Primary Care Nurse Practitioner Wage Differences by [625547]

Primary Care Nurse Practitioner Wage Differences by
Employment Setting
Yin Li, Ph.D.a,1,*, George “Mark” Holmes, Ph.D.b,2,3, Erin P. Fraher, Ph.D. MPPc,4,
Barbara A. Mark, Ph.D., RN, FAANd,5, Cheryl B. Jones, Ph.D., RN, FAANe,6
aResearch Assistant Professor, Nell Hodgson Woodruff School of Nursing, Emory University, Atlanta, GA 30322
bAssociate Professor, Department of Health Policy and Management , Gillings School of Global Public Health , Director of Cecil G. Sheps Center
for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill 27599-7411
cAssistant Professor, Department of Family Medicine, School of Medicine, Director of Program on Health and Workforce Research and Policy,
Cecil G. Sheps Center for Health Services Research, CB#7590, University of North Carolina at Chapel Hill, Chapel Hill 27599-7590
dSarah Frances Russell Distinguished Professor, School of Nursing, CB#7460, University of North Carolina at Chapel Hill, Chapel Hill,
NC 27599-7460
eProfessor Director, Hillman Scholars Program in Nursing Innovation, School of Nursing, CB#7460, University of North Carolina
at Chapel Hill, Chapel Hill, NC 27599-7460
ABSTRACT
Background: Previous studies reported that primary care nurse practitioners working in
p r i m a r yc a r es e t t i n g sm a ye a r nl e s st h a nt h o s ew o r k i n gi ns p e c i a l t yc a r es e t t i n g s .H o w -ever, few studies have examined why such wage difference exists.Purpose: This study used human capital theory to determine the degree to which the
wage differences between PCNPs working in primary care versus specialty care settingsis driven by the differences in PCNPs’ characteristics. Feasible generalized least squaresregression was used to examine the wage differences for PCNPs working in primarycare and specialty care settings.Methods: A cross-sectional, secondary data analysis was conducted using the restricted
file of 2012 National Sample Survey of Nurse Practitioners.
Findings: Oaxaca-Blinder decomposition technique was used to explore the factors con-
tributing to wage differences.The results suggested that hourly wages of PCNPs working in primary care settingswere, on average, 7.1% lower than PCNPs working in specialty care settings, holdingPCNPs’ socio-demographic, human capital, and employment characteristics con-stant. Approximately 4% of this wage difference was explained by PCNPs’ character-istics; but 96% of these differences were due to unexplained factors.Discussion: A large, unexplained wage difference exists between PCNPs working in
primary care and specialty care settings.ARTICLE INFO
Article history:Received 13 January 2018Received in revised form 18 June2018Accepted 27 June 2018Available online July 17, 2018.
Keywords:
Wage DifferenceNurse PractitionerPrimary CareSpecialty Care
*Corresponding author: Yin Li, Ph.D. Research Assistant Professor, Nell Hodgson Woodruff School of Nursing, Emory University,
Atlanta, GA 30322.
E-mail address: [anonimizat] (Y. Li).
0029-6554/$ -see front matter /C2112018 Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.outlook.2018.06.009
1Phone: (404)-727-8445
2Voice: (919) 807-1532
3Fax: (919) 966-1634
4Phone: (919) 966-5012
5Phone: (919) 843-6209
6Phone: (919) 966-5681Available online at www.sciencedirect.com
Nurs Outlook 66 (2018) 528 /C0538www.nursingoutlook.org

Cite this article: Li, Y., Holmes, G.M., Fraher, E.P., Mark, B.A., & Jones, C.B. (2018, November/December).
Primary Care Nurse Practitioner Wage Differences by Employment Setting. Nursing Outlook, 66(6),528/C0538. https://doi.org/10.1016/j.outlook.2018.06.009 .
Introduction
The number of primary care nurse practitioners (PCNPs) /C0
nurse practitioners (NPs) who are certified in a primarycare specialty /C0has increased by 30% in the past decade
in the United States because of the growing role they fill inproviding access to primary care services ( Josiah Macy Jr.
Foundation, 2016; U. S. Department of Health and HumanServices, 2016 ). Not all certified PCNPs practice in primary
care settings e.g. community health centers, physicianoffices, school health, home health, ambulatory care clin-ics, and hospital outpatient departments. Instead, theyoften practice in specialty care and “other” settings,
including hospitals, emergency departments, mental
health clinics, urgent care, and long-term care ( Chattopad-
hyay, Zangaro, & White, 2015; Keough, Stevenson, Marti-novich, Young, & Tanabe, 2011; Spetz, Fraher, Li, & Bates,2015). In 2011, 75% of the U.S. supply of NPs was certified
in a primary care specialty area, but only 49.0% of NPspracticed in a primary care setting ( Spetz et al., 2015 ).
One reason that PCNPs may not work in primary care
settings is that they earn lower wages working in pri-mary care than in specialty care settings ( Bodenheimer
& Bauer, 2016; Coomer, 2013; Jones & Gates, 2004;
McGregory, Niederjohn, & Peoples, 2009; Petterson, Phil-
lips, Bazemore, Burke, & Koinis, 2013; Schumacher &Hirsch, 1997; Walani, 2013 ). For NPs, Goolsby (2006)
used the 2004 data file of the National Nurse Practitioner
Sample Survey (NNPSS) conducted by the American
Academy of Nurse Practitioners and reported that,regardless of certification, the average hourly wage ofNPs working in primary care settings was $36.51, con-siderably lower than the $39.59 average wage for thoseworking in specialty care settings. In a later study,Goolsby (2009) examined data from the 2008 NNPSS file
and found that the average annual wage of NPs working
in primary care settings was $84,771, compared to$92,575 for NPs working in specialty care settings.
Although previous studies have reported that wage
differences exist for both RNs and NPs in primary careand specialty care settings, researchers have notexamined whether such wage differences exist forPCNPs specifically or examined why wages differ forPCNPs. Lower wages for NPs (both PCNPs and NPs cer-tified in other specialties) working in primary care set-tings have been associated with lower job satisfaction
(Pasaron, 2013 ), a greater likelihood of turnover
(De Milt, Fitzpatrick, & Sister Rita, 2011 ), or NP gradu-
ates seeking employment outside of primary care set-tings ( Budd, Wolf, & Haas, 2015 ).
This study investigated whether wage differences exist
for PCNPs who work in primary care versus specialtycare settings and examined the factors contributing tothese wage differences. Although such wage differencesbetween primary care and specialty physicians have been
the subject of previous studies and have been shown toaffect the recruitment and retention of physicians in pri-mary care specialties, no such study has examined poten-tial wage differences for nurse practitioners in differentsettings ( Heisler & Sarata, 2011; Wilder et al., 2010 ). Devel-
oping a better understanding of the sources of wage dif-
ferences is important to inf orm human resource policies
and payment incentives that affect the attractiveness ofprimary care relative to other specialties.
