Education work

zjlbjl112312p

HuWolniak.pdf

Home >Education homework help >Education work

The Review of Higher Education, Volume 36, Number 2, Winter 2013, pp. 211-233 (Article)

DOI: 10.1353/rhe.2013.0002

For additional information about this article

Access provided by Temple University (23 Sep 2015 17:52 GMT)

http://muse.jhu.edu/journals/rhe/summary/v036/36.2.hu.html

Hu & Wolniak / Student Engagement and Early Career Earnings 211

College Student Engagement and Early Career Earnings: Differences by Gender, Race/Ethnicity, and Academic Preparation Shouping Hu and Gregory C. Wolniak

Background

The quality of undergraduate education in the United States has con- stantly undergone scrutiny, while heightened skepticism and calls for higher education reform have recently intensified along with dissatisfaction about the current status of higher education (Arum & Roksa, 2011; National Commission on the Future of Higher Education, 2006). Promoting student engagement in educationally purposeful activities has been advocated as an effective way to transform undergraduate education (National Survey of

SHOUPING HU is Professor in the Department of Educational Leadership and Policy Stud- ies at Florida State University. GREGORY C. WOLNIAK is Senior Research Scientist in the Education and Child Development Department with NORC at the University of Chicago. Data for this research were provided by the Bill & Melinda Gates Foundation and coordinated by the Institute for the Higher Education Policy (IHEP). The research findings reflect the authors’ opinions and not necessarily those of the Gates Foundation or IHEP. Address queries to Shouping Hu at the Department of Educational Leadership and Policy Studies, College of Education, 1210H Stone Building, Florida State University, Tallahassee, FL 32306; telephone: (850) 644–6721; fax: (850) 644–1258; email: [email protected].

212 The Review of higheR educaTion Winter 2013

Student Engagement, 2004, 2005). Strong empirical evidence points to the promise of reform strategies centered around student engagement as the higher education literature unequivocally indicates that what matters most in student learning and personal development is what the students do in college (Astin, 1993; National Survey of Student Engagement, 2004, 2005; Pascarella & Terenzini, 1991, 2005). The concept of student engagement has historical roots in Astin’s theory of involvement (1999), as well as the concept of “quality of effort” put forth by Pace (1979) and Pascarella (1985). These concepts are theoretically and empirically associated with Chickering and Gamson’s (1987) Seven Principles for Good Practice in Undergraduate Educa- tion, including: (a) student-faculty contact, (b) cooperation among students, (c) active learning, (d) prompt feedback, (e) time on task, (f) communication of high expectations, and (g) respect of diverse talents and ways of learning.

Student engagement is an integral part of a quality education and plays an important role in many desirable college outcomes such as student learn- ing, academic performance, and persistence (Astin, 1993; Hu & Kuh, 2003; Kuh, Cruce, Shoup, Kinzie, & Gonyea, 2008; Pascarella & Terenzini, 1991, 2005). However, little is known about the relationship between student engagement and career or occupational outcomes following college. Given the prominent role of earnings in job satisfaction, socioeconomic status, and individual well-being, as well as evidence suggesting the ability to make more money and get a better job is a priority in students’ decisions to attend college (Astin, 1993), the lack of empirical evidence on the economic and career impacts of the college experience presents a void in our understanding of these important relationships.

Recent work by Hu and Wolniak (2010) demonstrated significant rela- tionships between measures of student engagement and labor market earn- ings following college. In this study, we sought to extend Hu and Wolniak’s (2010) earnings model of student engagement by focusing on student sub-populations. Across a vast array of postsecondary outcomes, evidence increasingly points to the fact that the effects of college and models of student development are not equally applicable to all students. Summarizing their extensive review of the college impact literature, Pascarella and Terenzini (2005) pointed to the increasing diversity in the college-going population to encourage greater attention to conditional effects; this approach offers the possibility of uncovering whether “any given college experience may have a different effect on different kinds of students” (p. 626).

It appears that the socioeconomic outcomes of college are highly condi- tional, while economic and sociological perspectives suggest that academic and social engagement, as well as subsequent labor market earnings, vary by student backgrounds (Knox, Lindsay, & Kolb, 1993; Pascarella & Terenzini, 2005; Thomas, 2003; Tinto, 1975). By focusing on the interactions between

Hu & Wolniak / Student Engagement and Early Career Earnings 213

academic and social dimensions of student engagement and students’ as- cribed and achieved characteristics prior to college, we examine whether the economic impacts of college are moderated by (or conditional on) student characteristics at college entry. In the sections below we provide an over- view of the theory and existing evidence that inform this study, followed by methods, results, and a discussion of the findings.

TheoreTical PersPecTives

In order to examine college student engagement in relation to postsecond- ary outcomes, we draw on economic and sociological perspectives related to human and social capital. These two perspectives are particularly relevant because student engagement in college activities has the feature of engage- ment in both the academic and social dimensions of campus life (Tinto, 1975, 1993).

Human capital is considered to be a set of skills that individuals acquire through education, training, and other means that improve health, productiv- ity, and therefore labor market earnings (Becker, 1994). In formalizing human capital theory, Becker (1994) applies the assumptions that schooling results in greater earnings and productivity because it provides “knowledge, skills, and a way of analyzing problems” (p. 19), and that individuals respond rationally to expected benefits and costs. Human capital theory therefore lends itself to understanding the earnings effects of postsecondary education by ground- ing educational experiences in the language of productivity-enhancement and investment returns (Becker, 1994; Paulsen, 2001). Job-related earnings following college provide a signal of the productive capabilities of students once they enter the labor market.

A positive relationship between student engagement during college and subsequent earnings would provide evidence of (a) the economic value tied to specific student behaviors, and (b) a policy rationale for promoting the kinds of educational programs and academic structures that facilitate purposeful involvement among students. Given the evidence suggesting a positive relationship between student engagement and student learning (Carini, Kuh, & Klein, 2006; Hu & Kuh, 2003; Kuh, Cruce, Shoup, Kinzie, & Gonyea, 2008; Pascarella & Terenzini, 1991, 2005), it is reasonable to expect that student engagement could be related to productivity in the labor market and, therefore, earnings. That is, student engagement in college could lead to attaining the kind of “knowledge, skills, and a way of analyzing problems” (Becker, 1994, p. 19) that accompany economic value in the labor market.

In addition to the human capital development that may result from en- gagement during college, sociological perspectives on student development add to our understanding of the overall impacts of student activities. De-

214 The Review of higheR educaTion Winter 2013

fined as social relations and the resources available through social networks (Coleman, 1988; Lin, 1999), social capital is central to understanding the impacts of the student experience and the college environment in relation to a variety of student-centered college outcomes. The extent to which students have access to social capital is largely related to the socioeconomic status of their families and to parent, peer, and friendship networks. Such social ties influence students’ educational and career choices, and contribute to their ability to achieve desirable educational outcomes. Similarly, the theory of peer influence suggests that academic outcomes are partially determined by the social environment accompanying the educational process (Hallinan, 1982; Sewell, Haller, & Portes, 1969), Further, models of student socializa- tion suggest that students’ values, aspirations, and educational choices are affected by the institutional context and interactions with a variety of different social agents, such as students, faculty, and staff (Weidman, 1989). Student engagement in college therefore has the potential to influence outcomes in the labor market directly through the accumulation of social capital and indirectly through its effects on educational and career decision-making.

