous quiz questions about fundamental scientific constructs, of which the scale
items represent a sample, and each construct is a replaceable i ndicator. A reflective measurement
model of scientific literacy may include additional latent variables that account for the co -variation
of items, such as an acquiescence response style and an assumed ignorance response style (Rughinis
and Toader 2010) .
The most usual computation of the individual score for the latent variable relies on counting the
number of correct answers to the quiz (see, for example, TNS OPINION & SOCIAL 2005b , p. 41 ;

50
Vlasceanu, Dușa, and Rughiniș 2010; National Science Board 2008 , p. 16 ). This method is easy to
implement. However, it intro duces errors in the estimate, since it grants each item an equal weight
in the final count, while their strength of causal association with the latent dimension may in fact
differ. At the same time, a count of correct answers ignores the inflation introduc ed by respondents’
acquiescence, understood as a disposition to answer “True” to all quiz items independently of their
specific content.
Item Response Theory (IRT) and Confirmatory Factor Analysis (CFA) provide means through which to
estimate reflective m odels more accurately. Miller uses a multiple group IRT method to estimate
respondent scores along a continuous latent trait, and, on this basis, he estimates country averages
for scientific literacy in Europe and in the US (Miller 1998b , pp. 212 -213, and Miller 2007c , pp. 4 -8)
and cohort averages for the US (Miller 2007a) . Pardo, Midden, and Miller (2002 , p. 11 ) use IRT as
well in order to estimate the level of knowledge of scientific constructs in the area of biotechnology.
While confirmatory factor analysis has been used to test the dimens ionality of the quiz scale (Miller
1998; Mill er 2007b; Miller 2007a; Bann and Schwerin 2004) , I did not encounter any CFA estimate of
average scores on the latent dimension. A CFA based reflective measurement model for scientific
literacy should also pass the tests of measurement invariance (Comșa 2010) , if there is an intention
to undertake cross -cultural comparisons.
A comparison of reflective and formative models of knowledge of scientific constructs.
The formative model does not seem relevant for the latent variable of general knowledge of
scientific constructs, if it is understood as a causal, underlying trait that produces answers to
interchangeable quiz items. Still, it may be relevant for knowledge of, or attitudes on scientific
constructs in s pecific thematic areas, where each construct may hold its own relevance. For
example, in Miller, Scott, and Okamoto 2006 (pp. 3 -4) the Ind ex of Genetic Knowledge is constructed
as a summative score, by counting correct answers to a quiz, under what seems to be a reflective
measurement model. On the contrary, the Attitude toward life index, which is also a summative
score that counts the pro -life answers to three questions about human life, could be understood as
following a formative model. It may be reasonably argued that the attitude toward life is not an
underlying trait that produces individual answers to the three topics but, the other w ay around, it is
the formative result of individual configurations of opinion on two key issues: the beginning of life,
measured by the first item, and the moral status of the embryo, measured reflectively by the next
two items.
While a summative score ma y approximate the score on a latent variable both under a reflective and
a formative model, the meaning of the observed variable differs according to its theoretical
specification. Reflective knowledge variables are interpreted as abilities or competences , while
formative knowledge variables are better interpreted as worldviews , if their scope is large, or
representations , if their scope is narrow. Whether a formative or a reflective model should be used
depends on the researchers’ choice.
Depending on the oretical considerations and, among others, on the degree to which high levels for
one of them compensate for lower levels for the others, the latent variable may be conceptualized
as continuous, ordered categorical or unordered categorical.
Moreover, star ting from the observation that specialists’ knowledge of scientific matters is
qualitatively different from lay people’s knowledge, which is more densely embedded in relational

51
and local contexts of significance, it may be argued that a given latent variab le can better be
operationalized under a formative model for the lay public, to match the fragmentary, composite
nature of knowledge, and under a reflective model for the specialists, to match the coherent,
systematic nature of their competences.
Table 5. A comparison of reflective and formative models for knowledge of scientific constructs
Issue Reflective model Formative model
Significance of the latent variable
of knowledge of scientific
constructs Competence
Ability Worldview
Represe ntation
Significance of latent variables
measuring knowledge of each
construct The true answer formulated by the
respondent when confronted with the quiz
item, in the interview situation The true knowledge of the respective
scientific constructs, prior to the
interview situation
Relationship between latent
knowledge of scientific
constructs, and knowledge of
each construct The competence influences the
respondents’ answers Knowledge of each construct shapes
respondents’ worldview
It can be estimated stat istically
by… Item Response Theory
Confirmatory Factor Analysis
Latent Class Analysis Structural Equation Models
(it requires insertion into a predictive
model for a dependent variable of
interest)
Shortcomings of summative
estimates Measurement model is often not explicit
Inflates knowledge estimates due to the
“True” response style
Incorporates random errors
Items are given equal weights despite
differential loadings of the indicators Measurement model is often not
explicit
Inflates knowledge estimates d ue to
the “True” response style
Incorporates random errors
Items are given equal weights despite
differential regression coefficients of
the latent variable on the indicators
Used for… Understanding the general public
knowledge of scientific constructs Modeling people’s worldviews or
representations in relation to specific
themes
Continuous or categorical latent
variable Usually modeled as continuous ability
May be modeled as categorical for
populations with highly uneven exposure
to science Depending on the theoretical model
Relevance of knowledge of
evolution as an indicator for the
latent operationalized variable Knowledge of evolution in the general
public is:
– Differently related to the underlying
ability for various religious groups
– Weakly related t o the underlying
ability for some conservative religious
groups Knowledge of evolution may be
relevant for several formative concepts,
such as:
– Acceptance of a scientific
worldview regarding religiously
contested concepts
– Human exceptionalism

The same difference between different types of knowledge of scientific constructs highlights another
hidden assumption of the reflective models discussed above, namely, that the latent variable is a
continuous ability or competence. It may be that a categorical con ceptualization would better fit the
discontinuous nature of the relationship between lay people without a background in science, and
trained specialists in scientific fields. It is also possible that the measurement model differs according
to other charact eristics of the social context. Working under the assumption that increased contact
with science, and prolonged scientific education introduce a qualitative change in knowledge of
scientific constructs, we can advance the hypothesis that a continuous abili ty is a proper
measurement model for countries or social contexts with relatively high exposure of the population
to scientific constructs, while a categorical model is better for contexts where people are
inconsistently exposed to science.

52
Table 5 above presents a comparison of the two measurement models for the vocabulary dimension
of civic scientific literacy.
4.3.2.2 Distinguishing constructs: q uiet versu s animated
Evolution is an example of an animated, existentially relevant scientif ic construct, and the next
sections focus on understanding its relevance for the NSF scale and the concept of scientific literacy,
in particular, and for public engagements with science in general. Apart from it, there are other such
constructs in the scal e. For example, “Cigarette smoking causes lung cancer” is an example of an
item that makes reference to matters of importance for many respondents – in this case, the good
and bad of their own and others’ smoking. The item on smoking has been eliminated fr om the short
NSF scale because it had a low loading on the underlying dimension (Bann and Schwerin 2004) and
low discrimination value in the IRT model, while also being one of the easiest items (idem). Besides
its unruly statistical behavior in the scale, there are other particularities of this item that are not
manifest in psychometric evaluations, such as the specific ways in which respondents actually
believe it to be true in general, but not true for them7.
The reproduction item, “It is the father’s gene which decides whether the baby is a boy or a gi rl,”
invokes a “gene” construct involved in emotionally charged conversations and decisions about
babies and reproductive strategies, having caused measurement predicaments too. Consequently,
the Oxford scale removed it (Durant, Evans, and Thomas 1992) . However, the NSF quiz still preserves
it, despite its low loading and discrimination power, but in order to maintain comparability of
summative scores with the longer earlier ve rsion of the quiz, given that it was the only item on
which women had systematically higher probabilities of correct answers than men (Bann and
Schwerin 2004) . This is another instance in which the everyday life relevance of a scientific construct
shapes its acquisition and appropriation processes, renderi ng it incommensurable with the more
remote scientific constructs.
The evolution item is particularly pertinent for evaluating the assumption that understanding
scientific constructs in lay knowledge is the same as believing in them. As the National Scienc e
Board’s report “Science and Engineering Indicators 2008” indicates, a considerable proportion of US
respondents are familiar with the scientific vocabulary of evolution, while declining to acknowledge
the fact of evolution. In the 2004 Michigan Survey of Consumer Attitudes, 74% of US respondents
agreed that “ According to the theory of evolution, human beings, as we know them today, developed
from earlier species of animals ”. In turn, only 42% agreed with the item formulated as a description
of the natural world, without the prefatory emphasis: “ Human beings, as we know them today,
developed from earlier species of animals ” (National Science Board 2008 , pp. 19 -20).
Confronted with the deletion of the evolution item from the 2010 NSF report, Miller answered that
“[p]art of being literate is to both understand and accept scientific constructs” (apud. Bhattacharjee
2010) . Then, the issue is whether a person who has a lay understanding of the scientific co ncept of
“evolution”, but does not accept it as descriptive for the world, is less scientifically literate than a
person who has the same lay understanding of the concept, but accepts it as descriptive of the
world. This matter raises definitional question s about the meaning of scientific literacy. If we define
scientific literacy as a general disposition to understand and believe in quiet scientific constructs,

7 Desantis (2003) offers an insightful account of how locally produced “collective rationalization” shapes beliefs
about scientific facts about cigar smoki ng.

53
than evolution seems not to be relevant for this measure, at least in the United States. If we d efine
scientific literacy as the belief in fundamental scientific constructs, including the quiet but also some
animated constructs, there seems to be no unique disposition to believe in them all. Beliefs in
animated constructs do not share the same proces s of causation with beliefs in quiet constructs.
Moreover, beliefs in animated constructs are not replaceable indicators, because each of them
illuminates particular, situational accomplishments for various publics in their work to understand
and live with science. Therefore, such an understanding of scientific literacy calls forth a definition in
terms of specific configurations of beliefs about scientific constructs, or possibly as worldviews,
instead of general individual dispositions.
The methodologica l implication of the distinction between animated and quiet scientific constructs
is that animated constructs may not be theoretically relevant for measuring the ability of the
respondent to understand (and believe in) quiet constructs, under a reflective measurement model,
which focuses on the underlying common cause of observed indicators. In addition, these constructs
may be even more interesting and telling about the public understanding of science than the
detached, classroom scientific facts. Their me asurements are meaningful not as interchangeable
indicators of respondents’ underlying abilities, but in themselves, as information about how science
is experienced in everyday life situations. Therefore, animated constructs are best measured either
on the ir own or as components of scientific worldviews.
To shed light on the incongruity of animated constructs in reflective models of scientific literacy, the
section below explicates the differences between concepts operationalized by reflective versus
format ive measurement models and presents the use of the reflective model for measuring scientific
literacy.
Evolution as an Animated Scientific Construct
Evolution is a key concept and established fact within the biological sciences, accounting for a
diverse bo dy of empirical evidence including observed transformations within living species,
similarity of structures in living and fossil species, or transitional structures observed in fossils (Gould
1983) . As Gould observes, “[ e]volutionists have been clear about this distinction between fact and
theory from the very beginning, if only because we have always acknowledged how far we are from
completely understanding the mechanisms (theory) by which evolution (fact) occurred” (Gould
1983 , p.2 ). Nevertheless, the fact of evolution has had a mixed reception among different public
groups, and it has constantly re -emerged at the forefront of public debates in volving the nature of
humankind and the relationship between religion and science.
Besides being a fundamental, textbook scientific construct, evolution has an existential relevance for
many people and is one of the very animated constructs, especially in the United States. A detailed
discussion of the understanding and use of evolution within common sense thinking or public
controversies lies beyond the scope of this paper. In what follows, I discuss evolution as an animated
scientific construct and briefl y highlight how it differs from the quieter constructs in the NSF quiz.
The scope of this distinction is to explore acceptance of evolution as a case in point for learning
about public engagement with science, as well as about the scientific literacy const ruct.
Unlike the working of the lasers, the movement of the continents, or the temperature of the centre
of the Earth, evolution is a highly debated concept, embedded in social interactions and
relationships. As compared to heavenly bodies or electrons, th e construct of evolution is used in