Conceptual Approach
This study is grounded in human capital theory, a the-ory that has been widely used to examine nurses’wages and the wage differences between certain nurs-ing groups. Human capital refers to an individual’s per-sonal characteristics (i.e., innate abilities andintelligence, acquired knowledge and education, job
skills and abilities, and work experiences) or behaviors
(e.g., job mobility) that affect their productivity and per-formance ( Becker, 1962, 2009; Currie & Madrian, 1999;
Kiker, 1966; Mincer, 1958; Schultz, 1961; Willis, 1985 ).
According to human capital theory, individuals makefinancial (e.g., tuition payments) and non-financial (e.g.,time) investments to acquire human capital, with theexpectation that they will receive a return on invest-ment at some future point ( Ehrenberg & Smith, 2010;
Schultz, 1961 ). Given that individuals’ investments in
human capital provide them knowledge, skills, and
abilities that allow them to do new and/or different
work, an important assumption of human capital the-ory is that greater human capital investments result inhigher productivity ( Mincer, 1958; Schultz, 1961 ).
Because productivity per se is difficult to observedirectly and measure, labor economists typically con-sider individuals’ wages as a proxy for productivity.Human capital theory thus posits that as human capitalincreases, wages also increase ( Becker, 2009 ).
Further, human capital theory acknowledges that, in
addition to human capital, individuals’ personal attrib-
utes can also affect their wages, including socio-demo-
graphic (e.g., gender, race, and marital status) andemployment characteristics (e.g., work setting, posi-tion, geographic region, or full-time or part-timeemployment) ( Becker, 2009, 2010 ). Thus, researchers
typically model individuals’ wages as a function oftheir human capital and personal attributes.
It is possible, however, that wage differences cannot be
fully explained by individuals’ human capital and per-sonal attributes; instead wage differences may be attrib-utable to other measurable and unmeasurable factors.
To identify sources of wage differences, Oaxaca (1973)Nurs Outlook 66 (2018) 528 /C0538 529

developed a widely used approach to decompose wage
differences into two main effects /C0endowment effects
and coefficient effects. Endowment effects are“explained factors” that reflect the contributions of indi-viduals’ human capital and personal attributes. Forexample, previous studies s uggested that nurses work-
ing in specialty care settings were paid more relative tonurses in primary care settings, because they havehigher cognitive ability (as measured by Armed ForceQualifying Test Scores) and greater years of experience
(e.g., organizational and occupational tenure) ( Schu-
macher & Hirsch, 1997 ). Coefficient effects are
“unexplained factors” that re flect factors not observed in
wage modeling. For instance, researchers in previousstudies speculated that nurses working in specialty caresettings were paid more to compensate for unpleasantworking conditions (e.g., irregular or late shifts, higherstress levels, and greater job hazards) or to cover theirpension and insurance, but these factors were not exam-ined in wage modeling due to data limitations ( Lehrer,
White, & Young, 1991; Schumacher & Hirsch, 1997 ).
Although there have been many applications of human
capital theory for RN wage differences ( Coomer, 2013;
Jones & Gates, 2004; Kalist, 2002; McGregory Jr, 2013; Spetz,2002; Walani, 2013 ), there have been none of NP differen-
ces. While Goolsby (2006, 2009 ) descriptively compared
the average NP wages between primary care and specialtycare settings, the analysis did not model NPs’ wages as afunction of human capital and personal attributes. Thus,little is known about whether the wage differencesbetween NPs working in primary care and specialty caresettings exist after controllin g for those factors. Although
researchers have speculated about reasons for wage dif-
ferences, no study has parsed differences into endowment
and coefficient effects. Guided by human capital theory,this study sought to examine wage differences for PCNPsemployed in primary care and specialty care settings; andexplore the extent to which endowment and coefficienteffects contribute to wage differences for PCNPs workingin primary care and specialty care settings. A deeperu n d e r s t a n d i n go ft h es o u r c eo fa n yw a g ed i f f e r e n c e sm a yenable policy makers and employers to make adjustmentsn e c e s s a r yt os u p p o r tam o r ee f fi c i e n tm a r k e t .
Methods
Design
This study used a secondary analysis of cross-sectionaldata, extracted from the restr icted file of the Department
of Health and Human Services, Health Resources andServices Administration (HRSA) National Sample Survey of
Nurse Practitioners (NSSNP) 2012. Data gathered in this
survey were obtained from a stratified sample of 22,000
actively licensed or certified U.S. NPs.
1The surveya c h i e v e dar e s p o n s er a t eo fa p p r o x i m a t e l y6 0 % ,f o ra
final sample of 12,923 NPs. Data gathered in this surveyincluded NP demographic, socioeconomic, educational,certification, employment, and practice characteristics(U.S. Department of Health and Human Services, 2014 ).
Sample
The sample for this study included NSSNP respondents
who met the following inclusion criteria: 1) practice
approval obtained through a state board of nursing; 2) cer-tified as a PCNP, in one of the following primary care spe-cialty areas /C0adult, family, geriatric, pediatric, or
women’s health
2; 3) worked for pay as an NP; and 4) prac-
ticed in a primary care or spe cialty care setting in their
principal NP job. Based on the categorization of NPs’ work-ing setting in NSSNP 2012, a primary care setting in this
study was defined as private physician office, private NPoffice, nurse managed clinic, retail clinic, ambulatory care,federal clinic, home care agency, community clinic, correc-
tional facility, health department, rural health clinic,
health maintenance organization, employee health, andschool health; specialty care settings were defined as mental
health center, urgent care clinic, hospital inpatient unit,hospital outpatient, hospita l emergency department, other
hospital settings, federal hospi tal (Military, Veteran Affairs,
India Health Services), hospita ls u r g i c a ls e t t i n g s ,h o s p i t a l
other clinic settings, and hospital non-clinical settings.
3
After applying these initial inclusion criteria, the sam-
ple was 8,175. Of these, 1,829 NPs responded “notapplicable” or “unknown” on relevant study variables.
Observations for these NPs were dropped or imputed for
analysis, leaving 7,532 observations as the final sampleused in this study.
4This PCNP final sample consisted of
5,793 (76.9%) observations in primary care settings and1,739 (23.1%) observations in specialty care settings.
Wage Models and Variables
Based on human capital theory and Mincer’s wagemodel ( Mincer, 1974 ), PCNP wages were modeled as a
function of PCNPs’ socio-demographic, human capital,
and employment characteristics of the general form:
1The NSSNP only surveyed NPs; employers were not included
as part of the survey.2In the final sample, 792 (10.5%) out of the 7,532 observations
hold more than one certification. Based on human capital theory,an NP holding more certifications may earn higher wages. There-fore, a binary variable of whether a PCNP holding multiple certifi-
cations (0=only one certification; and 1=more than one
certification) was created and included in a preliminary wagemodeling analysis. Because this variable was not statistically sig-
nificant associated with wages, however, it was not included in
our final wage modeling.