Evidence of Engagement’s Conditional Effects

The role of background characteristics such as gender, race/ethnicity, and academic achievement are commonly used in studies of status attainment and economic attainment (Coleman, Hoffer, & Kilgore, 1982; Sewell, Haller, & Ohlendorf, 1970). Researchers have also demonstrated that background fac- tors exert strong direct and indirect influences on educational and economic outcomes (Coleman, Hoffer, & Kilgore, 1982; Hearn, 1988, 1991; Karen, 1991; Sewell, Haller, & Ohlendorf, 1970), and that individual characteristics can influence the interrelationship between inputs and related outcomes.

Economic research has found that the rates of return on education vary significantly for students of different gender, race/ethnicity, and socioeco- nomic status (Perna, 2005). An earlier study (Thomas, 2000) recommends examining postsecondary earnings premiums by different subpopulations of students. Among college-educated individuals, studies have shown that minority groups earn significantly less than their White counterparts, and this pattern may be worsened by economic characteristics such as the minority representation in a given labor market sector (Tienda & Lii, 1987).

Research on college students has also indicated that the relationship between student experiences and a host of outcomes varies by student backgrounds. In a study on student behavioral outcomes, Nora, Cabrera, Hagedorn, and Pascarella (1996) found that student academic and social experiences in college had different effects on student college outcomes by ethnic and gender groups. Group differences were also identified by Ca- brera, Nora, Terenzini, Pascarella, and Hagedorn’s (1999) study of student

Hu & Wolniak / Student Engagement and Early Career Earnings 215

persistence and were further discussed in the work by Rendón, Jalomo, and Nora (2000).

Other college impact studies have identified conditional effects based on students’ demographic, socioeconomic, and academic backgrounds. For ex- ample, Sax, Bryant, and Harper (2005) found that student-faculty interactions are related to a wide range of college outcomes such as GPA and satisfaction but that men and women report different results. Kim and Sax (2009) fur- ther found different effects of student-faculty interaction on a set of college outcomes based on student race/ethnicity, social class, and first-generation status. Carini, Kuh, and Klein (2006) found that students with the lowest SAT scores benefited more from student engagement than those with high- est SAT scores; they claim that this finding somewhat “dovetails with some recent research” (p. 23), while cautioning that measurement issues in survey research could have contributed to their findings. In a study examining the relationship between student engagement and success in college, Kuh et al. (2008) found that student engagement in educationally purposeful activi- ties can benefit all students but tend to carry greater benefit for low-ability students and students of color.

Existing theory and empirical evidence suggest that gender, race/ethnicity, and academic preparation play a substantial role in understanding the effects of student engagement on many important college outcomes, and research- ers have encouraged increased analysis of gender and racial/ethnic-related conditional effects across many types of student-centered college outcomes (Pascarella & Terenzini, 2005; Reason, Terenzini, & Domingo, 2006). Building on past evidence and drawing on perspectives based on human capital and social capital, we hypothesize that the influence which student engagement exerts on professional and career outcomes is moderated by ascribed and achieved characteristics at college entry.

Purpose and Research Question

The purpose of this study is to examine whether student background characteristics moderate the relationship between student engagement in col- lege activities and labor market earnings in the years immediately following college. Building on past research and extending Hu and Wolniak’s (2010) earnings model of student engagement, we focus on student sub-populations to examine if the early career economic impacts of student engagement are conditional rather than general.

The research question guiding this study is: Are the relationships between measures of student engagement and earnings moderated by students’ gender, race/ethnicity, and academic preparation?

216 The Review of higheR educaTion Winter 2013

MeThod

Data

To address our research question, we sought data containing information on college student engagement and post-college labor market earnings, and we identified data fitting this unique criterion based on the longitudinal surveys of the 2001 cohort of applicants to the Gates Millennium Scholars (GMS) program. The longitudinal and tracking component of the GMS project was designed and conducted by NORC at the University of Chicago with support from the Bill & Melinda Gates Foundation (Lodato Nichols, Zimowski, Lodato, & Ghadialy, 2004).

NORC administered three waves of surveys to the 2001 GMS cohort, comprised of freshmen entering college in the 2001–2002 academic year. The initial data were collected through a base-year survey focused on student backgrounds and college choices. The first follow-up survey, conducted in 2004, gathered information on a wide range of student experiences during college. The final wave of data was gathered in 2006 with a second follow- up survey to collect information on several facets of post-college outcomes roughly five years after the GMS applicants had completed high school. The information on student engagement in college activities that we examine was collected as part of the first follow-up survey, and information on individual earnings originated from the second follow-up survey.

Applying the same standards and survey procedures used in the national surveys conducted by the National Center of Educational Statistics (NCES), NORC surveyed all GMS recipients and representative samples of nonrecipi- ents. The general patterns showed that female students had higher response rates than their male counterparts, and that Asian Americans had the highest response rates, while American Indians had the lowest (NORC, 2003, 2006). Response rates for all waves were above 50%, and NORC derived weighting strategies to compensate for survey nonresponse and study attrition that adjusted for differences between the survey respondents and the overall population of 2001 GMS applicants (Kuo et al., 2006; NORC, 2003, 2006).

The analytic sample for our study included students who had graduated from college at the time of the second follow-up survey, mirroring other studies of labor market earnings (e.g., Thomas, 2000, 2003). We used the longitudinal panel weights for all analyses because we drew the variables in the models from each of the three survey waves. The total sample con- sisted of 1,278 respondents, including GMS scholarship award recipients and nonrecipients. The sample reflects the full population of GMS scholars and nonrecipients in the 2001 cohort in terms of gender and racial/ethnic compositions (Kuo et al., 2006; NORC, 2003, 2006).

Altogether, the sample was predominately female (74% female, 26% male)

Hu & Wolniak / Student Engagement and Early Career Earnings 217

and was distributed across four racial/ethnic minority groups, including Af- rican Americans (37%), Hispanic Americans (25%), Asian/Pacific Islanders (30%), and American Indian/Alaskan Natives (8%). Among all the students, 15% were in the low-SAT/ACT group, 34% high-SAT/ACT group, and 51% middle-range SAT/ACT group. Students in the middle-range group had SAT scores of 990–1,230, and students in the low- and high- SAT/ACT groups had scores of below 990 and above 1,230 respectively.

The NORC data set for the GMS program is a uniquely constructed and rich source of information enabling the examination of student engagement in college activities in relation to early career earnings among students of different gender, race/ethnicity, and academic preparation. While our analyses were designed to address the research question centered on the relationships between student engagement in college activities, early career labor market earnings, and the moderating effect of student background characteristics, it is important to note that the focus of this study was not to evaluate the GMS program and our results should not be interpreted as an indication of GMS program effects.