54
other types of communication and in different relational settings. On the one hand, evolution is
often evoked in interactions that have some religious or political underpinnings; on the other hand,
evolution is called up on in contested communications that require participants to take a stance and
negotiate an affiliation. Of course, the significance of evolution in such public encounters does not
derive from its capacity to account for empirical data from biological exper iments or from fossil
records, all of them being of largely no significance for the daily lives of most of us. In contrast,
evolution engages the public because of what it means for issues such as the exceptionality of the
human species, the meaning of lif e, or the validity of religious knowledge claims about the world in
confrontation with science. Is it true that human beings are essentially of the same kind as other
animals – or is there something truly special about the human consciousness, spirit, or s oul? Is it true
that human beings have appeared by utter chance, or was our entrance in the world purposefully
created? Does religion have anything to say about the constitution of the empirical world, or should
it delegate all empirical claims to science? It is in such a landscape that standpoints in favor or
against evolution are framed (Grimm 2009), shaped and settled8.
If circulated as a symbol for community allegiances, the salience of evolution as a true or false
representation of the world increases in social confrontations. This has been historically the case in
the US, where the rejection of evolution is a currency of the far -right politically conservative groups
(Mazur 2005; Miller, Scott, and Okamoto 2006b) , and to a growing extent in Turkey (Edis 1994;
Cavuslu 2009; Hamee d 2008) , but less in other European countries.
The religious and political import of evolution comes in conjunction with its seemingly counter –
intuitive character. Evolution and natural selection appear to be at variance with some deeply
ingrained common -sense thinking heuristics, arguably more than other scientific constructs. Macro –
evolution and the concept of emergent properties call into question essentialism underpinning the
lay conceptualization of the species and their traits, while the idea of ran dom variation in natural
selection and the reliance on statistical thinking counters finalist and mechanistic reasoning (M. E.
Evans 2008; Poling and Evans 2004; Thagard and Findlay 2009) .
The substantial body of research on the acceptance of evolution in the United States, particularly in
education sciences, sheds light on the interwoven conflicts, controversies, doubts and certainties
that mediate the ways in which evoluti on is understood, judged, and sometimes accepted (Eve
Raymond and Dunn 1990; Scott 1997; Meadows, Doster, and Jackson 2000; Grimm 2009; Superfine
2009; Nadelson and Southerland 2010) . For many people from the United States, encounters with
evolutio n are radically different from their encounters with lasers or electrons. This is a theoretically
relevant consideration to be taken into account when deciding whether evolution is a good indicator
in a reflective operationalization of scientific literacy.

8 The position of the Roman Catholic Church, as presented in the Message to the Pontifical Academy of
Sciences of John Paul II (1997) , is an illuminating example for how solutions to the evolution debate are
proposed in order to accommodate both scientific evidence and theses abou t the meaning of life. While
accepting that “the theory of evolution is more than a hypothesis” (p. 382), and granting its explanatory status
in relation to observed biological data, John Paul II has maintained human exceptionality and the
purposefulness o f human existence by clarifying that “[w]ith man, then, we find ourselves in the presence of
an ontological difference, an ontological leap, one could say. The sciences of observation describe and
measure the multiple manifestations of life with increasing precision and correlate them with the time line.
The moment of transition to the spiritual cannot be the object of this kind of observation, which nevertheless
can discover at the experimental level a series of very valuable signs indicating what is speci fic to the human
being” (p. 383).

55
5 Overview of research contributions
In the previous pages I have presented my main research contributions and I have discussed their
significance in the context of the literature on surveys as a form of sociological knowledge.
My recent research has contri buted to the methodological literature in fields of ethnic studies and
public knowledge of science . I have argued that survey -based research is to a signific ant extent
independent from sociological theorizing in other social research traditions , at least i n these
domains, and that measurement practices and the statistical apparatus shape a distinctive theorizing
style.
Some of the practices of survey -based sociological research that are highly theoretically loaded , and
which I have illustrated in my researc h, include:
– The decontextualization of respondents’ answers, by erasing interviewers’ participation at
multiple moments of the research; therefore, answers are made to describe respondents ,
instead of being interactional event s.
– What actually happens in th e survey is taken to be an approximation of a truer reality that is
affected by error; the analytical focus in not on what happened, but on detecting invisible
phenomena that lie behind empirical occurrences; this orientation towards evidence is
supported by an ensemble of procedures that I have termed the ‘error -trait’ device,
including techniques of identifying and removing error, imputing missing values, and
estimating latent constructs;
o To this purpose, aggregation is considered as a tool for gaining p recision;
o Also, inter -individual variability is used to understand intra -individual variability, as
cross -sectional correlations are used as input data for estimating the central
tendencies and variability of individual traits (Borsboom, Mellenbergh, and van
Heerden 2003) .
– Analysis methods such as regression, path and structural equation models represent
individual behavior as influenced by relatively stable individual traits, located within the
individual mind, whi ch may be best understood in isolation from other traits and abstracted
from the situation of interaction; this understanding of social action is similar to the
psychometric perspective in psychology, created as a theoretical by -product of the
quantitative methodology (following the account of Danziger , 1990) .
– The terminology of statistical analysis encourage a quasi -causal vocabulary of effects and
influences, which is additionally supported by the imputatio n of survey answers as
‘individual traits’ .
– The questionnaire imposes its structure of relevance for the interview conversation , which is
then transferred in account of respondents’ lives and actions. If researchers are interested in
ethnicity, for example , people will be represented as bearers of ethnic labels and marks in all
aspects of their lives that are covered by the survey – such as education, employment, family
strategies, migration and so on. The issue is no longer whether and in what context ethn icity
is relevant for actions in these fields , but only ‘to what extent ’ – or, in an even more specific
terminology, what is the ‘effect size’. Or, i f researchers are interested in scientific literacy,

56
which researchers have defined to include familiarity with lasers, th en people’s responses to
questions about lasers are taken as indicative of a certain trait that they possess and that
also explains ( more or less) their behavior when encountering some instantiation of
scientific constructs . Whether and in w hat circumstances respondents have had anything to
do with lasers or the ‘laser’ construct is deemed irrelevant.
– The issue of measurement error largely displaces considerations on whether quantification
is possible and relevant for a specific social proces s, and what are the particular, contextual
and changing problems of quantification in a certain area of interest. The treatment of error
is regionalized into issues of a) scale coherence and data reduction, b) model goodness of fit,
and c) sampling error, to the expense of a substantive evaluation of the meaning of
quantities produced in research and their various fo rms and sources of imprecision (Haslam
and McGarty 2001) . Discussions of errors are formulated within a framework of practic ality,
which translates issues of quantification in/adequacy, either theoretical or moral, as matters
of quantifiable ‘approximation’, by searching (and often finding) a level of imprecision that is
declared largel y adequate for the task -at-hand.
– Researche rs rely on typologies that resonate with common reason concerns and
classifications, with scarce attention to the connections between first-order and second –
order constructs (Schütz 1953) ; in this process, moral arguments and classifications enter
the research universe and its thread of argumentations wit hout the required reflexivity and
attention to theoretical implications. For example, the debate of hetero – versus self –
identification of ethnic affiliation has been largely driven by ethical and political concerns
external to the field of theoretical disc ussions of ethnicity; in the same vein, the concept of
‘scientific literacy ’ has introduced school -specific concerns in research concerning
knowledge of scientific constructs for a large variety of publics , thus marginalizing theories
concerning the specif icity of common sense knowledge.
These practices in survey -based sociological research support a theoretical view of social action that
largely ignores concrete, empirical social interactions. While individuals are conceived as participants
in interaction and as shaped in interaction, by processes of socialization, survey -based sociological
research attempts to study individual behavior as a result of traits that reside within the individual.
These traits are represented as shared, variable in intensity and sometimes in quality, and
potentially combined into types; nonetheless , a main focus of attention remains the individual trait,
operationalized as a variable.
Of course, researchers use survey -based evidence to work with a plurality of theoretical
perspe ctives. Survey -based research practices may favor a particular type of theoretical outlook, but
they do not determine it. In practice, survey evidence is put to use in a large variety of theoretically
informed inquiries. Surveys may accommodate research in to social structures and research into
individual actions, affording empirical investigations of key theoretical questions in sociology such as
the reproduction of inequalities or the creation and re -creation of institutions. Still, one weak spot
remains t heir low sensitivity to social interaction as it unfolds methodically, in the situated here -and-
now. In this thesis and in my further projects I am interested to explore the potential of survey –
based sociological research as a resource for investigations t hat take into account the interaction
order and its links with the institutional order (Warfield Rawls 2011; Warfield Rawls 1987) .

57
In this thesis I have discussed the significance of my theoretical and methodological analyses as
contributions to a re -orientation of survey -based research towards a n interactionally -sensitive
theoretical outlook. Such a change in theoretical grounding may be accomplished at different levels
of complexity. In the first instance, we can achieve a compatibility of survey -based analysis with
theoretical inquiries that ar e based on an interactional view of social action and language use. In a
second instance, surveys may be designed in a novel research agenda. Several analytical orientations
that may encourage a theoretical awareness of the interaction order include : a) an alyzing and
interpreting data as collaborative outcomes of situated interaction, b) re -specifying the concept of
‘error’ in line with participants’ understanding of what counts for a mistake – such as typing errors,
or misunderstandings – instead of decomp osing the variability of meaningful answers into “true
value variability” and “error variability”; c) higher reflexivity on analysts’ use of categories by
empirically studying their variable contextual relevance in social interaction, d) increased theoreti cal
attention to typification and second -order typology construction, e) less focus on statistical
significance in favor of a preoccupation with substantive troubles of quantification and issues of
absolute and relative size, and f) attention to the relian ce on common reason categories and
concerns, to our commentaries on them, as analysts, and to how our findings are likely to be taken
over in common reason social knowledge claims.
In the following paragraphs I discuss several of the main changes in resear ch practice that can
support such an adaptation of the survey instrument to an interaction order theory.
The causal jargon
Probably the easiest point to start such an effort is the use of the survey -report rhetoric, focusing on
the quasi -causal statistical vocabulary . Presentations of research results are loaded with concepts
that have causal implications, with the occasional disclaimer , in ‘Discussion’ sections, to the effect
that such terms may not really refer to causal relationships. If we do not want t o argue that avowed
ethnicity “explains ” behavior, or that it has a “direct effect” on expressed beliefs, it is probably best
not to say so, not even figuratively.
Confirmatory versus exploratory inquiries
Compatibility also requires a change in mission. A frequent classification of quantitative and
qualitative research methods marks the first as confirmatory, providing definite answers for
sociological hypotheses, and the second as exploratory, opening questions for further quantitative
investigation. From an interactional theoretical perspective, the relationship is virtually reversed. To
the extent that any analysis can be considered confirmatory, or providing stable answers to
sociological questions, such an analysis would be looking closely at a corpus of evidence concerning
situated, concrete, rich interactions. The distant, de – and re -contextualized, aggregated interaction
outcomes that are presented in traditional survey -based research can only have a status of
exploratory inquiry, proposing typologie s for further validation and drawing attention to phenomena
that, by their apparent intensity when so quantified, indicate that some strongly methodical
activities are at work in producing them.
Measurement errors, missing data, and other anomalies
A secon d major change required for such a theoretical compatibility refers to the analytical status of
measurement errors, biases, uncertainty, and missing data. Mainstream quantitative research is
oriented towards reducing uncertainty as much as possible (Haslam and McGarty 2001) and

58
correcting really -occurring events so that they would reveal a hidden reality. Statistical procedures
of aggregation, correlation, and imputation are put in place to repair reality to the shape of a truer
reality. The loyalty of a researcher working under an interactional theory lies strictly with actual
occurrences, at the expense of the so -called ‘latent ’ phenomena. Inconsistencies,
misunderstandings, refusal of cooperation, irony, humor, deceit are part and parce l of social
interaction and, to the extent that they are present as evidence, they are to be treated as evidence,
instead of being erased from view.
In particular, if ethnicity is the topic of study, refusals to answer questions on ethnicity, or
unexpected answers are the very subject matter for the investigators. Situations in which ethnicity is
immediately recognizable are equally interesting, and even more so are situations in which ethnicity,
as an interactional concern, comes out as strange, misunderst ood, or even absent (a situation rather
unlikely if questionnaires are explicitly oriented towards ethnicity). If public knowledge of science is
the topic of study, “Don’t know” answers are qualitatively different from volunteered answers that
later prove to be wrong. One’s lack of engagement with a question is as informative as one’s
engagement, and misunderstanding is as informative as understanding.
A sample of interactions
Survey -based research has the potential to become part of an agenda of large scal e qualitative
research, generating information about social encounters that covers a variety of social settings
otherwise unavailable to researchers. The issue of representative design becomes central: such
encounters should be meaningful for participants in relation with their out -of-the-interview lives, in
order to allow for some sort of intelligible connection. This richness of information is valuable if
survey answers are interpreted as events, and not as descriptions of people. Surveys may be
designed to sample conversations, encounters, searches – and to report on efforts to communicate
with people across a large geographical and social space.