3These types of settings were not mutually exclusive, but
they were directly derived from the NSSNP 2012. The categoriza-
tion of primary care and specialty care settings were based on the
study of Spetz, Fraher, Li, and Bates (2015) .
4Data were dropped if missing data were less than or equal to
5% and imputed if greater than 5% of the 8,175 observations.
Information on the management of missing data can be obtained
from the corresponding author.530 Nurs Outlook 66 (2018) 528 /C0538

Wage ¼fS ;H;E;e ðȚ ð 1Ț
Where
Wage is the hourly PCNP wage, which was calculated
by using the pre-tax annual earnings from the princi-pal NP position (include overtime, on call earning,and bonuses) in 2011 divided by the product of thenumber of hours worked in the principal NP positionduring a typical week times 52 weeks;
Sis a vector of PCNP socio-demographic characteris-
tics, including gender (male or female), race (whiteor non-white), ethnicity (Hispanic or non-Hispanic),
and marital status (never married, married, or sepa-
rated, divorced, widowed, and other);
His a vector of PCNP human capital characteristics,
including NP education preparation (certificate pro-gram, Master’s degree, Post Master’s degree, or Doc-tor of Nursing Practice degree and other), years ofNP experience (2011 minus the year of completingthe initial NP education program), and intent toleave their position (no plan to leave, leave in 2012,leave in 1-2 years, or undecided or unknown);
Eis a vector of PCNP employment characteristics,
including type of their employment setting (pri-mary or specialty care), type of their position (NPposition in a clinic, other position that requiringNP credentialing, other position that not requiringNP credentialing, or other nursing position
5), spe-
cialty of their practice clinic (primary or specialtycare
6), degree of their practice independence (no
physician on site, worked with a physician as ateam member, supervised by a physician or otherrelationships with a physician), MetropolitanStatistical Areas [MSA] region of employment
(urban [ >= 50,000 population], large rural [10,000-
49,000 population], small rural [2,500-9,999 popula-tion], or isolated [ <2,500 population]), census
region of employment (New England, MiddleAtlantic, East North Central, West North Central,South Atlantic, East South Central, West SouthCentral, Mountain or Pacific), salary method (paidby annual salary, by the hour, by percentage ofbilling, or others), and whether holding more than
o n eR No rN Pj o b ;a n dei sr a n d o me r r o r .
Analysis
The following analytic model was estimated to exam-
ine PCNP wages and determine if there were wage dif-ferences for PCNPs in primary care and specialty caresettings:
ln WageðȚ ¼ațb
1X1țb2X2țb3X3ț…țɛ ð2Ț
where ln (wage) is the natural logarithm of hourly PCNP
wages ;ais the constant; bi(i=1 ,2,3…)is the coefficient of
each variable X i(i=1 ,2,3…);X i(i=1 ,2,3…)represents the
socio-demographic, human capital, and employmentvariables; and eis random error.
Feasible generalized least squares (FGLS) regres-
sion was used to estimate the model to overcomeproblems with heteroskedasticity in the data.Because statistically significant results were found, aChow-test was used to further examine whether the
“setting” variable had different effects on PCNPs in
primary care and specialty care settings. To estimatethe Chow test, two separate models for PCNPs /C0one
for those in primary care and another for those inspecialty care settings /C0were estimated using FGLS
in the same format as model (2).T h eC h o wt e s t
determined that there were structural differencesbetween these two models ( F= 2.02, p<0.01). This
finding suggested that use of the wage decomposi-tion approach was appropriate. Using these twomodels, we also predicted the average weighted level
wage differences between PCNPs in primary care and
specialty care settings.
Then we decomposed PCNP wages to determine the
origin of differences using the technique developed byOaxaca (1973) , and later refined by others ( Cotton,
1988; Holtmann & Idson, 1993; Reimers, 1983 ). We fol-
lowed Cotton’s model (1988) and used weighted X
iand
biterms because we used FGLS regression models to
examine PCNP wages, which predicted weighted Xi
andbi:
EY ipcw/C0/C1
/C0EY ispw/C0/C1
¼0:5SbipcwțbispwðȚ Xipcw/C0Xispw/C0/C1
ț0:5SXipcwțXispw/C0/C1
bipcw/C0bispwðȚ
ð3Ț
Where EðYipcwȚ/C0EðYispwȚrefers to the total effects, or
the predicted weighted logged wage differences ,which
is the sum of endowment effects and coefficienteffects;5These types of positions are mutually exclusive; also, the
survey did not ask if a NP held multiple positions. All of the typesof positions were included in the final analysis because it is possi-
ble that PCNPs could work in a position that does not require NP
credentialing, e.g., NPs working in a hospital can work in a RNstaff position but still hold a NP certification. The percentage ofthis group of NPs is very small in our sample (3%), which did not
significantly affect the analysis and results.
6Because NPs work in a primary care settings may not practice
in a primary care specialty . For example, for a NP working in a car-
diologist office, she/he works in a primary care settings but prac-
tice in a specialty care specialty. Primary care specialty refers to
internal medicine, family practice, geriatrics, general pediatrics,OB/GYN women’s health, and school health. Specialty care spe-
cialty refers to pediatric subspecialties, adolescent medicine, car-
diology, endocrinology, gastroenterology, hematology/oncology,infectious disease, pulmonary/respiratory, renal/nephrology,rheumatology, general surgery, urology, orthopedics, other surgi-
cal specialties, allergy & immunology, dermatology, emergency
care, hospitalist, intensive care, neonatal, neurology, occupa-tional health, palliative care/pain management, psychiatry/men-tal health, rehabilitation, urgent care, wound/ostomy, and
surgical (anesthesia, cardio, cardiothoracic, vascular, thoracic,
neurological, radiology).Nurs Outlook 66 (2018) 528 /C0538 531

0:5SðbipcwțbispwȚðXipcw/C0XispwȚrefers to the endow-
ment effects , indicating how much the wages of PCNPs
in specialty care settings would change if theirweighted characteristics, X
ispw, were exactly the same
as those of PCNPs in primary care settings,
Xipcw;0 :5SðXipcwțXispwȚðbipcw/C0bispwȚrefers to the coeffi-
cient effects , indicating how much the wages of PCNPs in
specialty care settings would change if the coefficient,b
ispw, representing each of their characteristics was
exactly the same as those of PCNPs in primary care set-
tings, bipcw. This wage decomposition technique deter-
mined the extent to which endowment effects andcoefficient effects contributed to the wage differences.
Findings
Descriptive statistics for the full PCNP sample and thetwo subsamples of PCNPs in primary care and spe-cialty care settings are shown in Table 1 . T-tests and
Chi-square tests were used to compare the two sub-samples of PCNPs. The unadjusted average wage forPCNPs in primary care and specialty care settings were
$43.7 and $47.2, respectively, indicating that PCNPs in
primary care on average earned 8% less than those inspecialty care settings. Compared with PCNPs workingin specialty care settings, PCNPs who worked in pri-mary care settings were: older, female, white, andmarried; had a lower level of education, more years ofexperience; and had no plans to leave their currentposition; worked as a NP, worked in a clinic thatfocused on a primary care specialty, practiced inde-pendently, practiced in a rural or isolated area, werepaid an annual salary (versus hourly wage), and held
more than one job.