Variables

The dependent variable was the natural log of annual earnings as re- ported by college graduates. Annual earnings were collected from survey questions asking non-enrolled college graduates to indicate the amount of income earned at their current job as well as the job’s pay scale (e.g., annu- ally, monthly, weekly, etc.), which we then converted to an annual basis for comparability.

Independent variables included an indicator of GMS scholarship award status (recipient = 1, nonrecipient applicant = 0) to control for the poten- tial influence of GMS program selection on student earnings. Based on the literature of the impact of college on students’ labor market outcomes (Pas- carella & Terenzini, 1991, 2005) and Hu and Wolniak’s (2010) recent study, we included four groups of variables in the analytic models to account for the influence of student backgrounds, institutional characteristics, major fields, and student engagement. The description and coding strategy for all variables are presented in Table 1, and we analyzed the overall sample accord- ing to the following student characteristics: gender (female as the reference group), race/ethnicity (Asian Americans as the reference group), SAT/ACT scores (middle-SAT/ACT as the reference group), parental educational level, and the number of Advanced Placement (AP) exams students have taken.

The full analytic model included a host of other variables shown in the literature to be important for understanding college impacts (Pascarella & Terenzini, 2005). Concerning the institutions student attended, we included measures of institutional control (private as the reference group) and selectiv- ity (based on Barron’s Profile of American Colleges, 2001) converted into three

218 The Review of higheR educaTion Winter 2013 T

a b

l e 1

d e

sc R

ip T

io n

a n

d c

o d

in g

o f v

a R

ia b

l e

V ar

ia bl

es cr

ip ti

on a

n d

C od

in g

L o

g o

f an

n u

al i

n co

m e

N at

u ra

l lo

g o

f gr

ad u

at es

’ a n

n u

al i

n co

m e

fi ve

y ea

rs a

ft er

h ig

h s

ch o

o l

gr ad

u at

io n

. G

M S

re ci

p ie

n ts

ec ip

ie n

t o

f G

M S

sc h

o la

rs h

ip (

Y es

= 1

, N o

= 0

) M

al e

A d

u m

m y

va ri

ab le

( Y

es =

1 , N

o =

0 )

A fr

ic an

A m

er ic

A d

u m

m y

va ri

ab le

( Y

es =

1 , N

o =

0 )

A m

er ic

an I

n d

ia n

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) H

is p

an ic

A m

er ic

A d

u m

m y

va ri

ab le

( Y

es =

1 , N

o =

0 )

A si

an A

m er

ic an

ef er

en ce

g ro

u p

P ar

en ta

l ed

u ca

ti o

n o

f b

ac h

el o

r o

r ab

o ve

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) L

o w

-S A

T /A

C T

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) M

id d

le -S

A T

/A C

R ef

er en

ce g

ro u

p H

ig h

-S A

T /A

C T

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) N

u m

b er

o f

A P

e xa

m s

T h

e n

u m

b er

o f

A P

e xa

m s

st u

d en

ts h

av e

ta ke

n : 0

( n

o n

e) , 1

( o

n e)

, 2 (

tw o

), 3

( th

re e)

, a n

d 4

( fo

u r

an d

m o

re )

L o

w s

el ec

ti vi

A d

u m

m y

va ri

ab le

( Y

es =

1 , N

o =

0 );

S el

ec ti

vi ty

n o

h ig

h er

t h

an 2

o n

B ar

ro n’

s (2

0 0

1 )

M id

d le

s el

ec ti

vi ty

ef er

en ce

g ro

u p

; S el

ec ti

vi ty

a t

3 o

r 4

o n

B ar

ro n’

s (2

0 0

1 )

H ig

h s

el ec

ti vi

A d

u m

m y

va ri

ab le

( Y

es =

1 , N

o =

0 );

S el

ec ti

vi ty

h ig

h er

t h

an 4

o n

B ar

ro n’

s (2

0 0

1 )

P u

b li

c in

st it

u ti

o n

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) ST

E M

m aj

o ra

d u

m m

y va

ri ab

le (

Y es

= 1

, N o

= 0

) A

ca d

em ic

e n

ga ge

m en

t sc

al e

Su m

m at

io n

o f

re sp

o n

se r

an gi

n g

fr o

m 1

( le

ss t

h an

o n

ce a

m o

n th

) to

6 (

fo u

r o

r m

o re

t im

es a

w ee

k )

o n

t h

es e

it em

) W

o rk

w it

h o

th er

s tu

d en

ts o

n s

ch o

o l

w o

rk o

u ts

id e

o f

cl as

(2 )

D is

cu ss

i d

ea s

fr o

m y

o u

r re

ad in

gs o

r cl

as se

s w

it h

s tu

d en

ts o

u ts

id e

o f

cl as

(3 )

D is

cu ss

i d

ea s

fr o

m y

o u

r re

ad in

gs o

r cl

as se

s w

it h

f ac

u lt

y o

u ts

id e

o f

cl as

(4 )

W o

rk h

ar d

er t

h an

y o

u t

h o

u gh

t to

m ee

t an

i n

st ru

ct o

r’ s

ex p

ec ta

ti o

n s

Hu & Wolniak / Student Engagement and Early Career Earnings 219 S

o ci

al e

n ga

ge m

en t

sc al

e Su

m m

at io

n o

f re

sp o

n se

s ra

n gi

n g

fr o

m 1

( n

ev er

) to

5 (

ve ry

o ft

en )

o n

t h

es e

it em

) P

ar ti

ci p

at e

in e

ve n

ts s

p o

n so

re d

b y

a fr

at er

n it

y o

r so

ro ri

) P

ar ti

ci p

at e

in r

es id

en ce

h al

l ac

ti vi

ti es

(3 )

P ar

ti ci

p at

e in

e ve

n ts

o r

ac ti

vi ti

es s

p o

n so

re d

b y

gr o

u p

s re

fl ec

ti n

g yo

u r

o w

n c

u lt

u ra

l h

er it

ag e

) P

ar ti

ci p

at e

in c

o m

m u

n it

y se

rv ic

e ac

ti vi

ti es

N ot

e: a

S T

E M

m aj

o r

fi el

d s

in cl

u d

e b

io lo

gy ,

en gi

n ee

ri n

g, c

o m

p u

te r

sc ie

n ce

, sc

ie n

ce ,

an d

m at

h .

N o

n -S

T E

M m

aj o

r fi

el d

s in

cl u

d e

so ci

al s

ci en

ce s,

h u

m an

it ie

s, e

d u

ca ti

o n

, p

ro fe

ss io

n al

, a n

d o

th er

220 The Review of higheR educaTion Winter 2013

categories (low selectivity for a Barron’s rating of 1 or 2, middle selectivity for ratings of 3 or 4, and high selectivity for ratings of 5 or 6; middle selectivity was the reference group).

The variables indicating major fields of study were biology, engineering, computer science, math, and science fields, grouped into a single category of STEM fields. In addition, students in social science, humanities, educa- tion, and professional fields were classified as non-STEM students (reference group). We used the variables described here mainly as control variables so that the relationship between student engagement and early career earnings could be reliably estimated without the confounding influence of college major.