59
6 Plans for future development
My current research plans are focused on developing a survey -based so ciological approach that
accommodates an interactional perspective, specifically for the study of subjective wellbeing – as
captured in common -reason and scientific terms such as happiness, satisfaction, quality time, flow,
meaningful life, or intrinsic mo tivation for an activity.
I have formulated two research projects that support this inquiry. The first project is concerned with
current research practices, in applied sociological investigations in the field of market research. The
second project aims to develop a methodology for quantifying subjective well -being as part of a
sociological inquiry into its interactional and institutional support, in present day Romania.
Table 6. Overview of my two current research projects
Project fe atures Igel LiSa
Implementation details Financed under the PN -II-RU-TE-2011 -2
program
2011 – 2014 Application under the ANCS Partnerships
Program, 2011
Currently under evaluation
Objective s Study of current research practices in
applied sociological inve stigations in the
field of market research
– Focus: use of quantification Development of a quantification
methodology that allows for theoretically
meaningful indicators of subjective well –
being
Main concepts Consumer motives, needs
Consumer satisfaction
Sociological imagination Subjective well -being
Happiness
Research questions concerning
quantification How do practitioners make use of
quantification techniques in market
research?
– What are the specific uses of
quantification, in relation to
their custome rs and their
professional objectives?
– How do they embed
quantification procedures in
the overall research process?
– What implicit and explicit
theories of consumer choice
and consumer behavior are
supported by practices of
quantification? How do different p ractices of
quantification answer to practical concerns
of various researchers and audiences?
How is quantitative evidence of happiness
embedded into larger research programs?
What is the theoretical relevance of
quantification processes? What theories of
happiness are explicitly and implicitly
supported?
Dialogue with other social
research threads Addressing t he academic – practitioner
divide in market research
Discussing the issue of d isciplinarity in
market research – the specificity of a
sociological perspective on researching
motives Engaging alternative traditions of
quantification of happiness:
a) Survey -based research on happiness
b) Experimental research on happiness
Trans -disciplinary dialogues Research findings are also oriented
towards:
a) so ciologists and social psychologists
that work as practitioners in market
research
b) sociology students interested in
market research Research findings are also oriented
towards:
a) respondents and other research
participants
b) public authorities and othe r
organizations interested in monitoring
quality of life
c) the general public and mass -media
d) students interested in studying
subjective wellbeing
Expected results Scientific publications
Study guide for students Scientific publications
Research report s addressing specific public
and professional concerns

60
6.1 Igel – Sociological imagination and disciplinary orientation in applied
social research. An inquiry into present -day market research in Romania
6.1.1 Scientific context: studies of science in applied and co rporate settings.
There is a rich body of social research investigating the production of scientific knowledge. The most
heated debates have centered on the contingent social construction of nature sciences in local
interaction situations (Collins 1983; Fine 1 996; K. D. Knorr -Cetina 1982; Latour and Woolgar 1986;
Karin D. Knorr -Cetina 1981; Shapin 1995; Krohn and van den Daele 1998) . The detailed descriptions
of science -in-the-making advanced by empirical investigations in research sites, especially in
laborat ory studies, has inspired a sociology of social scientific knowledge that is simultaneously
attentive to symmetries of natural, social and other kinds of knowledge, on the one hand, and to
assymetries between strands of sociological inquiry (Leahey 2008; Maynard and Schaeffer 2000) , on
the other hand . Sociological research on science, including sociologies of sociology, are drawn to
reflexively discuss the epistemic specificity and legitimacy of sociology itself. Our project takes over
the understanding of the heterogeneous, situational and interactional production of scientific
knowledge and employs it as a foundation for the investigation of corporate market research in
Romania. We thus relate to the research manifes to formulated by Penders, Verbakel, and Nelis
(2009) with regards to the social study of corporate science.
This projec t’s field of scrutiny, market research in Romania, is a Cinderella of science, displaying
manifold signs of weakness: an applied knowledge pursuit of for -profit organizations serving other
commercial clients, with a mosaic of instruments and theoretical mo dels crossing academic
disciplines, organized in an East -European country. At the same time, the very features that indict
market research as problematic from a normative positivist perspective also render it an important
subject matter for social inquirie s: it produces actionable knowledge about social actions, and it is
consequential for corporate business decisions, thus partaking in the creation of our material and
organizational environments.
A preliminary inquiry into the professional orientation of present -day Romanian market research
institutes, based on a review of their web -sites and several brief discussions with researchers and
managers, indicates three relevant organizational features: client diversity (marketing departments
that serve a single organization vs. research companies with multiple clients), size, and affiliation to
international corporations. Large companies offer a variety of services employing both qualitative
and quantitative methods, while some of the smaller ones are specialise d in a specific type of
methodology. Market research institutes may be part of larger international corporations. In this
case, they are significantly oriented towards a shared corporate research approach.
Present day market researchers in Romania have va rious educational backgrounds. Most of them
have graduated marketing, sociology, psychology, and other social sciences. The disciplinary
background of the research team is often presented on the company web site, which represents an
indicator of its signif icance, at minimum for public relations.
Therefore, we find ourselves part of a consistent tradition of scholarly reflection on applied social
research and, specifically, in the plentiful thread of work that relates academic and practitioner
knowledge in m arket research.
In the intense debate on actual and possible relationships between corporate market research and
academic scholarship, authors discuss possibilities and impossibilities of mutual relevance, while

61
charting and explaining differences and sim ilarities (Baines et al. 2009; Brannick and Coghlan 2006;
Brennan 2004; Brinberg and Hirschman 1986; Calof and Wright 2008; Catterall 1998; Cornelissen
2002; Razzaque 1998; Shugan 2004) . Compa risons direct attention to differences in the research
situations between applied and academic social research, invoking, among others, heterogeneous
research questions, strategies, and success criteria, audiences, available resources and constraints,
literary genres in reporting results, and ethical risks.
6.1.2 Research focus
We are particularly interested in the theoretical orientations of applied social research and market
research in particular. Previous research on theories in marketing has mainly discusse d theory as a
prerequisite of research design, looking for example at differences between practitioners and
academics in theorizing styles (Cornelissen 2002) , in affinity towards a positivist/empirist or a
relativist/constructivist epistemology (Razzaque 1998) , in understanding the value of qualitative
research (Catterall 1998) , or in research paths that connect concepts, met hods and empirical
evidence (Brannick and Coghlan 2006; Brinberg and Hirschman 1986) . The substantive theories that
orient marketing research have been discu ssed mostly in the context of its disciplinary classification;
for example, MacInnis and Folkes (2010) argue that consumer behavi or research is a
multidisciplinary subfield of marketing, identifying as its main theoretical strands the behavioral
decision theory, information processing psychology and consumer culture theory (p. 910). Hoffman
and Holbrook (2007) discuss the disciplinary focus of academic consumer research, di stinguishing
the “more macro level of sociology or anthropology” from the “more micro level of cognitive
psychology” (p. 514). Notably, the so -called micro – or interactionist perspectives of sociology and
anthropology are not mentioned in both analyses. Hu manistic, hermeneutic or social constructivist
approaches are often proposed for marketing research (see for example Arnold & Fischer, 1994;
Goulding, 2005; Hirschman, 1986) – but it is not clear how and if they are used in practice, mainly
because there is lit tle research on practitioners’ substantive theoretical orientations. Our project
aims to fill this gap.
We propose a novel concept as analytical tool when researching the sociological orientation of
market research: the interactional imagination . As a sta rting point in our approach, we define the
interactional imagination as a researcher’s disposition to attend to the situational and interactional
accomplishment of social action. Following the sociological debates on the specificity and autonomy
of an inte raction order (Goffman 1983; Warfield Rawls 1987) , we propose the concept of
interactional imagination as a sub -type of sociological imagination which may be consequential for
understanding the production of scientific knowledge in research sites (K. D. Knorr -Cetina 1982) and
for the daily professional decisions of soc ial researchers, both in qualitative or quantitative
investigations, academic or applied.
Sociological imagination, and interactional imagination in particular, may aptly be studied in relation
to two core scientific constructs in marketing research: motiv ation , in the theoretical register, and
focus group investigations , in the methodological register. The two constructs have a prominent
position in market research, they have been distinctively developed in sociology, as well as in other
disciplines, and t hey confront the practitioner with a wide repertoire of possible interpretations and
decisions in research design.
Research on motives has figured prominently in market research since its very beginnings, both in
psychologically unsophisticated surveys (G. Wagner 1938) and in innovative the oretical

62
developments. The motivational research school of Ernest Dichter has been “the most significant
area of consumer research in the 1950s” (Stern 2004) , and research on motives has continued to
develop after its decline, in various theoretical frameworks. In the meanwhile, between Wagner’s
(1938) critical discussion of reliance of motives in market research and Dichter’s development of
motivational research, Mills (1940) published his influential work on “Situated Actions and
Vocabularies of Motive”, which has shaped resear ch on motives across the discipline. His focus on
motives as shared vocabularies, methodically used to render action intelligible to oneself and to
others, has later been taken further in arguments that motives are of interest in sociology only as a
metho dical activity of motive ascription (Blum and McHugh 1971) , and it has been critiqued for
initiating an unjustifiable displacement of motives from sociological analyses (Campbell 1996) .
Despite their clear potential to serve as an argument against a search for internal, private states that
act as resorts for action (see for example Hopper, 2001) , vocabularies of motive are also employed in
research that relies on a conception of action driven by causally antecedent motives (Corey and
Wilson 1994; Monaghan 2002) . Sociological reflection on motives is relevant for understanding
consumer choices (see also the discu ssion of Campbell, 1998 on the rhetoric of need and want) and
for understanding conversations about motives, including interview questions (Bold en and Robinson
2011) – with undeniable methodological significance. Therefore, an empirical inquiry into the
theoretical and methodological treatment of motives in current market research is a particularly
germane approach in searching for disciplinary orientations and sociological imagination.
If motives have constituted a theoretical benchmark for market research, focus groups have been at
the core of its methodological repertoire, in particular in the qualitative approaches. There is a
wealth of theor etically -laden methodological discussion of this research method, which comes in a
variety of designs, purposes, and interpretations (Boddy 2005; Morgan 1996) . Focus groups may be
used as a “quick and cheap” way of gathering opinions or attitudes (Catterall, 1998, p. 72) or as a
delicately balanced method for u nderstanding interactional dynamics and situationally generated
orders (Kitzinger 1994) . Consequently , the method is often discussed, also in consumer research,
with reference to misuse and misinterpretation (Threlfall 1999) . As a research tool with sociological
tradition and rich interactional relevance, focus groups offer an o pportunity for understanding the
sociological and interactional imagination at work in market research.
6.1.3 Method ology
Our project consists of a sociological research on disciplinary orientations and sociological
imagination in market research, in the cities of Bucharest and Cluj -Napoca. We will rely mostly on
practitioners’ accounts of their research work, in individual and group discussions. Whenever
possible, given constraints of confidentiality in corporate market research, we shall also engage in
observat ion of research sites and activities, and documentary analysis of research reports and
handbooks.
Our methodology relies primarily on single occasion and repeated interviews with practitioners ,
joined by research workshops bringing together practitioners a nd academics, documentary analysis
of research texts and textbooks, and observation in market research organizations – thus following a
rich thread of sociological investigation of science via scientists’ accounts (Gilbert and Mulkay 1980;
Lee and Roth 2004; Mulkay and Gilbert 1982; Mitroff 1974) , and methodological inquiries into the
kinds of knowledge accesible by such methods (such as in Karin D. Knorr -Cetina, 1981; Mulkay & G.
Nigel Gilbert, 1983; S. W. Woolgar, 1976; Yearley, 1988) .