Wage differences between primary care and specialty
care settings
Because we assumed that the effects of PCNPs’
working setting on wages would differ betweenPCNPs in primary care and specialty care settings,we conducted the FGLS analysis for models of thefull sample and for separate models of PCNPs in pri-mary care and specialty care settings. Findings for
each statistically significant variable are presented
below.
7
Setting
The results for the full PCNP sample indicate thatthe wages of PCNPs in primary care settings were,on average, 7.1% lower (= 100*(exp( /C00.074) /C01))than
those of PCNPs in specialty care settings, holdingsocio-demographic, human capital, and employment
characteristics constant.Socio-demographics
For the full PCNP sample and two subsamples, male
PCNPs earned more than females. For PCNPs in spe-cialty care settings, PCNPs who were nonwhite earned
more than white PCNPs. For the full PCNP sample,
PCNPs who had never married earned less than thosePCNPs who were married.
Human Capital
For the full PCNP sample and PCNPs in primary caresettings, PCNPs prepared in a NP certificate programearned less than PCNPs with a master’s degree.
Years of experience had a significant positive associ-
ation with PCNP wages for both the full PCNP sampleand the two PCNP subsamples. Figure 1 illustrates
wage-experience curves for PCNPs in primary care andspecialty care settings. The curves were estimatedusing the average predicted wage for each PCNP givenhis/her years of experience. PCNPs in primary care set-tings earned, on average, $39.66/hour in their first yearas a NP. Early in their career, each additional year ofexperience led to a wage increase of approximately1.6%. Because of the nonlinear relationship between
PCNP experience and wage, the wage effect diminishes
gradually as experience increases, reaching a pointnear 20 years of experience where wages begin todecrease with further increases in experience. Themaximum wage predicted, $46.26, is approximately16.6% higher than entry-level wages.
The wage-experience curve for PCNPs in specialty
care settings is different than that of PCNPs in primarycare settings. The average entry-level wage for PCNPsin specialty care settings was $42.74/hour, which was7.8% higher than that of PCNPs in primary care set-
tings. Early in their careers, each additional year of
experience led to a wage increase of approximately1.4%, similar to the increase for PCNPs in primary caresettings. The wages of PCNPs in specialty care settingsalso peaked near 20 years of experience at the wage of$49.85/hour, which was 16.6% higher than their entry-level wage. Given their higher entry-level wages and asimilar rate increase as those of PCNPs in primary caresettings, the peak wage for PCNPs in specialty care set-tings was 7.8% higher than that of PCNPs in primarycare settings.
For the full sample and the subsample of PCNPs pri-
mary care settings, PCNPs who intended to leave theirjobs in the 1 to 2 years following the survey adminis-tration earned less than those who did not. However,this relationship did not hold for PCNPs working inspecialty care settings.
Employment
For all samples, PCNPs employed in RN staff positionsearned less than PCNPs employed in NP positions inclinical practice; and PCNPs who worked in clinicsfocused on primary care earned less than PCNPs work-ing in clinics focused on specialty care.
For the full sample and the subsample of PCNPs in
primary care settings, PCNPs supervised by physicians
7Detailed information of the regression models and results
can be obtained from the corresponding author.532 Nurs Outlook 66 (2018) 528 /C0538

Table 1 – Descriptive Statistics of PCNP Sample.
Variables Mean (SD) / n (%)
Full PCNP
Sample (n = 7,532)PCNPs in
Primary Care
Setting (n = 5,793, 76.9%)PCNPs in Specialty
Care Settings
(n = 1,739, 23.1%)p-Value
Dependent Variable
Hourly Wage (mean/median)a$44.5 (14.1)/42.7 $43.7 (14.1)/41.9 $47.2 (13.6)/45.7 0.00 *
Independent VariableSocio-demographic CharacteristicsAge 47.1 (10.8) 47.4 (10.9) 45.9 (10.6) 0.00 *
Gender
Female 7,042 (93.5%) 5,469 (94.4%) 1,573 (90.5%) 0.00 *
Male 490 (6.5%) 324 (5.6%) 166 (9.5%) 0.00 *
Race
White 6,630 (88.0%) 5,146 (88.8%) 1,484 (85.3%) 0.00 *
Non-white 902 (12.0%) 647 (11.2%) 255 (14.7%) 0.00 *
Ethnicity
Hispanic 263 (3.5%) 193 (3.3%) 70 (4.0%) 0.17
Non-Hispanic 7,269 (96.5%) 5,600 (96.7%) 1,669 (96.0%) 0.17Marital StatusNever married 752 (10.0%) 533 (9.2%) 219 (12.6%) 0.00 *
Married 5,531 (73.4%) 4,333 (74.8%) 1,198 (68.9%) 0.00 *
Separated, divorced, widowed, and other 1,249 (16.6%) 927 (16.0%) 322 (18.5%) 0.01 *
Human Capital CharacteristicsEducation level
Certificate 494 (6.6%) 443 (7.6%) 51 (2.9%) 0.00 *
Master 5,963 (79.2%) 4,559 (78.7%) 1,404 (80.7%) 0.07Post master 942 (12.5%) 692 (11.9%) 250 (14.4%) 0.01 *
DNP or other 133 (1.8%) 99 (1.7%) 34 (2.0%) 0.49
Experience of working as a NP 8.3 (10.5) 11 (8.6) 8.81 (7.12) 0.00 *
Intent to TurnoverNo plans to leave 5,062 (67.2%) 3,933 (67.9%) 1,129 (64.9%) 0.02 *
Leave in 2012 483 (6.4%) 360 (6.2%) 123 (7.1%) 0.2
Leave in next 1-2 years 958 (12.7%) 716 (12.4%) 242 (13.9%) 0.09Undecided or unknown 1,029 (13.7%) 784 (13.5%) 245 (14.1%) 0.56Employment Characteristics
Position
NP in clinic 6,781 (90.0%) 5,376 (92.8%) 1,405 (80.8%) 0.00 *
Other NP position 171 (2.3%) 126 (2.2%) 45 (2.6%) 0.31
Staff 254 (3.4%) 94 (1.6%) 160 (9.2%) 0.00 *
Other non-NP position 326 (4.3%) 197 (3.4%) 129 (7.4%) 0.00 *
Specialty of Clinics
b
Primary care specialty 3,849 (51.1%) 3,582 (61.8%) 267 (15.4%) 0.00 *
Specialty care specialty 3,004 (39.9%) 1,785 (30.8%) 1,219 (70.1%) 0.00 *
Other, or no specialty 679 (9.0%) 426 (7.4%) 253 (14.5%) 0.00 *
Relationship with physicianIndependent 785 (10.4%) 727 (12.5%) 58 (3.3%) 0.00 *
Collaborate with a physician 5,186 (68.9%) 4,162 (71.8%) 1,024 (58.9%) 0.00 *
Supervised by a physician 958 (12.7%) 539 (9.3%) 419 (24.1%) 0.00 *
Other relationship 603 (8.0%) 365 (6.3%) 238 (13.7%) 0.00 *
MSA Region
Urban 5,998 (79.6%) 4,530 (78.2%) 1,468 (84.4%) 0.00 *
Large rural 667 (8.9%) 560 (9.7%) 107 (6.2%) 0.00 *
Small rural 302 (4.0%) 264 (4.6%) 38 (2.2%) 0.00 *
Isolated 191 (2.5%) 173 (3.0%) 18 (1.0%) 0.00 *
Unknown 374 (5.0%) 266 (4.6%) 108 (6.2%) 0.01 *
Census RegionNew England 569 (7.6%) 467 (8.1%) 102 (5.9%) 0.00 *
Middle Atlantic 966 (12.8%) 667 (11.5%) 299 (17.2%) 0.00 *
East North Central 1,000 (13.3%) 765 (13.2%) 235 (13.5%) 0.74West North Central 569 (7.6%) 422 (7.3%) 147 (8.5%) 0.11South Atlantic 1,405 (18.7%) 1,061 (18.3%) 344 (19.8%) 0.17
East South Central 562 (7.5%) 447 (7.7%) 115 (6.6%) 0.13
West South Central 631 (8.4%) 517 (8.9%) 114 (6.6%) 0.00 *
Mountain 500 (6.6%) 421 (7.3%) 79 (4.5%) 0.00 *
Pacific 956 (12.7%) 760 (13.1%) 196 (11.3%) 0.04 *
(continued on next page )Nurs Outlook 66 (2018) 528 /C0538 533

earned less than PCNPs who practiced independently.