We included eight student engagement measures in the analysis repre- senting two engagement scales resulting from factor analysis (Table 2). The wording of the engagement measures was similar, but not identical, to that in the NSSE survey (National Survey of Student Engagement, 2004, 2005), resulting in a four-item scale measuring student academic engagement, with survey response options ranging from 1 (less than once a month) to 6 (four or more times a week). The four academic engagement items included: (a) Work with other students on school work outside of class; (b) Discuss ideas from your readings or classes with students outside of class; (c) Discuss ideas from your readings or classes with faculty outside of class; and (d) Work harder than you thought to meet an instructor’s expectations.

The additional four-item scale constructed to measure student social/com- munity engagement contained response options ranging from 1 (never) to 5 (very often), and included: (a) Participate in events sponsored by a fraternity

Table 2

facToR loadings and ReliabiliTy foR scaled MeasuRes of engageMenT

Factor Reliability Loading (Alpha)

Academic Engagement 0.75 Work with other students on school work outside of class 0.779 Discuss ideas from readings or classes with students outside of class 0.814 Discuss ideas from readings or classes with faculty outside of class 0.767 Work harder than thought to meet an instructor’s expectations 0.549

Social Engagement 0.78 Participate in events sponsored by a fraternity or sorority 0.627 Participate in residence hall activities 0.632 Participate in events by groups reflecting own cultural heritage 0.748 Participate in community service activities 0.753

Hu & Wolniak / Student Engagement and Early Career Earnings 221

or sorority, (b) Participate in residence hall activities, (c) Participate in events or activities sponsored by groups reflecting your own cultural heritage, and (d) Participate in community service activities. Alpha reliabilities ranged from .75 for academic engagement scale to .78 for social engagement scale, indicating acceptable psychometric property of scale reliability.

analysis

Using multiple regression techniques, our analysis consisted of two stages that addressed whether the effects of student engagement on early career earnings are conditional on selected student characteristics. Across all analyses, we applied a variation on the standard log-linear Mincerian (Mincer, 1974) earnings function which corrects for positively skewed earnings distributions and allows unstandardized regression coefficients to approximate percent of differences, or proportional changes, in earnings due to incremental changes in predictor variables (Björklund & Kjellström, 2002; Rosenfeld & Kalleberg, 1990).

Our first analytic stage was a preliminary examination of whether an- nual labor market earnings were influenced by interactions between levels of engagement (academic and social) and gender, racial/ethnic minority group, and academic preparation (based on SAT/ACT scores). In this step, we regressed the natural log of annual earnings on student background characteristics, major fields, institutional characteristics, and scaled engage- ment measures, followed by interaction (cross-product) terms. Following Pedhauzer’s (1982) well-established approach, if the addition of an inter- action term simultaneously improved model fit (a statistically significant increase in R2) while also having a significant net effect, we then moved on to the second analytic stage in which we disaggregated our sample by gender, race/ethnicity, and level of precollege academic preparation. We then used disaggregated subpopulations to estimate earnings models for different groups. When one of the background variables was used to disaggregate the sample, all other measures of student characteristics remained in the model as control variables.

It is worth noting that the results from the use of interaction (cross- product) terms are informative on what variables significantly moderate the relationship between student engagement and early career earnings, while the findings from the disaggregated sample can tell whether scaled student engagement measures are significantly related to earnings for dif- ferent student subpopulations. Together, these analyses illustrate a more comprehensive picture about the relative role of engagement measures (from the general model with interaction terms) and direct contributions of such engagement on early career earnings across different populations (from the conditional model).

222 The Review of higheR educaTion Winter 2013

resulTs

We began our analysis by examining the mean values of scaled measures of academic and social engagement, as well as annual income across the overall analytic sample (N = 1,278) by gender, race/ethnicity, and categories of precollege academic preparation. (Additional descriptive statistics appear in the Appendix.) Results from this descriptive analysis are shown in Table 3, indicating several important trends to consider when interpreting the multivariate results. In particular, we identified statistically significant gender differences in the mean values of academic engagement, social engagement, and annual income. Compared to their female counterparts, males were, on average, more engaged academically and enjoyed higher annual earnings in the labor market ($31,101 vs. $25,822) but were less socially engaged.

In terms of racial/ethnic minority group, we found statistically significant differences in average levels of both academic and social engagement, while annual income did not differ significantly by racial/ethnic groups. Across all racial/ethnic minority groups, Hispanics reported the highest average levels of engagement in academic activities, and African Americans reported the highest levels of social engagement. A pattern that prompts concern emerged among American Indians in our sample who showed low levels of both aca- demic and social engagement. With respect to academic preparation, there were significant differences in social engagement and annual income, but not in academic engagement. Interestingly, students in the high-SAT/ACT group had the lowest levels of engagement in social activities and students in the low-SAT/ACT group had the lowest early career earnings ($23,178 vs. $28,367 for the middle-SAT/ACT group and $27,917 for the high-SAT/ACT group).

Building on the examination of group means, our first stage of analysis involved running multiple regressions with interaction (cross-product) terms to establish evidence of group differences in the relationship between mea- sures of engagement and log annual earnings. The results from this analytic stage indicate that the inclusion of interaction terms significantly improved the variance explained in the regression model and that several interaction terms were statistically significant. Specifically, gender, race/ethnicity, and academic preparation significantly moderated the relationship between student engagement and early career earnings (Baron & Kenny, 1986). These results provided the empirical basis for disaggregating the sample by gender, racial/ethnic minority group, and precollege levels of academic preparation.

Running separate multiple regression models on disaggregated samples based on students’ gender, race/ethnicity, and academic preparation yielded evidence that academic engagement had a positive net effect on early career earnings among males, but no net effect among females. (See Table 4.) Social engagement positively influenced subsequent labor market earnings for fe- males but had no discernable effect for males in the presence of the model’s

Hu & Wolniak / Student Engagement and Early Career Earnings 223

T a

b l

e 3

T h

e s

c a

l e

d M

e a

su R

e s

o f e

n g

a g

e M

e n

T a

n d

e a

R ly

c a

R e

e R a

n n

u a

l e

a R

n in

g s

ca d

em ic

E n

ga ge

m en

oc ia

l E n

ga ge

m en

L og

E ar

n in

gs (

$ 2 00

N (

% )

M ea

.D .

M ea

.D .

M ea

.D .

F u

ll S

am p

1 ,2

7 8

( 1

0 0

% )

1 5

.4 7

4 .5

2 4

1 .0

0 8

.6 3

1 0

.2 0

7 (

$ 2

7 ,0

9 2

) 0

.6 0

8 G

en d

al e

3 3

1 (

2 5

.9 %

) 1

6 .8

8 2

.9 2

1 0

.5 6

2 .8

1 4

0 .3

4 5

( $

3 1

,1 0

1 )

0 .6

1 5

em al

e 9

4 7

( 7

4 .1

% )

1 4

.9 8

4 .6

1 8

1 .1

6 4

.8 7

1 0

.1 5

9 (

$ 2

5 ,8

2 2

) 0

.5 9

Si g.