63
6.1.4 Impact, relevance, applications
The main scientific impact of this project consists in developing a sociological understanding of
theory use and theoretical orientation in market research – as regards disciplinary distinctions and
sociological imagination. T he project will ellaborate the concept of interactional imagination, and it
will assess its theoretical value. Our research will also support the ellaboration of a methodological
reflection and report on analyzing interviews with practitioners, including r epeated interviews, in
order to understand applied social research.
Based on our inquiry, we will propose a curriculum module for sociology departments aimed at
market researchers and other applied research practitioners. This contribution to curriculum
development will support students’ understanding and meaningful use of sociological imagination in
applied social research.
6.2 LiSa – Gaps and bridges. Pursuing i ndividual life satisfaction and
happiness in the public sphere
6.2.1 Scientific context
There has been a gradual recognition, both in scientific and policy arenas, that economic indicators
do not capture everything about personal, organizational and societal well -being (Yan, 2008;
Easterlin, 2001; Boyce, Brown and Moore, 2010). There is a powerful tendency t o capture social
indicators of subjective well -being, alongside indicators that aim to measure the objective
circumstances of life, including the economic environment.
The urge for such an analysis is originating primarily in the famous two -years -old repo rt of J. Stiglitz,
A. Sen and J. -C. Fitoussi on the measurement of economic performance and social progress (Stiglitz,
Sen and Fitoussi, 2009). The issues raised by this report have been also taken over by OECD in its
Better Life Initiative (OECD, 2011). B oth projects underline that the standard macroeconomic
indicators fail to give an adequate account of individuals’ well -being and social progress, and that
policy making should focus not only on standard macroeconomic statistics, like those related to GDP,
for evaluating the current state of a society, but also on citizens’ well -being in order to ensure access
to opportunities and personal development. This tendency in social indicators is convergent with the
well consolidated focus at organizational levels for evaluating employee domain -specific and overall
happiness and life satisfaction. Moreover, a rich body of current survey research problematizes the
paradoxical relationships between economic status and subjective well -being at individual and
societal levels (see, for example, syntheses of previous research in Easterlin 2003, Veenhoven 2004,
or Kahneman and Deaton 2010).
What are the implications of these research threads and policy orientations for an enhanced public
and scientific understanding of th e ongoing transformations in the Romanian society? Unfortunately
there are no references to Romania in these reports, and, as researchers in social sciences, we are
confronted with a national shortage of data and information when attempting to undertake a
comparative analysis of the sort.
6.2.2 Research focus
The LiSa project engages the current global debate on the relevance of happiness and life
satisfaction in public arenas, aiming to contribute empirically and conceptually to its development.
We align our pr oject with currents of research and theoretical reflection that argue that happiness

64
and life satisfaction are worthwhile topics of personal and professional reflection, effort and also of
public policy (Veenhoven, 2004) – while acknowledging the considera ble challenges in
understanding and fostering subjective well -being (Kahneman 2003, 2006), both due to patterned
styles of individual judgment and action, and to the considerable diversity of worldviews and
lifestyles – challenges amplified by the limited predictability of public policy results.
Our research approach is innovative in four main respects:
– Empirical coverage : From an empirical point of view, Romania has been an almost uncharted
territory in this respect, due to lack of comparable information – as witnessed from the absence
of data on Romania in inter alia the recent OECD report. We plan to open this debate nationally
and internationally by collecting comparable national level data on subjective well -being
(including the measures employed in t he OECD report, such as the Cantril ladder and the
measurement of positive affect balance, as well as other relevant instruments).
– Focus on culturally informed agency : Conceptually, the current debate is focused on measuring
subjective well -being as a stat e of fact, interpreted as a personal reflection of previous and
current life experiences that, along with societal standards, shape individuals’ requirements and
expectations and, consequently, their subjective well -being (Pop, 2008). We propose to
investi gate happiness not only as an existent reality, but as an on -going accomplishment, by
looking at how people actively pursue it – following the research tradition on emotion work
opened by Russell Hochschild (1979). Therefore, we focus on how people conceiv e of their
present and future well -being and how they search for it: in what social spaces is happiness to
be found? What are the lay theories of happiness, and how do they orient social action – in the
present -day Romanian society? How do these lay theori es relate to the empirical configurations
observed in social research (Rughiniș, 2007)? What types and configurations of engagement with
the available social worlds are sought for, and how is the private / public divide constructed in
this pursuit? Specifi cally, how is happiness understood and pursued in particular professional
worlds, such as education?
– A configurational approach : We also acknowledge that subjective well -being is to be understood
not only unidimensionally, as a difference between positive and negative affects or evaluations,
but as a phenomenological configuration of happiness and unhappiness, satisfaction and
dis-satisfaction – since both tonalities are meaningful experiences that often co -exist in complex
world -views and self -definitions, creating tensions that energize action. Our project charts the
positive and negative experiences distinctly, and it inquires into how persons orient their actions
in relation to them.
– Focus on public phenomena : Moreover, we conceptualize happiness and lif e satisfaction as
essentially public, interactional phenomena – even if they are pursued in the private spheres of
life. We investigate subjective well -being as it is expressed in public interactions, and as it is
shaped by shared cultural theories, lay an d specialized, individual and organizational, about
living a good life. We inquire into the social practices (e.g.: embedded constructs, beliefs,
learning experiences) of pursuing happiness and life satisfaction, their distributions across social
spaces, w ith a focus on professional worlds of business and education, and their reliance on
shared discourses and vocabularies of motive – in the research tradition of Mills (1940) – of how
and where a good life is to be found.

65
The project thus relies on a twin in quiry:
– On individuals’ active pursuit of happiness and life satisfaction, across a multiplicity of private
and public spaces – and in relation to their position in a multidimensional social and economic
structure;
– On cultures and lay theories of subjective well-being, following them at individual, interactional,
organizational and social levels.
6.2.3 Methodology
Our theoretical and methodological premise is that individuals cannot be understood as isles, but as
participants in social worlds. As such, our lives d isplay both social influences, observable as stable
patterns and relationships, and agency, our own constant work of defining and answering life’s
situations. We rely on a large scale survey to observe social regularities and thus to gain insight into
the power of social structures to shape subjective well -being, and we conduct a thorough and
in-depth qualitative investigation to explore happiness -in-the-making, with a focus on the discursive
work in which individuals define, explain, justify, project and p lan for their happiness, or for others’
subjective wellbeing (Rughiniș, 2007).
The project therefore employs qualitative and quantitative methods designed to support one
another in identifying and focusing on the relevant empirical evidence. The main stage s of the
empirical research include: 1) an exploratory qualitative investigation, 2) instrument development
and testing, 3) a quantitative survey of the general population (2000 respondents), 4) in -depth
qualitative research, including individual interview s (120 respondents) and focus groups (4
interviews), and 5) a final wave of in -depth interviews pursuing the salient findings from quantitative
analysis.
6.2.4 Impact, relevance, applications
The project expected results include:
– A better understanding of the culturally informed pursuit of subjective well -being in the general
public and in specialized professional communities of practice (communicated in scientific
publications);
– A comprehensive model of the configurations of capital -related and subjective indi cators, and
inequalities in their distributions in present -day Romania (communicated in scientific
publications);
– Instruments for a sustained reflexivity on subjective well -being in academic and policy arenas,
including:
o Scientific instruments, such as mea surement scales and inventories, and datasets
available for further secondary analysis;
o Organizational instruments, such as structured themes for reflection on subjective
well-being in various types of events and evaluations;
o Policy -relevant instruments, i ncluding a report that discusses the measured indicators
and their utility for policy design;
o Online instruments for the engagement of the general public, developed on the project
site.

66
7 References

Ahmed, Patr icia, Cynthia Feliciano, and Rebecca Jean Emigh. 2006. “Internal and External Ethnic
Assessments in Eastern Europe.” California Center for Population Research On -Line Working
Paper Series . http://papers.ccpr.ucla.edu/
Ajzen, Icek. 1991. “The theory of p lanned behavior.” Organizational Behavior and Human Decision
Processes 50 (2): 179 -211. doi:10.1016/0749 -5978(91)90020 -T.
http://linkinghub.elsevier.com/retrieve/pii/074959789190020T.
——— . 2011. “The theory of planned behaviour: reactions and reflections.” Psychology & health 26
(9) (September): 1113 -27. doi:10.1080/08870446.2011.613995.
http://www.ncbi.nlm.nih.gov/pubmed/21929476.
Allum, N., P. Sturgis, D. Tabourazi, and I. Brunton -Smith. 2008. “Science knowledge and attitudes
across cultures: a meta -analy sis.” Public Understanding of Science 17 (1) (January 1): 35 -54.
doi:10.1177/0963662506070159. http://pus.sagepub.com/cgi/content/abstract/17/1/35.
Allum, Nick, Patrick Sturgis, Dimitra Tabourazi, and Ian Brunton -Smith. 2008. “Science Knowledge
and Attitud es across Cultures: a Meta -Analysis.” Public Understanding of Science 17: 35 -54.
American Sociological Association. 2003. The Importance of Collecting Data and Doing Social
Scientific Research on Race. Washington, DC: American Sociological Association .
http://www2.asanet.org/media/asa_race_statement.pdf.
Antaki, Charles. 2006. “Producing a ‘cognition’.” Discourse Studies 8 (1) (February 1): 9 -15.
doi:10.1177/1461445606059545. http://dis.sagepub.com/cgi/content/abstract/8/1/9.
Arnold, Stephen J, and Eileen F ischer. 1994. “Hermeneutics and Consumer Research.” Journal of
Consumer Research: An Interdisciplinary Quarterly 21 (1): 55 -70.
http://ideas.repec.org/a/ucp/jconrs/v21y1994i1p55 -70.html.
Aspinall, Peter J. 1997. “The conceptual basis of ethnic group termin ology and classifications.” Social
Science & Medicine 45 (5) (September): 689 -698. doi:10.1016/S0277 -9536(96)00386 -3.
http://dx.doi.org/10.1016/S0277 -9536(96)00386 -3.
——— . 2001. “Operationalising the collection of ethnicity data in studies of the sociology of health
and illness.” Sociology of Health & Illness 23 (6) (November): 829 -862. doi:10.1111/1467 –
9566.00277. http://doi.wiley.com/10.1111/1467 -9566.00277.
——— . 2009. “‘Mixed race’, ‘mixed origins’ or what? Generic terminology for the multiple
racial/eth nic group population.” Anthropology Today 25 (2) (April): 3 -8. doi:10.1111/j.1467 –
8322.2009.00653.x. http://doi.wiley.com/10.1111/j.1467 -8322.2009.00653.x.
——— . 2012. “Answer Formats in British Census and Survey Ethnicity Questions: Does Open
Response Bett er Capture ‘Superdiversity’?” Sociology 46 (2) (February 2): 354 -364.
doi:10.1177/0038038511419195.
http://soc.sagepub.com/cgi/content/abstract/0038038511419195v1.

67
Augoustinos, M., and D. Every. 2007. “The Language of ‘Race’ and Prejudice: A Discourse of D enial,
Reason, and Liberal -Practical Politics.” Journal of Language and Social Psychology 26 (2) (June
1): 123 -141. doi:10.1177/0261927X07300075.
http://jls.sagepub.com/cgi/content/abstract/26/2/123.
Auyero, J ., and D . Swistun. 2008. “The Social Production o f Toxic Uncertainty.” American Sociological
Review 73 (3): 357 -379. doi:10.1177/000312240807300301.
http://asr.sagepub.com/cgi/doi/10.1177/000312240807300301.
Babusik, Ferenc. 2004. “Legitimacy, Statistics and Research Methodology Who Is Romani in Hungary
Today and What Are We (Not) Allowed to Know About Roma.” Roma Rights Quarterly 1: 14 -18.
Baines, Paul R., Ross Brennan, Mark Gill, and Roger Mortimore. 2009. “Examining the
academic/commercial divide in marketing research.” European Journal of Marketing 43
(11/12): 1289 -1299. doi:10.1108/03090560910989894.
http://www.emeraldinsight.com/journals.htm?issn=0309 –
0566&volume=43&issue=11/12&articleid=1822064&show=html.
Bann, Carla M., and Michael J. Schwerin. 2004. “Public knowledge and attitudes scale constructi on:
development of short forms.” RTI International .
http://www.rti.org/pubs/PKAS_Short_Form_Report.pdf .
Barth, Fredrick. 1969. Introduction. In Ethnic groups and boundaries. The social organization of
culture difference , ed. Fredrick Barth. Boston: Little Brown & Co .
Bauer, Martin. 1996. “Socio -Demographic Correlates of DK -Responses in Knowledge Surveys: Self –
attributed Ignorance of Science.” Social Science Information 35 (1): 39 -68.
——— . 2009. “The Evolution of Public Understanding of Science Discourse and Comparative
Evidence .” Science, Technology & Society 14 (2): 221 -240.
Bauer, Martin W. 2009. “The Evolution of Public Understanding of Science —Discourse and
Comparative Evidence.” Science Technology Society 14 (2): 221 -240.
Bauer, Martin W., Nick Allu m, and Steve Miller. 2007. “What can we learn from 25 years of PUS
survey research? Liberating and expanding the agenda.” Public Understanding of Science 16:
79-95.
Bhattacharjee, Yudhijit. 2010. Evolution, big bang polls omitted from NSF report. Science I nsider .
http://news.sciencemag.org/scienceinsider/2010/04/evolution -big-bang -polls -omitted.html.
Bhopal, Raj. 2006. “Race and Ethnicity: Responsible Use from Epidemiological and Public Health
Perspectives.” Journal of Law, Medicine & Ethics : 500 -07.
Billie t, Jaak. 2010. Controlling for measurement error and construct equivalence concepts in cross –
nation research: used methodologies and examples from ESS . 9th ESS Train . Ljubljana.
Billig, M. 1999a. “Whose Terms? Whose Ordinariness? Rhetoric and Ideology in C onversation
Analysis.” Discourse & Society 10 (4): 543 -558. doi:10.1177/0957926599010004005.
http://das.sagepub.com/cgi/doi/10.1177/0957926599010004005.