However, this relationship did not hold for PCNPs inspecialty care settings.
For PCNPs in specialty care settings, PCNPs in large
rural areas earned 6.5% less than those in urban areas.For the full sample and the subsample of PCNPs in pri-mary care settings, PCNPs employed in all regions
earned more than those employed in the South Atlantic
region, except PCNPs working in the West North Centraland East South Central regions. For PCNPs in specialtycare settings, those employed in New England, MiddleAtlantic, West South Central, and Pacific regions earnedmore than PCNPs employed in the South Atlantic region.
For the full sample and the two subsamples, PCNPs
who were paid on an hourly basis earned more thanthose who were paid an annual salary. For the fullsample and the subsample of PCNPs in primary caresettings, PCNPs who held more than one NP or RN
position earned more than PCNPs who did not. How-
ever, this relationship did not hold for PCNPs in spe-cialty care settings.
Predicted Wages Differences
The average predicted weighted wages for PCNPs in pri-mary care and specialty care settings were $43.80/hourand $47.93/hour, respectively. The total wage differencewas $4.13, indicating that PCNPs in primary care settingsearned 9.4% less than those in specialty care settings.
Sources of Wage Differences
Recall that we used a wage decomposition technique andd e c o m p o s e dt h ew a g ed i f f e r e nces into two components
/C0the “endowment” effect /C0the effect due to differencesin PCNPs’ characteristics /C0and the “coefficient” effects.
8
The total effects were /C00.1117, indicating the wages of
PCNPs in primary care settings were 10.6% lower (100*[ exp
(¡0.1117) /C01]) than the wages of PCNPs working in spe-
cialty care settings. The endo wment effects were esti-
mated to be /C00.0045, which explained 4% of the total
wage differences. That is, PCNP s in specialty care settings
would earn 0.45% less (100*[ exp(/C00.0045) /C01]) if their
socio-demographic, human capital, and employmentcharacteristics were exactly the same as those of PCNPsworking in primary care settings. PCNPs in primary caresettings had greater endowments than PCNPs in specialtycare settings for some characteristics that are associatedwith higher wages (i.e., PCNPs had more experience, andfewer worked in staff position s). Specifically, if PCNPs in
specialty care settings had the same years of experienceor the same percent of PCNPs worked in a staff positionas those working in primary c are settings, their wages
would increase approximately 2.2% or 1.2%, respectively.
However, PCNPs in primary care settings also possessedfewer endowments for some other characteristics thatwere also associated with lower wages, such as fewerPCNPs were male, fewer PCNPs worked in a specialty careclinic, and fewer PCNPs held more than one job. If thesame percent of PCNPs in specialty care settings weremale, worked in a specialty care clinic , or held more thanone job, their wages would decrease about 0.4%, 1.7%, or
0.3%, respectively.
The coefficient effects were /C00.1073, which explained
96% of the total wage differences. That is, PCNPs in spe-
cialty care settings would earn 10.2% less (100*[ expTable 1 – ( Continued )
Variables Mean (SD) / n (%)
Full PCNP
Sample (n = 7,532)PCNPs in
Primary CareSetting (n = 5,793, 76.9%)PCNPs in Specialty
Care Settings(n = 1,739, 23.1%)p-Value
Unknown 374 (5.0%) 266 (4.6%) 108 (6.2%) 0.01 *
How PCNPs are paid
Annual salary 4,488 (59.6%) 3,524 (60.8%) 964 (55.4%) 0.00 *
By the hour 2,076 (27.6%) 1,568 (27.1%) 508 (29.2%) 0.08Percentage of billing 266 (3.5%) 246 (4.2%) 20 (1.2%) 0.00 *
Other, or percent billing plus salary/hourly 702 (9.3%) 455 (7.9%) 247 (14.2%) 0.00 *
Hold more than one position
No 5,784 (76.8%) 4,549 (78.5%) 1,235 (71.0%) 0.00 *
Yes 1,748 (23.2%) 1,244 (21.5%) 504 (29.0%) 0.00 *
* Significant at the level of p <0.05.
aPrevious researchers have suggested recoding wage values that are less than $5 or greater than $100 as exactly $5 or $100,
respectively ( Jones & Gates, 2004 ). Because the wage calculation used in this study may yield some values that are not meaning-
ful estimations, the wages of 56 observations (2.0% of the sample) were recoded as $5 or $100. Eight wage estimations were
recoded as $5 and 48 estimations were recoded as $100. Wages ranged from $0.01 to $721.15.
bClinics with a primary care specialty refer to those focus on internal medicine, family practice, geriatrics, general pediatrics,
OB/GYN women’s health, and school health. Clinics with specialty care specialty refer to those focus on pediatric subspecialties,
adolescent medicine, cardiology, endocrinology, gastroenterology, hematology/oncology, infectious disease, pulmonary/respira-
tory, renal/nephrology, rheumatology, general surgery, urology, orthopedics, other surgical specialties, allergy & immunology,dermatology, emergency care, hospitalist, intensive care, neonatal, neurology, occupational health, palliative care/pain manage-
ment, psychiatry/mental health, rehabilitation, urgent care, wound/ostomy, and surgical (anesthesia, cardio, cardiothoracic,
vascular, thoracic, neurological, radiology).