** *

R ac

e/ E

th n

ic it

A fr

ic an

A m

er ic

4 7

8 (

3 7

.4 %

) 1

5 .4

9 2

.8 0

1 1

.8 9

3 .5

7 5

0 .1

9 5

( $

2 6

,7 6

9 )

0 .5

2 7

m er

ic an

I n

d ia

1 0

0 (

7 .8

% )

1 4

.1 6

4 .6

2 3

.8 6

2 .4

0 3

0 .1

0 8

( $

2 4

,5 3

8 )

0 .6

7 2

is p

an ic

2 4

( 2

5 .4

% )

1 5

.7 8

4 .4

2 0

0 .8

9 0

.6 8

1 0

.1 8

1 (

$ 2

6 ,3

9 6

) 0

.5 2

A si

an A

m er

ic an

7 6

( 2

9 .4

% )

1 5

.5 3

4 .1

6 6

0 .5

5 5

.6 3

1 0

.2 7

0 (

$ 2

8 ,8

5 3

) 0

.7 3

S i g.

** *

A ca

d em

ic P

re p

ar at

io n

L o

w S

A T

/A C

2 6

3 (

2 0

.6 %

) 1

5 .7

3 5

.4 3

1 1

.5 0

4 .5

3 5

0 .0

5 1

( $

2 3

,1 7

8 )

0 .4

0 4

id d

le S

A T

/A C

6 5

2 (

5 1

.0 %

) 1

5 .4

3 8

.7 2

1 1

.2 0

3 .4

1 3

0 .2

5 3

( $

2 8

,3 6

7 )

0 .6

9 0

ig h

S A

T /A

C T

6 3

( 2

8 .4

% )

1 5

.3 5

4 .2

2 2

0 .2

9 4

.1 8

1 0

.2 3

7 (

$ 2

7 ,9

1 7

) 0

.5 5

Si g.

** *

N ot

e: S

ta ti

st ic

al s

ig n

if ic

an ce

o f

gr o

u p

m ea

n d

if fe

re n

ce s:

* p

< 0

.0 5

, * *p

< 0

.0 1

, * **

p <

0 .0

0 1

224 The Review of higheR educaTion Winter 2013

T a

b l

e 4

T o

T a

l a

n d

c o

n d

iT io

n a

l M

o d

e l e

sT iM

a T

e s

o f T

h e R

e l

a T

io n

sh ip

e T

w e

e n

g a

g e

M e

n T a

n d

a n

n u

a l e

a R

n in

g s

(l o

g $

2 0

0 6 )

st im

at ed

E ff

ec t

E st

im at

ed E

ff ec

t of

ca d

em ic

E n

ga ge

m en

oc ia

l E n

ga ge

m en

t T

ot al

a n

d C

on d

it io

n al

M od

el s

od el

R 2

B et

et a

T ot

al S

am p

le P

op u

la ti

on (

N =

1 ,2

7 8

) 0

.0 9

4 **

* 0

.0 0

0 .0

3 6

0 .0

1 9

.1 1

** *

G en

d er

C on

d it

io n

al M

od el

em al

e (n

= 9

4 7

) 0

.0 7

2 **

* -0

.0 0

-0 .0

1 6

0 .0

2 4

.1 5

** *

al e

(n =

3 3

1 )

0 .2

7 1

** *

0 .0

4 8

.3 0

** *

# #

-0

.0 2

-0 .1

1 3

# #

R ac

e/ E

th n

ic it

y C

on d

it io

n al

M od

el s

A fr

ic an

A m

er ic

an (

n =

4 7

5 )

0 .2

2 2

** *

0 .0

1 1

.0 9

.0 1

0 .0

9 6

*†

m er

ic an

I n

d ia

n (

n =

1 0

0 )

0 .5

9 1

** *

0 .0

7 9

.5 4

** *†

-0

.1 7

-0 .6

1 6

*† †

†

is p

an ic

( n

= 3

2 4

)

0 .2

1 2

** *

0 .0

2 4

.2 0

** *

0 .0

0 7

.0 5

†

si an

A m

er ic

an (

n =

3 7

6 )

0 .1

7 6

** *

0 .0

1 2

.0 6

.0 8

0 .4

2 4

S A

T /A

C T

C on

d it

io n

al M

od el

o w

( n

= 2

6 3

) 0

.2 7

0 **

* -0

.0 3

-0 .4

2 7

*^ ^

-0 .0

0 4

-0

.0 5

id d

le (

n =

6 5

2 )

0 .1

1 4

** *

0 .0

2 5

.1 7

** *

0 .0

1 9

.0 9

H ig

h (

n =

3 6

3 )

0 .2

0 0

** *

0 .0

0 4

.0 2

0 .0

1 9

.1 0

N ot

e: I

n a

d d

it io

n t

o t

h e

ac ad

em ic

a n

d s

o ci

al e

n ga

ge m

en t

sc al

es ,

th e

re gr

es si

o n

m o

d el

s w

er e

sp ec

if ie

d t

o c

o n

ta in

t h

e va

ri ab

le s

sh o

w n

i n

T ab

le 1

. *p

< 0

.0 5

, **

p <

0 .0

1 ,

** *p

< 0

.0 0

1 # E

st im

at ed

e ff

ec t

is s

ig n

if ic

an tl

y d

if fe

re n

t fr

o m

f em

al es

, # p

< 0

.0 5

, # # p

< 0

.0 1

, # #

# p

< 0

.0 0

1 .

† E

st im

at ed

e ff

ec t

is s

ig n

if ic

an tl

y d

if fe

re n

t fr

o m

A si

an A

m er

ic an

s, †

p <

0 .0

5 , †

† p

< 0

.0 1

, † †

† p

< 0

.0 0

1 .

^ E

st im

at ed

e ff

ec t

is s

ig n

if ic

an tl

y d

if fe

re n

t fr

o m

t h

e m

id d

le S

A T

/A C

T a

ch ie

ve rs

, ^ p

< 0

.0 5

, ^ ^ p

< 0

.0 1

, ^ ^

^ p

< 0

.0 0

1 .

Hu & Wolniak / Student Engagement and Early Career Earnings 225

control variables. This finding points to existing gender differences in the impact of academic and social engagement on early career earnings, where academic engagement appears to have greater explanatory power among men and social engagement has greater explanatory power among women.

The results of examining disaggregated samples based on racial/ethnic minority groups show that academic engagement has a positive and statisti- cally significant impact on earnings among American Indian and Hispanic students, but no net effect among African Americans or Asian Americans. Social engagement had a positive and significant impact on earnings among African American and Asian American students. Alternatively, all other fac- tors being equal, American Indian students who were more socially engaged during college experienced lower annual earnings once they were in the labor market, such that social engagement had a negative effect on earnings among American Indians. Academic engagement had a significant influence on the earnings of American Indian students, but not for their Asian American or African American counterparts. Social engagement had a positive role in determining early career earnings among African American and an even more pronounced positive influence among Asian Americans.