68
——— . 1999b. “Critical discourse analysis and conversation analysis: an exchange between Michael
Billig and Emanuel A. Schegloff.” Discourse and Society 10 (4): 543 -582.
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Critical+discourse+analysis+a
nd+conversation+analysis:+an+exchange+between+Michael+Billig+and+Emanuel+A.+Schegloff
#0.
Billig, Mi chael. 2008. “The language of critical discourse analysis: the case of nominalization.”
Discourse & Society 19 (6): 783 -800. doi:10.1177/0957926508095894.
http://das.sagepub.com/cgi/doi/10.1177/0957926508095894.
Blanchard, Chris, and et. al. 2003. “Ethnici ty and the Theory of Planned Behavior in the Exercise
Domain.” American Journal of Health Behavior 27 (6): 579 -591.
Blum, Alan F., and Peter McHugh. 1971. “The Social Ascription of Motives.” American Sociological
Review 36 (1) (December): 98 -109. http://ww w.jstor.org/pss/2093510.
Boddy, Clive. 2005. “A rose by any other name may smell as sweet but ‘group discussion’ is not
another name for a ‘focus group’ nor should it be.” Qualitative Market Research: An
International Journal 8 (3) (January): 248 -255. doi: 10.1108/13522750510603325.
http://www.emeraldinsight.com/journals.htm?issn=1352 –
2752&volume=8&issue=3&articleid=1505932&show=html.
Bolden, Galina B., and Jeffrey D. Robinson. 2011. “Soliciting Accounts With Why -Interrogatives in
Conversation.” Journal of C ommunication 61 (1) (February): 94 -119. doi:10.1111/j.1460 –
2466.2010.01528.x. http://doi.wiley.com/10.1111/j.1460 -2466.2010.01528.x.
Bollen, Kenneth, and Richard Lennox. 1991. “Conventional Wisdom on Measurement: A Structural
Equation Perspective.” Psychol ogical Bulletin 110 (2): 305 -314.
Borsboom, Denny, Gideon J. Mellenbergh, and Jaap van Heerden. 2003. “The theoretical status of
latent variables.” Psychological review 110 (2): 305 -314.
Brannick, Teresa, and David Coghlan. 2006. “To know and to do: Academ ics’ and practitioners’
approaches to management research.” Irish Journal of Management 26 (2): 1 -23.
http://www.iamireland.com/book/IJM_26_2_Final.pdf#page=9.
Brennan, Ross. 2004. “Should we worry about an ‘academic -practitioner divide’ in marketing?”
Mar keting Intelligence & Planning 22 (5) (January): 492 -500. doi:10.1108/02634500410551879.
http://www.emeraldinsight.com/journals.htm?issn=0263 –
4503&volume=22&issue=5&articleid=854687&show=html.
Briciu, Cosmin. 2007. Cartierul Dodeni -Bicaz. In Nevoi și resur se în comunitățile de romi , ed. Malina
Voicu, 27. Bucuresti: Fundatia Soros Romania.
Brinberg, David, and Elizabeth C. Hirschman. 1986. “Multiple Orientations for the Conduct of
Marketing Research: An Analysis of the Academic/Practitioner Distinction.” Journal of
Marketing 50 (4): 161 -173. http://faculty.fuqua.duke.edu/~willm/Classes/PhD/PhD_2007 –
2008/MarketingVersionOfRSvBS_0 -3.pdf.

69
Brown, Graham K., and Arnim Langer. 2010. “Conceptualizing and Measuring Ethnicity.” Oxford
Development Studies 38 (4) (Decem ber): 411 -436. doi:10.1080/13600818.2010.525629.
http://dx.doi.org/10.1080/13600818.2010.525629.
Brubaker, Rogers. 2002. “Ethnicity without groups.” Archives Europeenes de Sociologie 18 (2): 163 –
89.
——— . 2004. Ethnicity without groups . Cambridge: Harvard U niversity Press .
Brubaker, Rogers, Mara Loveman, and Peter Stamatov. 2004. Ethnicity as Cognition . vol. 33. Theory
and Society .
Bryan, Angela, Monica S. Ruiz, and Dennis O’Neill. 2003. “HIV -Related Behaviors Among Prison
Inmates: A Theory of Planned Beha vior Analysis.” Journal of Applied Social Psychology 33 (12):
2565 -2586.
Burton, Jonathan, Alita Nandi, and Lucinda Platt. 2010. “Measuring ethnicity: challenges and
opportunities for survey research.” Ethnic and Racial Studies 33 (8) (September): 1332 -1349.
doi:10.1080/01419870903527801.
http://www.tandfonline.com/doi/abs/10.1080/01419870903527801.
Calof, Jonathan L., and Sheila Wright. 2008. “Competitive intelligence: A practitioner, academic and
inter -disciplinary perspective.” European Journal of Market ing 42 (7/8): 717 -730.
doi:10.1108/03090560810877114. http://www.emeraldinsight.com/journals.htm?issn=0309 –
0566&volume=42&issue=7/8&articleid=1734885&show=html.
Campbell, Colin. 1996. “On the Concept of Motive in Sociology.” Sociology 30 (1) (February): 10 1-
114. doi:10.1177/0038038596030001007.
http://soc.sagepub.com/cgi/content/abstract/30/1/101.
——— . 1998. “Consumption and the rhetorics of need and want.” Journal of Design History 11 (3)
(March): 235 -246. doi:10.1093/jdh/11.3.235.
http://jdh.oxfordjournal s.org/cgi/content/abstract/11/3/235.
Catterall, Miriam. 1998. “Academics, practitioners and qualitative market research.” Qualitative
Market Research: An International Journal 1 (2) (January): 69 -76.
doi:10.1108/13522759810214244. http://www.emeraldinsight .com/journals.htm?issn=1352 –
2752&volume=1&issue=2&articleid=858356&show=html.
Cavuslu, E. 2009. Clash of discourses: The discourses on evolution and creationism in Turkey .
http://vssr.info/docs/2009_participants/Cavuslu -Clash of Discourses.pdf .
Chandra, K., and S. Wilkinson. 2008. “Measuring the Effect of ‘Ethnicity’.” Comparative Political
Studies 41 (4 -5) (April 1): 515 -563. doi:10.1177/0010414007313240.
http://cps.sagepub.com/cgi/content/abstract/41/4 -5/515.
Cicourel, Aaron. 1964. Method and Measuremen t in Sociology . New York: The Free Press.
Collins, H M. 1983. “The Sociology of Scientific Knowledge: Studies of Contemporary Science.”
Annual Review of Sociology 9 (1) (August): 265 -285.
doi:10.1146/annurev.so.09.080183.001405.

70
http://www.annualreviews.or g/doi/abs/10.1146/annurev.so.09.080183.001405?journalCode=
soc.
Comșa, Mircea. 2010. “How to compare means of latent variables across countries and waves:
testing for invariance measurement. An application using Eastern European societies.”
Sociológia – Slovak Sociological Review 42 (6): 639 -669.
Connolly, P. 2011. “Us ing survey data to explore preschool children’s ethnic awareness and
attitudes.” Journal of Early Childhood Research 9 (2) (June 23): 175 -187.
doi:10.1177/1476718X10387898. http://ecr.sagepub.com/cgi/content/abstract/9/2/175.
Conrad, F G, and M F Schober. 2000. “Clarifying question meaning in a household telephone survey.”
Public Opinion Quarterly 64 (1): 1 -28. http://www.ncbi.nlm.nih.gov/pubmed/10810073.
Corey, Robert J., and David T. Wilson. 1994. “Using Motivation as a Basis for Understanding and
Improvi ng Direct Marketing Relationships.” Journal of Direct Marketing 8 (4): 28 -39.
Cornelissen, J. 2002. “Academic and Practitioner Theories of Marketing.” Marketing Theory 2 (1)
(March): 133 -143. doi:10.1177/1470593102002001647.
http://mtq.sagepub.com/cgi/cont ent/abstract/2/1/133.
Covrig, Anca. 2004. “Why Roma do not Declare their Identity – Careful Decision or Unpremeditated
Refusal?” Journal for the Study of Religions and Ideologies (8): 90 -101.
http://jsri.ro/ojs/index.php/jsri/article/viewFile/186/186.
Cozm a, Irina, Cosima Rughiniș, and Leonard Sultănescu. 2003. “Impactul social al politicilor publice în
cartierul Zăbrăuți. Studiu de caz.” Revista de Asistență Socială (1): 75 -83.
Csepeli, György, and David Simon. 2004. “Construction of Roma identity in Easte rn and Central
Europe: Perception and self – identification.” Journal of Ethnic and Migration Studies 30 (1):
129-150. http://www.csepeli.hu/elearning/cikkek/csepeli_simon.pdf.
Danziger, Kurt. 1990. Constructing the Subject: Historical Origins of Psychologi cal Research .
Cambridge studies in the history of psychology . Cambridge University Press.
http://www.loc.gov/catdir/description/cam032/89022160.html.
Desantis, Alan D. 2003. “A couple of white guys sitting around talking: The collective rationalization
of cigar smoking .” Journal of Contemporary Ethnography 32 (4): 432 -466.
Diamantopoulos, Adamantios, and Heidi M. Winklhofer. 2001. “Index Construction with Formative
Indicators: An Alternative to Scale Development .” Journal of Marketing Research 38 (2): 269 –
277.
Duminică, Gelu, Simona Lupu, and Cosima Rughiniș. 2009. Persoanele de etnie romă în situații de
risc. In Riscuri și inechități sociale în România , ed. Marian Preda, 226 -232. Iași: Polirom.
Durant, John, Geoffrey Evans, and Geoffrey Thomas. 1992. “Public Un derstanding of Science in
Britain: the role of medicine in the popular representation of science.” Public Understanding of
Science 1: 161 –182.