8Detailed information of this wage decomposition model and
results can be obtained from the corresponding author.534 Nurs Outlook 66 (2018) 528 /C0538

(/C00.1073) /C01]) if the coefficients for their socio-demo-
graphic, human capital, and employment characteris-tics were exactly the same as those of PCNPs in primarycare settings. These effects were mainly due to the vari-
able of PCNPs’ independent practice. Specifically, PCNPs
who worked in primary care settings and were super-vised by a physician earned, on average, less than thoseworking independently; however this relationship wasnot statistically significant for PCNPs in specialty caresettings (as discussed early). If PCNPs in specialty caresettings had the same coefficients of PCNPs’ indepen-dence of practice as PCNPs in primary care settings,their wages would decrease 7.0%.
Discussion
This study found that PCNPs in primary care settings
earned significantly less than PCNPs in specialty caresettings, and most of these wage differences could not
be explained by endowment effects, i.e., PCNPs’ humancapital and personal attributes. PCNPs in primary careand specialty care settings possessed different endow-
ments, but these differences did not significantly con-
tribute to wage differences for these two groups. Anexplanation for this finding is that, although PCNPs inprimary care possessed greater human capital endow-ments (e.g., more years of experience) than PCNPs inspecialty care, they possessed other characteristics thatare known to be associated with lower wages, such asbeing predominantly female, and working only one job,which effectively offset any potential wage gains ( Jones
& Gates, 2004; Kalist, 2002 ). In fact, differences in total
endowment effects were small for these two groups.
The observed wage differences between PCNPs in pri-
mary care and specialty care settings can thus be attrib-uted to coefficient effects. Specifically, PCNPs’ practiceindependence was a significant factor contributing tothese effects. Generally, PCNPs working in a job whereFigure 1 – Wage-experience profiles for PCNPs working in primary care settings and those working in specialtycare settings. The lines indicate whether the experience was within the 25th /C075th percentiles (solid and
thicker lines), 10th /C090th percentiles (dashed lines), or above the 90th percentiles (dots). For PCNPs working in
primary care settings, about 25% to 75% had 4 to 15 years of experience and had expected wages between$41.94/hour and $45.83/hour. For PCNPs working in specialty care settings, approximately 25% to 75% had 3 to13 years of experience and earned wages between $44.62/hour and $49.94/hour. Note: Wages were the average
prediction, using the Duan’s smearing estimators, for each subsample of PCNPs.Nurs Outlook 66 (2018) 528 /C0538 535

they have greater levels of independence take on more
risk and thus, earn higher wages. Interestingly, this rela-tionship held for PCNPs in primary care settings but notfor those in specialty care settings. This finding may bebecause only a small number of PCNPs reported practicingindependently in specialty care settings, which makes itdifficult to detect an association between PCNP wages andtheir practice independence. Another explanation is thatPCNPs in specialty care settings are compensated basedon other factors that were not examined in this study.
Using the wage decomposition technique, this study not
only examined the sources of wages differences but alsoidentified both non-modifiable and modifiable factors thatmay be useful to managers of primary care organizationsand policy makers. The non-modifiable factors mayinclude PCNPs’ demographic characteristics. For example,we found that female PCNPs earned less than male PCNPsh o l d i n go t h e rf a c t o r sc o n s t a n t ;a n dt h i sr e s u l ti sc o n s i s -tent with previous studies ( Greene, El-Banna, Briggs, &
Park, 2017; Jones & Gates, 2004; Kalist, 2002; Muench, Sin-delar, Busch, & Buerhaus, 2015 ). Although gender is gener-
ally non-modifiable, these results may still suggest that
managers and policy makers examine their pay structuresto eliminate gender-based wage differences.
We also found some factors that potentially could be
modified by policy makers at the practice, system andpublic policy levels. First, since we found that PCNPs whopracticed independently earned higher wages than thosew h ow e r es u p e r v i s e db yap h y sician, policy-makers
should consider whether a change in PCNP practice inde-pendence, and a potential wage increase, would helpaddress the shortage of provid ers in primary care settings
(Poghosyan, Liu, Shang, & D’Aunno, 2015 ). Second, our
results indicate that PCNPs working in rural areas earned
less than those working in ur ban areas, a finding which
is in contrast to wages for primary care physicians, inw h i c hc a s et h o s ew h ow o r ki nr u r a la r e a so na v e r a g eearn more than those working in urban areas(Medicare Payment Advisory Commission, 2012 ). One
r e a s o nf o rt h i si st h a tM e d i c a r ep h y s i c i a n sr e c e i v eap a y -ment bonus of 10% if they practice in rural areas(Centers for Medicare and Medicaid Services, 2014 ). Thus,
policy makers and managers in rural, primary care set-tings should consider providing such bonus payments
and other incentives to attrac t PCNPs to practice in rural
communities, where nurses often serve as the main pro-viders of care ( Josiah Macy Jr. Foundation, 2016 ). Third,
this study found that PCNPs in primary care settings whoplanned to leave their jobs earned less than their coun-terparts in specialty care settings. If policy-makers wishto retain PCNPs in primary care settings, actions are
needed to change payment policies for PCNPs.
The findings presented in Figure 1 show that regardless
of setting, PCNPs’ peak wages were only 16.6% higherthan entry-level wages, which may indicate wage com-
pression in the PCNP labor market. Wage compression
refers to a phenomenon whereby small differencesbetween individuals’ peak wages and entry-level wagesresult in the underemployment of experienced and pro-ductive workers ( Pierce, Freund, Luikart, & Fondren, 1991 ).Wage compression, a l ong-standing problem in nursing, is
due to the lack of financial recognition for experience andproductivity and is significantly associated with nurses’job satisfaction and retention ( Greipp, 2003 ). Others have
noted that RN salaries are is likely to increase by less than69% throughout their career, compared with 109% foraccountants and 184% for engineers ( Evans & Carlson,
1992; Lynn & Redman, 2006 ). It may be even worse in the
PCNP workforce because PCNP wages had increased lessthan the 27% observed for nurses in general ( Jones &
Gates, 2004 ). Nurses may be more likely to leave their posi-
tion or even leave the profession altogether if they see lim-ited opportunities for wage increases during their career(N o o n e y ,U n r u h ,&Y o r e ,2 0 1 0 ). Thus, solving wage com-
pression may help increasing t he returns to PCNPs’ invest-
ments on their human capital, recognizing PCNPs’productivity and further ma intain their employment in
primary care settings.
There are other potentially modifiable factors that
were not captured in this analysis. First, PCNPs in spe-cialty care settings are more likely than those in primary
care settings to work on a night or late shift or have to
take call and thus, earn higher wages ( Schumacher &
Hirsch, 1997 ). Managers of primary care practices may
review their payment structure and improve PCNPs’compensations for working overtime. Also, under theMedicare claims payment structure, PCNPs who work inhospitals are usually paid a fixed salary, but PCNPs whowork in physician offices are reimbursed at 85% to 100%of physician fees ( Chapman, Wides, & Spetz, 2010 ).