Focusing on the models run on samples disaggregated by precollege aca- demic preparation, we found that academic engagement and social engage- ment were both positive predictors of early career earnings for students in the middle-SAT/ACT group, while social engagement had a positive influence on earnings among the most academically prepared students (e.g., those in the high-SAT/ACT group). Interestingly, after controlling for all the other variables in the model, academic engagement was significantly and negatively associated with early career earnings for the least academically prepared stu- dents (e.g., the low-SAT/ACT group). Thus, it appears that academic engage- ment has a negative influence on earnings among the low-SAT/ACT group, a positive influence among middle-SAT/ACT scorers, and an insignificant influence among students who achieved the highest SAT/ACT scores. Social engagement did not have a substantially different role for students in three groups but was statistically significant for students in high-SAT/ACT group.

liMiTaTions

There are four limitations in this study. First, because the outcome in this study is early career earnings, results may differ when considering longer- term earnings; subsequent follow-up surveys by the Gates Foundation could facilitate examination of longer-term socio-economic outcomes.

Second, the data did not contain sufficient information to differentiate the employment type of college graduates, which can be an important indicator of career attainment and earnings for college graduates.

226 The Review of higheR educaTion Winter 2013

Third, even though we included a number of institution-level variables (e.g., institutional selectivity, control, etc.) as covariates in our multivariate models, it is possible that additional institutional characteristics are con- founding the results and could be added to the analysis.

Finally, earnings models and labor market studies suggest controlling for local labor market conditions based on geographic regions and occupations. While we recognize the importance of such measures in modeling earnings, our interest in this study was mainly on the effects of engagement on early career earnings. To examine the overall influences of student engagement on early career earnings while maintaining the parsimony of our models, we decided not to include additional state or regional variables in our analyses. When interpreting the results of our analyses, it is important to consider that student engagement in college may have influenced students’ occupational choices and in turn affect their earning power in the labor market.

It is also worthwhile to consider the generalizability of findings from this study to the broad higher education sector. The GMS scholars are a group of high-achieving low-income minority students selected according to the criteria determined by the Gates Foundation, and the nonrecipients were also applicants to the GMS programs with comparable qualifications. Scholars and nonrecipient applicants are therefore different from the general college student population. However, the racial/ethnic composition of our sample approximately reflects the national population of students of color who graduated from four-year institutions within a six-year time frame. Based on the national composition of undergraduates who enrolled in a four-year institution and who graduated within six years of entering college across all racial/ethnic groups, 57% were female and 43% were male. Among non- White graduates, 36% were classified as Black non-Hispanic, 29% Hispanic, 34% Asian or Pacific Islander, and 1% American Indian or Alaska Native. Thus, while it is important to recognize that the analytic sample used in this study is disproportionally female in comparison to the broader postsecondary system, the racial/ethnic make-up mirrors the national composition of college graduates (IPEDS, 2011). We hope this study can start a conversation on the long-term effects of the college experience and that more datasets containing rich information on student engagement and labor market outcomes will emerge to further strengthen inquiry in this direction.

conclusions and iMPlicaTions

The results from this study show that student academic and social en- gagement play distinct roles in determining the early career earnings that are conditional on students’ gender, race/ethnicity, and precollege academic preparation. At least two conclusions stem from these findings.

Hu & Wolniak / Student Engagement and Early Career Earnings 227

First, it appears that the effects of student engagement on earnings in the early stages of students’ careers depend on the nature or type of engagement in which students participate. In other words, engagement in the form of academic activities influences earnings differently than engagement in social activities. Researchers have long conceptualized that students in college must navigate both the academic and social systems in order to achieve successful educational outcomes (Tinto, 1993). Our findings substantiate the distinc- tive role of student engagement across both academic and social systems as reflected in the influence that both types of activities exert on earnings in the years immediately following college. It is therefore critical to differentiate academic engagement and social engagement when explaining career and labor market outcomes.

Second, the effects of academic engagement and social engagement vary by students’ gender, race/ethnicity, and academic preparations. Given the ever-growing body of evidence showing that college impact models dif- fer across a range of individual background characteristics (Pascarella & Terenzini, 2001, 2005), our findings add support for examining conditional models when explaining post-college outcomes. In particular, our findings indicate that the earnings effects of engagement are conditional on gender, such that men disproportionately experience an earnings benefit from being academically engaged during college while women uniquely benefit from social engagement during college.

Among students with lower levels of academic preparation, engage- ment during college appears to have little influence on early career earn- ings. This finding marks a dramatic deviation from past research using non-labor market outcomes (Carini, Kuh, & Klein, 2006; Kuh et al, 2009), where students with the lowest SAT scores were found to benefit more from student engagement than those with higher SAT scores. As the researchers acknowledged (Carini, Kuh, & Klein, 2006), it may be that a ceiling effect confounded previous studies based on student GPAs or self-reported learn- ing outcomes, whereas earnings in labor market do not have that limitation. Nevertheless, this is an interesting contrast that warrants recognition and further investigation.

In summary, by focusing on differences in students’ background char- acteristics, our study adds to the field’s understanding of the influence that academic engagement and social engagement have on early career earnings. As past studies have demonstrated (Cabrera, Nora, Terenzini, Pascarella, & Hagedorn, 1999; Nora, Cabrera, Hagedorn, & Pascarella, 1996; Pascarella & Terenzini, 2005; Reason, Terenzini, & Domingo, 2006), characteristics such as gender, race/ethnicity, and other ascribed and achieved qualities can moderate the impacts of college, and the effects of college experiences are often condi- tional rather than general. It is increasingly important for educational policy

228 The Review of higheR educaTion Winter 2013

and practice to take into consideration the diversity of today’s college students so that programs and interventions may be designed in a manner that may translate desirable program effects to students of different backgrounds. It appears that this concept holds true for labor market outcomes in the years immediately following college. As researchers put forth more efforts toward understanding the economic implications of student engagement (Hu & Wolniak, 2010), it is increasingly important to consider the extent to which student engagement is conditional on specific student characteristics as a way to understand how distinct facets of the college experience may serve to compensate for, or possibly reinforce, precollege differences. The kinds of conditional effects uncovered in this study may contribute to an improved understanding of student engagement from both human capital and social capital perspectives.

Students identify the ability to make more money and the ability to get a better job as the two most important issues in their decision to enter col- lege (Astin, 1993), and it is essential to recognize that labor market earnings provide a signal of occupational skill development. However, earnings are only one of many desirable college outcomes that students and the society as a whole are concerned about. We should therefore also consider the role of student academic and social engagement on other desirable outcomes when shaping educational policy and practice. With future studies, we en- courage scholars to consider the underlying mechanisms by which student engagement may influence the formation of career dispositions, occupational preferences, job searching, and vocational decision-making.

Implications for future research include building on our findings by examining how academic and social dimensions of student engagement may contribute to the accumulation of human capital and social capital across the broad college-going population and how students from differ- ent backgrounds use such forms of capital when making career decisions. For instance, our findings suggest that American Indian students who are more engaged in social activities tend to earn less in the labor market after college graduation. It is possible that such students who are more socially engaged on campus are more committed to social issues and thus choose careers with more service orientation but lower financial returns. While our data did not support an exploration of this possibility, it is a topic worthy of further investigation.