71
Durnescu, Ioan, Cristian Lazar, and Roger Shaw. 2002. “Incidence and Characteristics of Rroma Men
in Romanian Pri sons.” The Howard Journal of Criminal Justice 41 (3) (July): 237 -244.
doi:10.1111/1468 -2311.00239. http://www.blackwell -synergy.com/links/doi/10.1111/1468 –
2311.00239.
Edis, Taner. 1994. “Islamic creationism In Turkey.” Creation/Evolution 34 (1).
http://www 2.truman.edu/~edis/writings/articles/islamic.html .
Edwards, Derek. 2003. Analyzing racial discourse; the discursive psychology of mind -world
relationships. In Analyzing Race Talk: Multidisciplinary Approaches to the Interview , ed. H. van
den Berg, M. Weth erell, and H. Houtkoop -Steenstra, 31 -48. Cambridge, UK: Cambridge
University Press.
Edwards, Jeffrey R., and Richard P. Bagozzi. 2000. “On the nature and direction of relationships
between constructs and measures .” Psychological Methods 5 (2): 155 -174.
Emigh, Rebecca, and Iván Szelényi. 2001. Poverty, ethnicity and gender in Eastern Europe during the
market transition Westport : Praeger Publishers .
Evans, Geoffrey, and John Durant. 1995. “The relationship between knowledge and attitudes in the
public unde rstanding of science in Britain .” Public Understanding of Science 4 (1): 57 -74.
Evans, Margaret E. 2008. Conceptual change and evolutionary biology: A developmental analysis. In
International Handbook of Research on Conceptual Change , ed. Stella Vosniado u, 263 -294.
New -York, NY: Routledge .
Eve Raymond, A., and Dana Dunn. 1990. “Psychic powers, astrology & creationism in the classroom?
Evidence of pseudoscientific beliefs among high school biology & life science teachers.” The
American Biology Teacher 52 (1): 10 -21.
Fayard, Pierre. 1992. “Let’s stop persecuting people who don't think like Galileo!” Public
Understanding of Science 1 (1): 15 -16. doi:10.1088/0963 -6625/1/1/003.
http://pus.sagepub.com.
Fine, Arthur. 1996. Science Made Up: Constructivist Sociolo gy of Scientific Knowledge. In The
Disunity of Science: Boundaries, Contexts, and Power , ed. Peter Galison and David Stump, 231 –
254. Stanford: Stanford University Press. http://faculty.washington.edu/afine/SCIENCE MADE
UP.pdf.
Fleck, Gabor, and Cosima Rugh iniș, eds. 2008. Come Closer. Inclusion and Exclusion of Roma in
Present Day Romanian Society . Bucharest: Human Dynamics.
http://books.google.com/books?id=ck_kFYKjeBkC&pgis=1.
Fleck, Gábor, and Cosima Rughinis. 2008. Come closer. Inclusion and exclusion of Roma in present –
day Romanian society . Human Dynamics . http://books.google.com/ .
Gabriel Bădescu, and et.al. 2007. Barometrul incluziunii romilor / Roma inclusion barometer .
http://www.osf.ro/en/publicatii.php?cat=16 .
Garfinkel, Harold. 1945. Color Tr ouble. In Primer for white folks , ed. Bucklin Moon, 269 -286. Garden
City, New York: Doubleday, Doran and Co., Inc.

72
Gilbert, Nigel, and Michael Mulkay. 1980. Contexts of Scientific Discourse: Social Accounting in
Experimental Papers. In The Social Process o f Scientific Investigation. Sociology of the Sciences ,
ed. Karin D. Knorr, Roger Krohn, and Richard Whitley, 269 -294. Dordrecht: D. Reidel Publishing
Company.
Godin, Benoit, and Yves Gingras. 2000. “What is scientific and technological culture and how is i t
measured? A multidimensional model.” Public Understanding of Science 9 (1): 43 -58.
doi:10.1088/0963 -6625/9/1/303. http://pus.sagepub.com/cgi/content/abstract/9/1/43.
Goffman, Erving. 1983. “The Interaction Order: American Sociological Association, 1982 P residential
Address.” American Sociological Review 48 (1): 1 -17.
Gould, Stephen Jay. 1983. Evolution as Fact and Theory . In Further Reflections in Natural History ,
ed. Hen’s Teeth and Horse's Toes S. J. Gould, 253 -262. New York : W. W. Norton & Company.
http://courses.washington.edu/anth599/Evolution as Fact and Theory Gould 1981.pdf .
Goulding, Christina. 2005. “Grounded theory, ethnography and phenomenology: A comparative
analysis of three qualitative strategies for marketing research.” European Journal of Marketing
39 (3/4) (January): 294 -308. doi:10.1108/03090560510581782.
http://www.emeraldinsight.com/journals.htm?issn=0309 –
0566&volume=39&issue=3/4&articleid=1463194&show=html.
Grigoraș, Vlad. 2007. Buhuși -Mocirla. In Nevoi și resurse în comunitățile de romi , ed. Malina Voicu,
51. Bucuresti: Fundatia Soros Romania.
Grimm, Joshua. 2009. “‘Teach the Controversy’. The Relationship Between Sources and Frames in
Reporting the Intellig ent Design Debate .” Science Communication 31 (2): 167 -186.
Hajioff, Steve, and Martin MCKee. 2000. “The health of the Roma people: a review of the published
literature.” Journal of Epidemiology and Community Health 54: 864 –69.
Hale, Henry. 2004. “Explaini ng Ethnicity.” Comparative Political Studies 37: 458 485.
Hameed, Salman. 2008. “Bracing for Islamic creationism .” Science 322: Bracing for Islamic
creationism.
Hammersley, Martyn. 2003. “Conversation Analysis and Discourse Analysis: Methods or Paradigms? ”
Discourse & Society 14 (6): 751 -781. doi:10.1177/09579265030146004.
http://das.sagepub.com/cgi/doi/10.1177/09579265030146004.
——— . 2010. Is Social Measurement Possible, and Is It Necessary? In The SAGE Handbook of
Measurement , ed. Geoffrey Walford, Eric Tucker, and Madhu Viswanathan, 409 -426. London:
SAGE.
Hammond, Kenneth R. 1998. Ecological Validity: Then and Now.
http://www.albany.edu/cpr/brunswik/notes/essay2.html.
Hanquinet, Laurie, and et.al. 2006. Exploratory Comparative Study on Definitions and
Operationalisations of the Notion of Ethnic Disadvantage and Discrimination ISPO -K.U. Leuven,
GERME -ULB, Koning Boudewijnstichting . http://www.kbs -frb.be/ .

73
Hansen, Alan D. 2005. “A Practical Task: Ethnicity as a Resource in Social Interaction.” Research o n
Language & Social Interaction 38 (1) (January): 63 -104. doi:10.1207/s15327973rlsi3801_3.
http://dx.doi.org/10.1207/s15327973rlsi3801_3.
Haslam, S A, and C McGarty. 2001. “A 100 years of certitude? Social psychology, the experimental
method and the manage ment of.” British Journal of Social Psychology 40: 1 -21.
http://www.ingentaconnect.com/content/bpsoc/bjsp/2001/00000040/00000001/art00001.
Hayes, Bernadette, and Vicki Tariq. 2000. “Gender Differences in Scientific Knowledge and Attitudes
toward Science: a Comparative Study of Four Anglo -American Nations.” Public Understanding
of Science 9 433 -447.
Hirschman, Elizabeth C. 1986. “Humanistic Inquiry in Marketing Research: Philosophy, Method, and
Criteria.” Journal of Marketing Research 23 (3) (October): 237 -249.
http://www.jstor.org.libgate.library.nuigalway.ie/stable/info/3151482.
Hoffman, Donna L., and Morris B. Holbrook. 1993. “The Intellectual Structure of Consumer Research:
A Bibliometric Study of Author Cocitations in the First 15 Years of the Journal of Consumer
Research.” The Journal of Consumer Research 19 (4) (November): 505 -517.
http://www.jstor.org/pss/2489437.
Hopper, Joseph. 2001. “The Symbolic Origins of Conflict in Divorce.” Journal of Marriage and Family
63 (2) (May): 430 -445. doi:10.1111/j.174 1-3737.2001.00430.x.
http://doi.wiley.com/10.1111/j.1741 -3737.2001.00430.x.
Houtkoop -Steenstra, Hanneke. 2000. Interaction and the standardized survey interview . Cambridge,
UK: Cambridge University Press.
John Paul II. 1997. “Message to the Pontifical Acad emy of Sciences.” The Quarterly Review of Biology
72 (4): 381 -383.
Kim, Kwang -ki, and Tim Berard. 2009. “Typification in Society and Social Science: The Continuing
Relevance of Schutz’s Social Phenomenology.” Human Studies 32 (3) (December 10): 263 -289.
doi:10.1007/s10746 -009-9120 -6. http://www.springerlink.com/content/108125553h25w370/.
Kitzinger, Jenny. 1994. “The Methodology of Focsu Groups: The Importance of Interaction between
Research Participants.” Sociology of Health & Illness 16 (1): 103 -121.
Knorr -Cetina, K. D. 1982. “Scientific Communities or Transepistemic Arenas of Research? A Critique
of Quasi -Economic Models of Science.” Social Studies of Science 12 (1) (February): 101 -130.
doi:10.1177/030631282012001005. http://sss.sagepub.com/cgi/content/abs tract/12/1/101.
Knorr -Cetina, Karin D. 1981. The Manufacture of Knowledge . Oxford: Pergamon Press.
Krohn, W., and W. van den Daele. 1998. “Science as an agent of change: finalization and
experimental implementation.” Social Science Information 37 (1) (Marc h): 191 -222.
doi:10.1177/053901898037001009. http://ssi.sagepub.com/cgi/content/abstract/37/1/191.
Kwan, Kwong, and Karl Liem. 2000. “The Internal -External Ethnic Identity Measure: Factor -Analytic
Structures Based on a Sample of Chinese Americans’ .” Educa tional and Psychological
Measurement 60 (1): 142 -152.

74
Ladányi, János, and Iván Szelényi. 2001. “The social construction of Roma ethnicity in Bulgaria,
Romania and Hungary during market transition.” Review of Sociology 7: 71 -89.
——— . 2002. “The nature and s ocial determinants of Roma poverty – A cross -national comparison .”
Review of Sociology 8 (2): 75 –96.
Langdridge, Darren, Paschal Sheeran, and Kevin J. Connolly. 2007. “Analyzing Additional Variables in
the Theory of Reasoned Action.” Journal of Applied So cial Psychology 37 (8) (August): 1884 –
1913. doi:10.1111/j.1559 -1816.2007.00242.x. http://www.blackwell –
synergy.com/doi/abs/10.1111/j.1559 -1816.2007.00242.x.
Larionescu, Maria, Cosima Rughinis, and Sorin Radulescu. 1999. Cu ochii minerului. Reforma
mineritu lui in Romania . Bucharest: Gnosis.
Latour, Bruno, and Steve Woolgar. 1986. Laboratory Life. The Construction of Scientific Facts .
Princeton: Princeton University Press.
Laugksch, Rudiger C. 2000. “Scientific literacy: A Conceptual Overview.” Science Educat ion 84 (1)
(January 13): 71 -94. doi:10.1002/(SICI)1098 -237X(200001)84:1<71::AID -SCE6>3.0.CO;2 -C.
http://adsabs.harvard.edu/abs/2000SciEd..84…71L.
Leagans, Jean P. 1964. “A Concept of Needs.” Journal of Cooperative Extension 2: 89 -96.
http://www.joe.org/j oe/1964summer/1964 -2-a4.pdf.
Leahey, Erin. 2008. “Methodological Memes and Mores: Toward a Sociology of Social Research.”
Annual Review of Sociology 34 (July): 33 -53.
http://www.annualreviews.org.libgate.library.nuigalway.ie/doi/abs/10.1146/annurev.soc.34. 0
40507.134731.
Lee, Yew -Jin, and Wolff -Michael Roth. 2004. “Making a Scientist: Discursive ‘Doing’ of Identity and
Self-Presentation During Research Interviews.” Forum Qualitative Sozialforschung / Forum:
Qualitative Social Research 5 (1). http://www.quali tative –
research.net/index.php/fqs/article/view/655/1418.
Lott, Bernice, and Susan Saxon. 2002. “The influence of ethnicity, social class, and context on
judgments about U.S. women.” The Journal of social psychology 142 (4) (August): 481 -99.
doi:10.1080/002 24540209603913. http://www.ncbi.nlm.nih.gov/pubmed/12153124.
MacInnis, Deborah J., and Valerie S. Folkes. 2010. “The Disciplinary Status of Consumer Behavior: A
Sociology of Science Perspective on Key Controversies.” Journal of Consumer Research: An
Interd isciplinary Quarterly 36 (6): 899 -914.
http://ideas.repec.org/a/ucp/jconrs/v36y2010i6p899 -914.html.
Madden, Thomas J., Pamela Scholder Ellen, and Icek Ajzen. 1992. “A Comparison of the Theory of
Planned Behavior and the Theory of Reasoned Action.” Personality and Social Psychology
Bulletin 18 (1): 3 -9. doi:10.1177/0146167292181001.
http://psp.sagepub.com/cgi/doi/10.1177/0146167292181001.
Mateos, Pablo, Alex Singleton, and Paul Longley. 2009. “Uncertainty in the Analysis of Ethnicity
Classifications: Issues of Extent and Aggregation of Ethnic Groups.” Journal of Ethnic and