Although little is known about whether PCNPs workingin primary care settings earn less under such a paymentsystem, the findings suggest that PCNP wages are influ-
enced by payment policies. Examining how the payment
structure of the third-party payer affects PCNP wagesmay help narrow the wage difference between primarycare and specialty care settings.
M o r e o v e r ,P C N P sw h ow o r ki ns p e c i a l t yc a r es e t t i n g s
are typically supervised by a specialist and may earnmore than those working in primary care settings whoare supervised by a primary care physician, because spe-cialists, on average, earn more than primary care physi-cians and may pay PCNPs more. Previous studies havedocumented that physicians working in primary care set-
tings earn lower wages than those working in specialty
care settings ( Shih & Konrad, 2007; Simon & Born, 1996 ).
Therefore, the wage differences between working in pri-mary care and specialty care are not unique to nursingp e rs e ,b u ta l s ot op h y s i c i a n s . An important policy consid-
eration is therefore to narrow the wage gap between spe-cialist and primary care physicians, which may, in turn,eliminate some of the wage differences between NPsworking in specialty versus primary care settings.
The results of this study should be interpreted in light
of certain limitations. Using self-reported survey data
may affect the accuracy of results, as variables in the
wage model may contain measurement error. For exam-ple, measuring years of experience using the survey yearminus the year when PCNPs received their NP educationmay not accurately reflect PCNPs’ years of experience.536 Nurs Outlook 66 (2018) 528 /C0538

Also, the analytic methods of modeling PCNP wages may
be problematic due to potential model specificationerror. For example, NP wages may be influenced byreceiving extra pay for shif twork, or for working full-
time versus part-time on their position, but these varia-
bles were not included in the 2012 NSSNP.
Also, because NSSNP data were cross-sectional and
gathered in 2011, the results of this study reflect PCNPwages at only one point in time, and do not reflect anywage changes that may have occurred subsequently.
Additionally, we do not know whether the wage differen-
ces for PCNPs in primary care and specialty care settingsobserved in this study have changed over time.
Numerous policy changes have occurred since the
implementation of the Consensus Model for AdvancedPractice Registered Nurse (APRN) in 2008, which wasimplemented to clarify the four areas of APRN licen-sure, accreditation, certification, and education(National Council of State Boards of Nursing, 2008 ). A
subsequent analysis of how this policy change hasimpacted NP wages is warranted, if and when data
from future surveys are available. Despite these con-
cerns, this dataset was the most currently availabledataset representing the NP workforce at the time thisstudy was conducted. Therefore, this paper representsan important step in describing PCNP wages, examiningthe setting-based PCNP wage differences, and attempt-ing to explain why these differences may exist.
Conclusions
This study reported that PCNPs working in primarycare settings earned, on average, considerably less
than PCNPs working in specialty care settings. These
wage differences were not fully explained by PCNPs’socio-demographic, human capital, and employmentcharacteristics, but were largely due to unexplainedfactors. These differences may instead reflect the dif-ferent working environments and payment policiesbetween primary care and specialty care settings.Future research is needed to explore these factors.
Acknowledgement
We thank the National Center for Health WorkforceAnalysis, Health Resources and Services Administration,US Department of Health and Human Services for accessto the National Sample Survey of Nurse Practitioners.
REFERENCES
Becker, G. S. (1962). Investment in human capital: A
theoretical analysis. The Journal of Political Economy,
70, 9–49.Becker, G. S. (2009). Human capital: A theoretical and empiri-
cal analysis, with special reference to education . University
of Chicago Press, Chicago, IL.
Becker, G. S. (2010). The economics of discrimination . Univer-
sity of Chicago Press, Chicago, IL.
Bodenheimer, T., & Bauer, L. (2016). Rethinking the
Primary Care Workforce – An Expanded Role forNurses. N Engl J Med, 375 (11), 1015–1017, doi: 10.1056/
NEJMp1606869 .
Budd, G., Wolf, A., & Haas, R. (2015). Addressing the pri-
mary care workforce: A study of nurse practitioner stu-dents’ plans after graduation. The Journal of nursing
education, 54 (3), 130–136.
Centers for Medicare and Medicaid Services. (2014). Physi-
cian Bonuses. Retrieved from https://www.cms.gov/
Medicare/Medicare-Fee-for-Service-Payment/HPSAP
SAPhysicianBonuses/ .
Chapman, S. A., Wides, C. D., & Spetz, J. (2010). Payment
Regulations for Advanced Practice Nurses: Implica-
tions for Primary Care. Policy, Politics, & Nursing Practice,
11(2), 89–98, doi: 10.1177/1527154410382458 .
C h a t t o p a d h y a y ,A . ,Z a n g a r o ,G .A . ,&W h i t e ,K .M .
(2015). Practice Patterns and Characteristics of
Nurse Practitioners in the United States: Results
From the 2012 National Sample Survey of NursePractitioners. The Journal for Nurse Practitioners, 11 (2),
170–177.
Coomer, N. M. (2013). Historical data indicates a wage pre-
mium for black registered nurses. Nurs Econ, 31 (5), 254–
259.
Cotton, J. (1988). On the decomposition of wage differen-
tials. The review of economics and statistics, 70 , 236–243.
Currie, J., & Madrian, B. C. (1999). Health, health insurance
and the labor market. Handbook of labor economics, 3 ,
3309–3416.
De Milt, D. G., Fitzpatrick, J. J., & Sister Rita, M. (2011).
Nurse practitioners’ job satisfaction and intent to leavecurrent positions, the nursing profession, and the
nurse practitioner role as a direct care provider. JA m
Acad Nurse Pract, 23 (1), 42–50, doi: 10.1111/j.1745-
7599.2010.00570.x .
Ehrenberg, R. G., & Smith, R. S. (2010). Modern labor economics:
Theory and Policy . Prentice Hall, Upper Saddle River, NJ.
Evans, S. A., & Carlson, R. (1992). Nurse-physician collabo-
ration: Solving the nursing shortage crisis. Journal of the
American College of Cardiology, 20 (7), 1669–1673.
Goolsby, M. J. (2006). 2004 AANP National Nurse Practi-
tioner Sample Survey, part III: NP income and benefits.J Am Acad Nurse Pract, 18 (1), 2–5, doi: 10.1111/j.1745-
7599.2005.00094.x .
Goolsby, M. J. (2009). 2008 AANP National NP Compensa-
tion Survey. J Am Acad Nurse Pract, 21 (3), 186–188,
doi:10.1111/j.1745-7599.2009.00400.x .
Greene, J., El-Banna, M. M., Briggs, L. A., & Park, J. (2017).
Gender differences in nurse practitioner salaries. JA m
Assoc Nurse Pract, 29 (11), 667–672, doi: 10.1002/2327-
6924.12512 .