Concerns over the quality of undergraduate education occupy a promi- nent role in the current educational and political dialogue, and emerging research empirically questions the amount of learning that is taking place on campuses nationwide (Arum & Roksa, 2011; Pascarella, Blaich, Martin, & Hanson, 2011). Together these trends are fueling calls for accountability in higher education and demanding that our colleges and universities provide

Hu & Wolniak / Student Engagement and Early Career Earnings 229

all students with a quality education that translates into productive careers and lasting professional development.

Labor market earning is one measure of occupational productivity and economic success, and our findings contribute new evidence on the broad value of college student engagement across diverse student populations. The relationships between student engagement and early career earnings that we have demonstrated among students of color—and the implications these findings have for research, policy, and practice—may contribute to the de- velopment of social theories that explain racial/ethnic differences. Ultimately, this study provides new evidence to improve our broad understanding of how the academic and social engagement of students of color translates into economic success once they enter the labor market. It may also contribute to the dialogue surrounding the broad value of the college student experience.

230 The Review of higheR educaTion Winter 2013

a p

p e

n d

sa M

p l

e c

o M

p o

si T

io n

b y g

e n

d e

R , R

a c

e /e

T h

n ic

iT y , a

n d

a c

a d

e M

ic p

R e

pa R

a T

io n

er ce

n t

G en

d er

er ce

n t

R ac

e/ E

th n

ic it

er ce

n t

A ca

d em

ic P

re p

ar at

io n

F u

M en

om en

fr ic

A m

er ic

is p

an ic

si an

M id

d le

ig h

ar ia

bl es

am p

A m

er ic

d ia

A m

er ic

G M

S R

ec ip

ie n

2 5

.2 %

7 .9

2 4

.3 %

1 .3

5 7

.3 %

3 .3

1 4

.8 %

4 .3

2 1

.9 %

1 .8

% M

al e

(f em

al e

= 0

) 2

5 .9

1 0

0 .0

1 5

.2 %

8 .9

3 3

.7 %

2 .0

1 2

.0 %

6 .7

3 4

.6 %

A fr

ic an

A m

er ic

3 7

.4 %

1 .9

4 2

.8 %

0 0

.0 %

2 .3

2 9

.1 %

2 .3

% A

m er

ic an

I n

d ia

7 .8

8 .8

7 .5

0 .0

1 0

0 .0

1 3

.4 %

.1 %

.3 %

H is

p an

2 5

.3 %

3 .0

2 2

.7 %

.0 %

0 0

.0 %

3 .6

2 4

.8 %

7 .5

% A

si an

A m

er ic

2 9

.5 %

6 .3

4 4

.4 %

.0 %

0 0

.0 %

0 .7

3 8

.0 %

6 .9

% P

ar en

ta l

ed u

ca ti

o n

o f

b ac

h el

o r

ab o

3 1

.5 %

7 .4

3 2

.9 %

1 .4

4 6

.4 %

2 .7

3 5

.3 %

2 .6

2 7

.6 %

7 .8

% L

o w

-S A

T /A

C T

0 .6

9 .5

2 4

.5 %

3 .7

3 5

.1 %

9 .2

1 4

.0 %

0 0

.0 %

M id

d le

-S A

T /A

C T

1 .0

5 2

.6 %

0 .4

6 6

.9 %

2 .9

4 9

.9 %

1 .2

0 .0

1 0

0 .0

% H

ig h

-S A

T /A

C T

8 .4

3 7

.9 %

5 .1

9 .4

1 2

.0 %

0 .9

5 4

.8 %

.0 %

0 0

.0 %

L o

w s

el ec

ti vi

1 0

.5 %

1 .7

1 0

.1 %

4 .8

1 3

.3 %

0 .0

4 .8

2 8

.2 %

.8 %

.5 %

M id

d le

s el

ec ti

vi ty

5 .5

4 3

.6 %

6 .1

5 7

.3 %

8 .4

4 0

.9 %

0 .9

6 1

.7 %

4 .3

1 7

.9 %

H ig

h s

el ec

ti vi

4 4

.0 %

4 .7

4 3

.8 %

7 .9

2 8

.3 %

9 .1

6 4

.3 %

0 .1

3 7

.9 %

9 .6

% P

u b

li c

in st

it u

ti o

ri va

te i

n st

it u

ti o

n =

0 )

5 6

.9 %

5 .7

5 3

.4 %

7 .1

6 7

.8 %

8 .9

6 0

.5 %

6 .9

5 4

.8 %

5 .9

% ST

E M

m aj

o r

o n

-S T

E M

= 0

) 3

7 .3

4 9

.4 %

3 .0

3 9

.5 %

2 .5

2 6

.9 %

7 .3

3 7

.1 %

6 .0

3 9

.7 %

T ot

1 0

0 .0

2 5

.9 %

4 .1

3 7

.4 %

.8 %

5 .4

2 9

.4 %

0 .6

5 1

.0 %

8 .4

(N =

1 ,2

7 8

) (n

= 3

3 1

) (n

= 9

4 7

) (n

= 4

7 8

) (n

= 1

0 0

) (n

= 3

2 4

) (n

= 3

7 6

) (n

= 2

6 3

) (n

= 6

5 2

) (n

= 3

6 3

)

Hu & Wolniak / Student Engagement and Early Career Earnings 231

references

Arum, R., & Roksa, J. (2011). Academically adrift: Limited learning on college campuses. Chicago: University of Chicago Press.

Astin, A. W. (1993). What matters in college: Four critical years revisited. San Fran- cisco: Jossey-Bass.

Astin, A. W. (1999). Student involvement: A developmental theory for higher educa- tion. Journal of College Student Development, 40, 518–529.

Baron, R., & Kenny, D. (1986). The moderator-mediator distinction in social psycho- logical research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Barron’s Profiles of American Colleges. (2001). Hauppage, NY: Barron’s Educational Series.

Becker, G. S. (1994). Human capital: A theoretical and empirical analysis with special reference to education (3rd ed.). Chicago: University of Chicago Press.

Björklund, A., & Kjellström, C. (2002). Estimating the return to investments in education: How useful is the standard Mincerian equation? Economics of Education Review, 21, 195–210.

Cabrera, A. F., Nora, A., Terenzini, P. T., Pascarella, E. T., & Hagedorn, L. S. (1999). Campus racial climates and the adjustment of students to college: A compari- son between White students and African American students. Journal of Higher Education, 70, 134–160.

Carini, R. M., Kuh, G. D., & Klein, S. P. (2006). Student engagement and student learning: Testing the linkages. Research in Higher Education, 47, 1–32.

Chickering, A. W., & Gamson, Z. (1987). Seven principles of good practice in under- graduate education. American Association of Higher Education Bulletin, 39, 3–7.

Coleman, J. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, S95–S120.

Coleman, J., Hoffer, T., & Kilgore, S. (1982). High school achievement: Catholic, public, and private schools compared. New York: Basic Books.