75
Migration Studies 35 (9) (November): 1437 -1460. doi:10.1080/13691830903125919.
http://dx.doi.org/10.1080/13691830903125919.
Maynard, Douglas W, Jeremy Freese, and Nora Cate Schaeffe r. 2010. “Requests, Blocking Moves,
and Rational (Inter)action in Survey Introductions.” American sociological review 75 (5)
(October 1): 791 -8. doi:10.1177/0003122410379582.
http://asr.sagepub.com/content/75/5/791.full.
Maynard, Douglas W., and Nora Cate Schaeffer. 2000. “Toward a Sociology of Social Scientific
Knowledge: Survey Research and Ethnomethodology’s Asymmetric Alternates.” Social Studies
of Science 30 (3) (June): 323 -370. doi:10.1177/030631200030003001.
http://sss.sagepub.com/cgi/content/abstrac t/30/3/323.
Mays, Vickie M., and et. al. 2003. “Classification of race and ethnicity: Implications for Public Health
.” Annual Review of Public Health 24: 83 –110.
Mazur, A. 2005. “Believers and disbelievers in evolution .” Politics and the Life Sciences 23 (2 ): 55 –
61.
Meadows, Lee, Elizabeth Doster, and David F. Jackson. 2000. “Managing the conflict between
evolution & religion .” The American Biology Teacher 62 (2): 102 -107 .
Miller, Jon D. 1983. “Scientific Literacy: A Conceptual and Empirical Review.” Daedalus 112 (2): 29 –
48.
——— . 1998. “The measurement of civic scientific literacy.” Public Understanding of Science 7 (3):
203-223. doi:10.1088/0963 -6625/7/3/001.
http://pus.sagepub.com/cgi/content/abstract/7/3/203.
——— . 2006a. Civic scientific literacy in Europe and the United States. In Annual meeting of the
World Association for Public Opinion Research . Montreal.
http://www.arcsfoundation.org/Pittsburgh/JMiller.pdf.
——— . 2006b. Civic scientific literacy in Europe and the United States. Montreal.
www.arcs foundation.org/Pittsburgh/JMiller.pdf .
——— . 2007a. Civic Scientific Literacy Across the Life Cycle. In American Association for the
Advancement of Science . San Francisco, California. http://ucll.msu.edu.
——— . 2007b. The Public Understanding of Science in Europe and the United States.
http://ucll.msu.edu/files_ucll.msu.edu/docs/miller -science -europe.doc .
——— . 2007c. The Public Understanding of Science in Europe and the United States. In Annual
meeting of the American Association for the Advancement of Sci ence . San Francisco.
http://ucll.msu.edu/files_ucll.msu.edu/docs/miller -science -europe.doc.
Miller, Jon D., Eugenie C. Scott, and Shinji Okamoto. 2006a. Supporting online material for public
acceptance of evolution. http://www.sciencemag.org/cgi/data/313/5 788/765/DC1/1 .
——— . 2006b. “Public acceptance of evolution .” Science 313: 765 -766 .

76
Mills, C. Wright. 1940. “Situated Actions and Vocabularies of Motive.” American Sociological Review
5 (6) (February): 904 -913. http://www.jstor.org/pss/2084524.
Mitroff, Ian I. 1974. “Norms and Counter -Norms in a Select Group of the Apollo Moon Scientists: A
Case Study of the Ambivalence of Scientists.” American Sociological Review 39 (4) (December):
579-595. http://www.jstor.org/pss/2094423.
Moerman, Michael. 1974. Accomp lishing ethnicity. In Ethnomethodology , ed. Roy Turner, 54 -68.
Baltimore: Penguin Education.
Monaghan, Lee F. 2002. “Vocabularies of motive for illicit steroid use among bodybuilders.” Social
Science & Medicine 55: 695 -708.
http://www.ingentaconnect.com/co ntent/els/02779536/2002/00000055/00000005/art00195.
Moore, Robert J. 2004. “Managing Troubles in Answering Survey Questions: Respondents’ Uses of
Projective Reporting.” Social Psychology Quarterly 67 (1) (March 1): 50 -69.
doi:10.1177/019027250406700106. ht tp://spq.sagepub.com/content/67/1/50.abstract.
Morgan, David L. 1996. “Focus Groups.” Annual Review of Sociology 22 (1) (August): 129 -152.
doi:10.1146/annurev.soc.22.1.129.
http://www.annualreviews.org/doi/abs/10.1146/annurev.soc.22.1.129?journalCode=soc.
Mulkay, Michael, and G. N. Gilbert. 1982. “Accounting for Error: How Scientists Construct their Social
World when they Account for Correct and Incorrect Belief.” Sociology 16 (2) (May): 165 -183.
doi:10.1177/0038038582016002001. http://soc.sagepub.com/cgi/c ontent/abstract/16/2/165.
Mulkay, Michael, and G. Nigel Gilbert. 1983. “Scientists’ Theory Talk.” The Canadian Journal of
Sociology 8 (2) (October): 179 -197.
http://www.jstor.org.libgate.library.nuigalway.ie/stable/info/3340125.
Máté -Toth, András, and Cosi ma Rughiniș. 2011. Spaces and borders. Current research on religion in
Central and Eastern Europe . Berlin: Walter De Gruyter.
Nadelson, Louis S., and Sherry A. Southerland. 2010. “Examining the interaction of acceptance and
understanding: How does the rela tionship change with a focus on macroevolution? .”
Evolution: Education and Outreach 3 82 -88.
National Science Board. 2008. “Science and engineering indicators 2008. Chapter 7.” Science and
technology: Public attitudes and understanding .
http://www.nsf.go v/statistics/seind08/pdf/c07.pdf .
Pardo, Rafael, and Félix Calvo. 2004. “The Cognitive Dimension of Public Perceptions of Science:
Methodological Issues.” Public Understanding of Science 13 (3) (July 1): 203 -227.
doi:10.1177/0963662504045002. http://pus.s agepub.com/cgi/content/abstract/13/3/203.
Pardo, Rafael, Cees Midden, and Jon D Miller. 2002. “Attitudes toward biotechnology in the
European Union.” Journal of Biotechnology 98: 9 –24.
Pellebon, D. A. 2000. “Influences of Ethnicity, Interracial Climate, an d Racial Majority in School on
Adolescent Ethnic Identity.” Children & Schools 22 (1) (January 1): 9 -20. doi:10.1093/cs/22.1.9.
http://cs.oxfordjournals.org/content/22/1/9.abstract.

77
Penders, B., J. M. A. Verbakel, and A. Nelis. 2009. “The Social Study of C orporate Science: A Research
Manifesto.” Bulletin of Science, Technology & Society 29 (6) (November): 439 -446.
doi:10.1177/0270467609349047. http://bst.sagepub.com/cgi/content/abstract/29/6/439.
Phinney, Jean S., and Anthony D. Ong. 2007. “Conceptualizatio n and measurement of ethnic identity:
Current status and future directions.” Journal of Counseling Psychology 54 (3): 271 -281.
Poling, Devereaux A., and E. Margaret Evans. 2004. “Religious belief, scientific expertise, and folk
ecology .” Journal of Cognit ion and Culture 4: 485 -524 .
Pop, Lucian, and Cosima Rughiniș. 2000. Capital social și dezvoltare comunitară – Studiu de caz. In
Diagnoza problemelor sociale comunitare , ed. Elena Zamfir and Marian Preda, 114 -152.
Bucharest: Editura Expert.
Potter, Jonatha n. 2003a. “Discursive Psychology: Between Method and Paradigm.” Discourse &
Society 14 (6): 783 -794. doi:10.1177/09579265030146005.
http://das.sagepub.com/cgi/doi/10.1177/09579265030146005.
——— . 2003b. “Review EssayStudying the Standardized Survey as Inter action.” Qualitative Research
3 (2) (August 1): 269 -278. doi:10.1177/14687941030032007. http://qrj.sagepub.com.
Preda, Marian. 1993. Conflictul din localitatea Mihail Kogălniceanu. In Țiganii – Între ignorare și
îngrijorare , ed. Catalin Zamfir and Elena Zamfir, 178. Bucharest: Alternative.
Prieto -Flores, Òscar. 2009. “ ’Does the canonical theory of assimilation expl ain the Roma case? Some
evidence from Central and Eastern Europe .” Ethnic and Racial Studies 32 (8): 1387 — 1405 .
Razzaque, Mohammed Abdur. 1998. Scientific Method, Marketing Theory Development and
Academic vs Practitioner Orientation: A Review. Journal of Marketing .
http://business.nmsu.edu/~mhyman/M670_Articles/Razzaque_JMTP_1998.pdf.
Raț, Cristina. 2005. “Romanian Roma, State Transfers, and Poverty.” International Journal of
Sociology 35 (3): 85 -115. http://academos.ro/sites/default/files/biblio -docs/243/p –
romanian_roma_rat.pdf.
Roth, Wolff -Michael, and Stuart Lee. 2002. “Scientific li teracy as collective praxis.” Public
Understanding of Science 11 (1): 33 -56. doi:10.1088/0963 -6625/11/1/302.
http://pus.sagepub.com/cgi/content/abstract/11/1/33.
Roulston, Kathryn. 2006. “Close encounters of the ‘CA’ kind: a review of literature analysing talk in
research interviews.” Qualitative Research 6 (4) (November 1): 515 -534.
doi:10.1177/1468794106068021. http://qrj.sagepub.com/cgi/content/abstract/6/4/515.
Rughinis, Cosima, and Roxana Toader. 2010. “Education and scientific knowledge in European
societies. Exploring measurement issues in general population surveys. .” Studia Universitatis
Babes -Bolyai – Sociology (1): 175 -202.
http://www.unibuc.ro/prof/rughinis_a_c/docs/res/2011ianStudiaSociologia_2010 -1.pdf .
Rughiniș, Cosima. 2001. Comment s’organ isant pour un meilleur environnement dans un quartier
d’exclus a Bucarest. In Femmes dans les crises urbaines , ed. François Hainard and Christine
Verschuur, 231 -270. Paris: Karthala.

78
——— . 2002a. “Experiences of Beneficiary Participation in Development Proj ects for Roma People.”
Romanian Journal of Sociology (1-2): 116 -137.
——— . 2002b. “Valori europene în relațiile intime. Studiu comparativ.” Sociologie Românească (1-
2): 38 -75.
——— . 2002c. “Sunt problemele cu care se confruntă romii de natură etnică sau soci ală?” Revista de
Asistență Socială (4-5): 117 -121.
——— . 2004a. Cunoaștere incomodă: Intervenții sociale în comunități defavorizate în România anilor
2000 . Bucharest: Printech.
——— . 2004b. Social housing and Roma residents in Romania. Budapest, January.
http://pdc.ceu.hu/archive/00001919/01/rughinis.pdf.
——— . 2005. Genre et citoyenneté : accès à l’électricité, droit au bail et démission de l’Etat. In
Mouvements de quartier et environnements urbains -La prise de pouvoir des femmes dans les
pays du Sud et de l’ Est, ed. François Hainard and Christine Verschuur, 187 -215. Paris: Karthala.
——— . 2006. Religiozitate și scepticism religios. In Viața socială în România urbană , ed. Dumitru
Sandu, 123 -133. Iași: Polirom.
——— . 2007a. Explicația sociologică . Iași: Polirom.
——— . 2007b. “„Filosofia medie”. Studiul concepțiilor despre fericire și iubire prin anchete
sociologice cantitative”.” Sociologie Românească (4): 65 -81.
——— . 2010. “The forest behind the bar charts: bridging quantitative and qualitative research on
Roma/Ti gani in contemporary Romania.” Patterns of Prejudice 44 (4) (September): 337 -367.
doi:10.1080/0031322X.2010.510716.
http://www.informaworld.com/smpp/content~content=a926971107~db=all~jumptype=rss.
——— . 2011a. “Quantitative tales of ethnic differentiation: measuring and using Roma/Gypsy
ethnicity in statistical analyses.” Ethnic and Racial Studies 34 (4) (April): 594 -619.
doi:10.1080/01419870.2010.514055.
http://www.informaworld.com/smpp/content~db=all~content=a928117662.
——— . 2011b. “A Lucky Answer to a Fai r Question: Conceptual, Methodological, and Moral
Implications of Including Items on Human Evolution in Scientific Literacy Surveys.” Science
Communication (May 19). doi:10.1177/1075547011408927.
http://scx.sagepub.com/cgi/content/abstract/1075547011408927 v1.
Rughiniș, Cosima, and Iulia Răuțu. 2009. “Of priests and politics. Measuring separation of Church and
State in present -day Romania.” Studia Sociologia (2): 57 -77.
http://www.studia.ubbcluj.ro/download/pdf/466.pdf.
Rughiniș, Cosima, and Roxana Toader. 2010. “Education and Scientific Knowledge in European
Societies. Exploring Measurement Issues in General Population Surveys.” Studia Sociologia (1):
175-202. http://www.studia.ubbcluj.ro/download/pdf/521.pdf.