Greipp, M. E. (2003). Salary compression: Its effect on
nurse recruitment and retention. Journal of Nursing
Administration, 33 (6), 321–323.
Heisler, E. J., & Sarata, A. K. (2011). physician supply and the
patient protection and affordable care act . Retrieved from
Holtmann, A. G., & Idson, T. L. (1993). Wage determination
of registered nurses in proprietary and nonprofit nurs-
ing homes. Journal of Human resources, 28 , 55–79.
Jones, C. B., & Gates, M. (2004). Gender-based wage differen-
tials in a predominantly female profession: Observationsfrom nursing. Economics of Education Review, 23 (6), 615–631.Nurs Outlook 66 (2018) 528 /C0538 537

Josiah Macy Jr. Foundation. (2016). Registered Nurses: Part-
ners in Transforming Primary Care: Recommendations from
the Macy Foundation Conference on Preparing RegisteredNurses for Enhanced Roles in Primary Care. Atlanta, GA.
Kalist, D. E. (2002). The gender earnings gap in the RN
labor market. Nursing economic$, 20 (4), 155–162.
Keough, V. A., Stevenson, A., Martinovich, Z., Young, R., &
Tanabe, P. (2011). Nurse practitioner certification and
practice settings: Implications for education and prac-
tice. J Nurs Scholarsh, 43 (2), 195–202, doi: 10.1111/j.1547-
5069.2011.01395.x .
Kiker, B. F. (1966). The historical roots of the concept of
human capital. The Journal of Political Economy, 74 ,4 8 1 – 4 9 9 .
Lehrer, E. L., White, W. D., & Young, W. B. (1991). The three
avenues to a registered nurse license: A comparativeanalysis. Journal of Human resources, 26 , 362–379.
Lynn, M. R., & Redman, R. W. (2006). Staff nurses and their
solutions to the nursing shortage. West J Nurs Res, 28 (6),
678–693.
McGregory, R. C. Jr (2013). An Analysis of Black /C0White
Wage Differences in Nursing: Wage Gap or Wage Pre-mium. The Review of Black Political Economy, 40 (1), 31–37.
McGregory, R. C., Niederjohn, M. S., & Peoples, J. (2009). Do
part-time/full-time compensation differentials for
nurses vary between the private and public sector.Applied Economics, 41 (4), 537–546.
Medicare Payment Advisory Commission. (2012). Report to
the Congress: Medicare Payment Policy Retrieved from.
Mincer, J. A. (1958). Investment in human capital and per-
sonal income distribution. The Journal of Political Econ-
omy, 66 , 281–302.
Mincer, J. A. (1974). Schooling, experience, and earnings .N a t i o n a l
Brureau of Economic Research Books, Cambridge, MA.
Muench, U., Sindelar, J., Busch, S. H., & Buerhaus, P. I.
(2015). Salary differences between male and female
registered nurses in the United States. Jama, 313 (12),
1265–1267, doi: 10.1001/jama.2015.1487 .
National Council of State Boards of Nursing. (2008). Con-
sensus model for APRN regulation: Licensure, accredi-
tation, certification & education.
Nooney, J. G., Unruh, L., & Yore, M. M. (2010). Should I stay
or should I go? Career change and labor force separa-
tion among registered nurses in the US. Social Science &
Medicine, 70 (12), 1874–1881.
Oaxaca, R. (1973). Male-female wage differentials in
urban labor markets. International economic review, 14 ,
693–709.
Pasaron, R. (2013). Nurse practitioner job satisfaction:
Looking for successful outcomes. J Clin Nurs, 22 (17-18),
2593–2604, doi: 10.1111/j.1365-2702.2012.04331.x .
P e t t e r s o n ,S .M . ,P h i l l i p s ,R .L .J r . ,B a z e m o r e ,A .W . ,
B u r k e ,T .B . ,&K o i n i s ,G .T .( 2 0 1 3 ) .R e l y i n go nN P sa n dP A sd o e sn o ta v o i dt h en e e df o rp o l i c ysolutions for primary care. Am Fam Physician, 88(4),
230.
Pierce, S. F., Freund, C. M., Luikart, C., & Fondren, L. (1991).
Nurses Employed in Nonnursing Fields: Is Nursing Los-ing Its Best and Brightest. Journal of Nursing Administra-
tion, 21 (6), 29–34.
Poghosyan, L., Liu, J., Shang, J., & D’Aunno, T. (2017). Prac-
tice environments and job satisfaction and turnover
intentions of nurse practitioners: Implications for pri-
mary care workforce capacity. Health Care Manage Rev,
42
(2), 162–171, doi: 10.1097/hmr.0000000000000094 .
Reimers, C. W. (1983). Labor market discrimination
against Hispanic and black men. The review of economics
and statistics, 65 , 570–579.
Schultz, T. W. (1961). Investment in human capital. The
American economic review, 51 , 1–17.
Schumacher, E. J., & Hirsch, B. T. (1997). Compensating
differentials and unmeasured ability in the labor mar-ket for nurses: Why do hospitals pay more. Industrial
and Labor Relations Review, 50 , 557–579.
Shih, Y. C., & Konrad, T. R. (2007). Factors associated
with the income distribution of full-time physicians:A quantile regression approach. Health Serv Res, 42 (5),
1895–1925, doi: 10.1111/j.1475-6773.2006.00690.x .
Simon, C. J., & Born, P. H. (1996). Physician earnings in a
changing managed care environment. Health Affairs, 15
(3), 124–133, doi: 10.1377/hlthaff.15.3.124 .
Spetz, J. (2002). The value of education in a licensed pro-
fession: The choice of associate or baccalaureatedegrees in nursing. Economics of Education Review, 21 (1),
73–85.
Spetz, J., Fraher, E., Li, Y., & Bates, T. (2015). How Many
Nurse Practitioners Provide Primary Care? It DependsOn How You Count Them. Medical Care Research and
Review, 72 (3), 359–75 1077558715579868 .
U. S. Department of Health and Human Services. (2016).
State-Level Projections of Supply and Demand for PrimaryCare Practitioners: 2013-2025 . Retrieved from Rockville,
Maryland.:
U.S. Department of Health and Human Services. (2014).
The 2012 National Sample Survey of Nurse Practitioners:Variables Definitions Document . Retrieved from Rockville,
Maryland:
Walani, S. R. (2013). Earnings of the Internationally Edu-
cated Nurses in the U.S. Labor Market. Nursing research,
62(3), 169–177.
Wilder, V., Dodoo, M. S., Phillips, R. L. Jr, Teevan, B.,
Bazemore, A. W., Petterson, S. M., & Xierali, I. (2010).Income disparities shape medical student specialty
choice. Am Fam Physician, 82 (6), 601.
Willis, R. J. (1985). Wage determinants: A survey and rein-
terpretation of human capital earnings functions. Eco-
nomics Research Center/NORC, 1 , 525–602.538 Nurs Outlook 66 (2018) 528 /C0538