Hallinan, M. T. (1982). The peer influence process. Studies in Educational Evalua- tion, 7, 285–306.

Hearn, J. C. (1988). Attendance at higher-cost colleges: Ascribed, socioeconomic, and academic influences on student enrollment patterns. Economics of Education Review, 7, 65–76.

Hearn, J. C. (1991). Academic and nonacademic influences on the college destinations of 1980 high school graduates. Sociology of Education, 64, 158–171.

Hu, S., & Kuh, G. D. (2003). Maximizing what students get out of college: Testing a learning productivity model. Journal of College Student Development, 44, 185–203.

Hu, S., & Wolniak, G. (2010). Initial evidence of the influence of college student en- gagement on early career earnings. Research in Higher Education, 51, 750–766.

IPEDS Data Center. (2011). Retrieved on March 30, 2011, from http://nces.ed.gov/ ipeds/datacenter).

Karen, D. (1991). The politics of class, race, and gender: Access to higher education in the United States, 1960–1986. American Journal of Education, 99, 208–237.

232 The Review of higheR educaTion Winter 2013

Kim, Y. K., & Sax, L. J. (2009). Student-faculty interaction in research universities: Differences by student gender, race, social class, and first-generation status. Research in Higher Education, 50, 437–459.

Knox, W. E., Lindsay, P., & Kolb, M. N. (1993). Does college make a difference? Long- term changes in activities and attitudes. Westport, CT: Greenwood Press.

Kuh, G. D., Cruce, T. M., Shoup, R., Kinzie, J., & Gonyea, R. M. (2008). Unmasking the effects of student engagement on first-year college grades and persistence. Journal of Higher Education, 79, 540–563.

Kuo, V., Wilds, D., Garcia, P. B., Trent, W. T., Nichols, L., Bronwyn, Z., et al. (2006). The Gates Millennium Scholars research program: A conceptual framework. Chicago: National Opinion Research Center at the University of Chicago.

Lodato Nichols, B., Zimowski, M., Lodato, R. M., & Ghadialy, R. (2004). The survey of diverse students. In E. P. St. John (Ed.), Improving access and college success for diverse students: Studies of the Gates Millennium Scholars Program (pp. 23–44). Readings on equal education, Vol. 20. New York: AMS Press.

Lin, N. (1999). Social networks and status attainment. Annual Review of Sociology, 25, 467–487.

Mincer, J. (1974). Schooling, experience, and earnings. New York: National Bureau of Economic Research.

National Commission on the Future of Higher Education (2006). A test of leadership: Charting the future of U.S. higher education. Washington, DC: US Department of Education.

National Opinion Research Center (2003). Gates Millennium Scholars tracking and longitudinal study: Year 1 final report. Chicago: The National Opinion Research Center (NORC) at the University of Chicago.

National Opinion Research Center (2006). The inaugural cohort of the Gates Mil- lennium Scholars: The first few years after high school. Chicago: The National Opinion Research Center (NORC) at the University of Chicago.

National Survey of Student Engagement. (2004). Student engagement: Pathways to collegiate success. Bloomington: Indiana University Center for Postsecondary Research.

National Survey of Student Engagement. (2005). Student engagement: Exploring different dimensions of student engagement. Bloomington: Indiana University Center for Postsecondary Research.

Nora, A., Cabrera, A. F., Hagedorn, L., & Pascarella, E. T. (1996). Differential impacts of academic and social experiences on college-related behavioral outcomes across different ethnic and gender groups at four-year institutions. Research in Higher Education, 37, 427–452.

NORC. See National Opinion Research Center. Pace, R. C. (1979). Measuring outcomes of college. San Francisco: Jossey-Bass. Pascarella, E. T. (1985). College environmental influences on learning and cognitive

development: A critical review and synthesis. In J. C. Smart (Ed.), Higher educa- tion: Handbook of theory and research (Vol. 1., pp. 1–62). New York: Agathon.

Pascarella, E. T., Blaich, C., Martin, G. L., & Hanson, J. M. (2011). How robust are the findings of Academically Adrift? Change: The Magazine of Higher Learn- ing, 43, 20–24.

Hu & Wolniak / Student Engagement and Early Career Earnings 233

Pascarella, E. T., & Terenzini, P. T. (1991). How college affects students: Findings and insights from twenty years of research. San Francisco: Jossey-Bass.

Pascarella, E. T., & Terenzini, P. T. (2005). How college affects students: A third decade of research. San Francisco: Jossey-Bass.

Paulsen, M. B. (2001). The economics of human capital and investment in higher education. In M. B. Paulsen & J. C. Smart (Eds.), The finance of higher educa- tion: Theory, research, policy, and practice (pp. 55–94). New York: Agathon Press.

Pedhazur, E. J. (1982). Multiple regression in behavioral research: Explanation and prediction. New York: Holt, Rinehart and Winston.

Perna, L. W. (2005). The benefits of higher education: Sex, racial/ethnic, and socio- economic group differences. Review of Higher Education, 29(1), 23–52.

Reason, R. D., Terenzini, P. T., & Domingo, R. J. (2006). First things first: Developing academic competence in the first year of college. Research in Higher Educa- tion, 47, 149–175.

Rendón, L. I., Jalomo, R. E., & Nora, A. (2000). Theoretical considerations in the study of minority student retention. In J. Braxton (Ed.), Rethinking the student departure puzzle: New theory and research on college student retention (pp. 127–156). Nashville, TN: Vanderbilt University Press.

Rosenfeld, R. A., & Kalleberg, A. L. (1990). A cross-national comparison of the gender gap in income. American Journal of Sociology, 96, 69–106.

Sax, L. J., Bryant, A. N., & Harper, C. E. (2005). The differential effects of student- faculty interaction on college outcomes for women and men. Journal of College Student Development, 46, 642–659.

Sewell, W. H., Haller, A. O., & Ohlendorf, G. W. (1970). The educational and early occupational status attainment process: Replication and revision. American Sociological Review, 35, 1014–1027.

Sewell, W. H., Haller, A. O., & Portes, A. (1969). The educational and early occupa- tional attainment process. American Sociological Review, 34, 82–92.

Thomas, S. (2000). Deferred costs and economic returns to college major, quality, and performance. Research in Higher Education, 41, 281–313.

Thomas, S. (2003). Longer-term economic effects of college selectivity and control. Research in Higher Education, 44, 263–299.

Tienda, M., & Lii, D. (1987). Minority concentration and earnings inequality: Blacks, Hispanics, and Asians compared. American Journal of Sociology, 93, 141–165.

Tinto, V. (1975). Drop out from higher education: A theoretical synthesis of recent research. Review of Educational Research, 63, 89–125.

Tinto, V. (1993). Leaving college: Rethinking causes and cures of student attrition (2nd ed.). Chicago: The University of Chicago Press.

Weidman, J. C. (1989). Undergraduate socialization: A conceptual approach. In J. C. Smart (Ed.), Higher education: Handbook of theory and research (pp. 289–322). New York: Agathon Press.