79
Sacks, Harvey. 1995. Harvey Sacks: lectures on conve rsation . Ed. Gail Jefferson. Malden: John Wiley
& Sons. http://books.google.com/books?hl=en&lr=&id=6b0RimyejTkC&pgis=1.
Salat, Levente, and Valér Veres. 2009. Ethnic minorities and local public administration in Romania .
http://adatbank.transindex.ro/vend eg/htmlk/pdf4098.pdf .
Sandu, Dumitru. 2005. Comunitățile de romi din România. O hartă a sărăciei comunitare prin
sondajul PROROMI . http://www.anr.gov.ro/site/Biblioteca.html .
Saris, Willem E., and Irmtraud N. Gallhofer. 2007. Design, evaluation, and an alysis of questionnaires
for survey research Hoboken, New Jersey: Wiley -Interscience .
Schaeffer, Nora Cate, and Stanley Presser. 2003. “The Science of Asking Questions.” Annual Review
of Sociology 29 (1): 65 -88. doi:10.1146/annurev.soc.29.110702.110112.
http://arjournals.annualreviews.org/doi/abs/10.1146/annurev.soc.29.110702.110112.
Schegloff, E A. 1999. “Naivete vs Sophistication or Discipline vs Self -Indulgence: A Rejoinder to Billig.”
Discourse & Society 10 (4): 577 -582. doi:10.1177/0957926599010004008 .
http://das.sagepub.com/cgi/doi/10.1177/0957926599010004008.
Schegloff, Emanuel A. 1999. “Schegloff’s Texts' as `Billig's Data': A Critical Reply.” Discourse & Society
10 (4): 558 -572. doi:10.1177/0957926599010004006.
http://das.sagepub.com/cgi/doi/10.117 7/0957926599010004006.
Schilling -Estes, Natalie. 2004. “Constructing ethnicity in interaction.” Journal of Sociolinguistics 8 (2)
(June 8): 163 -195. doi:10.1111/j.1467 -9841.2004.00257.x.
http://doi.wiley.com/10.1111/j.1467 -9841.2004.00257.x.
Schober, Micha el F, and Frederick G Conrad. 1997. “Does Conversational Interviewing Reduce Survey
Measurement Error?” Public Opinion Quarterly 61 (4): 576 -602. doi:10.1086/297818.
http://poq.oupjournals.org/cgi/doi/10.1086/297818.
Schütz, Alfred. 1953. “Common -sense and scientific interpretation of human action.” Philosophy and
Phenomenological Research 14 (1): 1 -38.
Scott, Eugenie C. 1997. “Antievolution and creationism in the United States.” Annual Review of
Anthropology 26: 263 -289.
Serban, Monica. 1998. “Aspecte ale atașamentului față de comunitatea locală. Studiu comparativ
asupra a două comunități de rromi căldărari bogați.” Revista de Cercetări Sociale (3-4): 151 –
179.
Shapin, Steven. 1995. “Here and Everywhere – Sociology of Scientific Knowledge.” Annual Review of
Sociology 21: 289 -321. http://dash.harvard.edu/handle/1/3219885.
Sheppard, Blair H, Jon Hartwick, and Paul R Warshaw. 1988. “The Theory of Reasoned Action: A
Meta -Analysis of Past Research with Recommendations for Modifications and Future
Research.” Journa l of Consumer Research 15 (3): 325 -343. doi:10.1086/209170.
http://www.journals.uchicago.edu/doi/abs/10.1086/209170.

80
Shugan, Steven M. 2004. “Editorial: Consulting, Research, and Consulting Research.” Marketing
Science 23 (2) (March): 173 -179. doi:10.1287/ mksc.1040.0078.
http://mktsci.journal.informs.org/cgi/content/abstract/23/2/173.
Simon, Patrick. 2011. “Collecting ethnic statistics in Europe: a review.” Ethnic and Racial Studies
(September 26): 1 -26. doi:10.1080/01419870.2011.607507.
http://dx.doi.org/1 0.1080/01419870.2011.607507.
Singh, Swaran. 1997. “Ethnicity in psychiatric epidemiology: need for precision.” The British Journal
of Psychiatry 171: 305 -308.
Smith, Dorothy E. 1974. Theorizing as Ideology. In Ethnomethodology , ed. Roy Turner, 41 -44.
Balti more: Penguin Education.
Smith, George Davey. 2000. “Learning to Live With Complexity: Ethnicity, Socioeconomic Position,
and Health in Britain and the United States.” American Journal of Public Health 90 (11): 1694 –
1698.
Smith, Tom W. 2008. “An Experiment al Comparison of Methods of Measuring Ethnicity.” Field
Methods 20 (2) (March 10): 171 -178. doi:10.1177/1525822X08316644.
http://fmx.sagepub.com/cgi/content/abstract/20/2/171.
Steinberg, Laurence, and Anne Fletcher. 1998. Data Analytic Strategies in Resear ch on Ethnic
Minority Youth . In Studying Minority Adolescents. Conceptual, Methodological and Theoretical
Issues , ed. Vonnie McLoyd and Laurence Steinberg, 279 -294. Mahwah, New Jersey: Lawrence
Erlbaum Associates .
Stephan, Cookie White, and Walter G. St ephan. 2000. “The measurement of racial and ethnic
identity.” International Journal of Intercultural Relations 24 (5) (September): 541 -552.
doi:10.1016/S0147 -1767(00)00016 -X. http://dx.doi.org/10.1016/S0147 -1767(00)00016 -X.
Stern, Barbara B. 2004. “The Imp ortance of Being Ernest: Commemorating Dichter’s Contribution to
Advertising Research.” Journal of Advertising Research 44 (02) (June): 165 -169.
http://journals.cambridge.org/abstract_S0021849904040127.
Stewart, Michael. 2008. Approaches to Roma and Gypsie s from within social anthropology with
particular reference to the Anglo -Saxon traditions. In Az etnicitás empirikus kutatásának
dilemmái . Research Institute of Ethnic and National Minorities, Hungarian Academy of
Sciences.
http://www.mtaki.hu/docs/080611_ 12/teljes/michael_stewart_anthropology_and_the_roma.
pdf.
Suchman, Lucy, and Brigitte Jordan. 1990. “Interactional troubles in face -to-face survey interviews.”
Journal of the American Statistical Association 85 (409): 232 -241.
http://eprints.lancs.ac.uk/270 37/.
Superfine, Benjamin Michael. 2009. “The evolving role of the courts in educational policy: The
tension between judicial, scientific, and democratic decision making in Kitzmiller v. Dover. .”
American Educational Research Journal 46 (4): 898 -923 .

81
Szelényi, Iván. 2001. Poverty under post -communist capitalism – the effects of class and ethnicity in a
cross -national comparison. In “Unity and Diversity” , 51. Bruges.
ftp.cordis.europa.eu/pub/belgium/docs/szelenyi_paper.pdf .
Székeli, Mária, György Csepeli, and Antal Örkény. 2003. “Ambitious education: the role of family,
school and friends in the development of successful Romany life courses.” Romani Studies 13
(1): 53 -72.
TNS OPINION & SOCIAL. 2005a. Eurobarometer 63.1 Science and technology, social values, and
services of general interest. Basic English questionnaire Brussels .
——— . 2005b. Eurobarometer 63.1 Science and technology, social values, and services of general
interest . Brusseles.
Thagard, Paul, and Scott Findlay. 2009. “Getting to Darwin: Obstac les to accepting evolution by
natural selection.” Science & Education 19: 625 -636.
Threlfall, Denise K. 1999. “Using focus groups as a consumer research tool.” Journal of Marketing
Practice: Applied Marketing Science 5 (4): 102 -105.
http://www.ingentaconne ct.com/content/mcb/155/1999/00000005/00000004/art00001.
Trafimow, D. 2009. “The Theory of Reasoned Action: A Case Study of Falsification in Psychology.”
Theory & Psychology 19 (4) (August 14): 501 -518. doi:10.1177/0959354309336319.
http://tap.sagepub.com/c gi/content/abstract/19/4/501.
UNDP – United Nations Development Program. 2003. Avoiding the dependency trap.
http://roma.undp.sk/ .
Viswanathan, Madhu. 2010. Understanding the Intangibles of Measurement in the Social Sciences. In
The SAGE Handbook of Meas urement , ed. Geoffrey Walford, Eric Tucker, and Madhu
Viswanathan, 285 -311. London: SAGE.
Vlăsceanu, Laz ăr. 2011. Cunoaștere și cunoaștere științifică. In Sociologie , ed. Laz ăr Vlăsceanu, 536 –
565. Iasi: Polirom.
Vlăsceanu, Lazar, Adrian Dușa, and Cosima Ru ghiniș. 2010. Publicul și știința. Știință și societate.
Interese și percepții ale publicului privind cercetarea științifică. și rezultatele cercetării .
http://stisoc.ro/psip.html .
Vlăsceanu, Lazăr. 2011. Cunoaștere și cunoaștere științifică. In Sociologi e, ed. Lazăr Vlăsceanu, 536 –
565. Iași: Polirom.
Voiculescu, Cerasela. 2002. “Construcții identitare la rromii din Sângerogiu de Mureș.” Sociologie
Românească (1-2): 100 -125.
Wagner, Gertrud. 1938. “Market Research: A Critical Study.” The Review of Economic Studies 5 (2)
(December): 132 -138.
Wagner, Wolfgang. 2007. “Vernacular science knowledge: its role in everyday life communication.”
Public Understanding of Science 16: 7 -22.

82
Warfield Rawls, Anne. 1987. “The Interaction Order Sui Generis: Goffman’s Contrib ution to Social
Theory.” Sociological Theory 5 (2): 136 -149.
——— . 2011. “Wittgenstein, Durkheim, Garfinkel and Winch: Constitutive Orders of Sensemaking.”
Journal for the Theory of Social Behaviour 41 (4): 396 -418.
Warfield Rawls, Anne, and Gary David. 200 6. “Accountably Other: Trust, Reciprocity and Exclusion in
a Context of Situated Practice.” Human Studies 28: 469 -497.
Wilkinson, S. 2011. “Constructing ethnicity statistics in talk -in-interaction: Producing the ‘White
European’.” Discourse & Society 22 (3 ) (April 28): 343 -361. doi:10.1177/0957926510395446.
http://das.sagepub.com/cgi/content/abstract/22/3/343.
Wimmer, Andreas. 2008. “The Making and Unmaking of Ethnic Boundaries: A Multilevel Process
Theory .” American Journal of Sociology 113: 970 –1022.
Wink, Paul. 1997. “Beyond Ethnic Differences: Contextualizing the Influence of Ethnicity on
Individualism and Collectivism.” Journal of Social Issues 53 (2) (April 9): 329 -349.
doi:10.1111/j.1540 -4560.1997.tb02446.x. http://doi.wiley.com/10.1111/j.1540 –
4560.1 997.tb02446.x.
Winker, Margareta. 2006. “Race and Ethnicity in Medical Research: Requirements Meet Reality.”
Journal of Law, Medicine & Ethics Fall 2006: 520 -25.
Woolgar, S.W. 1976. “Writing an Intellectual History of Scientific Development: The Use of Dis covery
Accounts.” Social Studies of Science 6 (3-4) (September): 395 -422.
doi:10.1177/030631277600600306. http://sss.sagepub.com.
Wynne, Brian. 1992. “Public understanding of science research: new horizons or hall of mirrors?”
Public Understanding of Scien ce 1 (1): 37 -43. doi:10.1088/0963 -6625/1/1/008.
http://pus.sagepub.com/cgi/pdf_extract/1/1/37.
Yearley, Steven. 1988. “Settling Accounts: Action, Accounts and Sociological Explanation.” British
Journal of Sociology 39 (4) (December): 578 -599.
http://www.js tor.org.libgate.library.nuigalway.ie/stable/info/590502.
Zamfir , Cătălin, and Preda Marian. 2002. Romii în România . Ed. Zamfir Catalin and Preda Marian.
Bucharest : Editura Expert .
Zamfir, Cătălin, and Cosima Rughiniș. 2000. Mecanismele sociale ale dezvol tării sociale comunitare –
Studiu de caz al comunității Zăbrăuți, București. In Diagnoza problemelor sociale comunitare ,
ed. Elena Zamfir and Marian Preda, 49 -90. Bucharest: Editura Expert.
Zamfir, Elena, and Catalin Zamfir (Eds) . 1993. Țiganii – Între ignorare și îngrijorare . Bucharest :
Alternative.