2-Pane Combined
Full Summaries Sorted

Responsibility or Liability? Student Loan Debt and Time Use in College


Student Loan Debt and Time Use in College

Responsibility or Liability? Student Loan Debt and Time Use in College
Natasha Yurk Quadlin and Daniel Rudel, Indiana University

Recent research shows that student loans may affect persistence and completion among undergraduates, but few have examined the association between debt and everyday college experiences. We use data from the National Longitudinal Survey of Freshmen to assess the relationship between student loan debt and the foun- dation of campus life—time use. On a basic level, we find that debt is related to stu- dents’ time spent on activities such as working for pay, consuming media, and athletics. We then focus on results from a latent class analysis, which suggest that college activ- ities coalesce into three distinct student lifestyles that are differentially associated with debt. Although many debt-free students develop active non-academic agendas in the “Socially Engaged” lifestyle, indebted students may exhibit one of two divergent types: the “Serious Students,” who spend their time on academics and working for pay, or the “Inactive,” who are comparatively uninvolved in campus life. These patterns suggest that either student loan debt is a stratifying mechanism for students’ college experi- ences, or that it acts as a proxy for students’ selection into particular time use patterns.

Student loan debt has become a pressing and pervasive social issue. In the past several decades, cumulative student debt has grown steadily, such that it is now the most substantial type of non-mortgage debt in the United States. Nearly 40 million Americans have accrued student debt, with 60 percent of all borrowers owing at least $10,000 and another 30 percent owing $25,000 or more. Young adults, in particular, have assumed considerable financial responsibilities, as over 40 percent of all 25-year-olds have student loan debt (Lee 2013). Many college students are taking on substantial monetary burdens in one of the most unstable phases of the life course.

Sociologists are only beginning to examine consequences for indebted students, including effects on persistence and completion (Dwyer, Hodson, and McCloud

The authors would like to extend special thanks to Brian Powell for his exceptional support and feed- back. They also thank Donna Eder, Joe Johnston, J. Scott Long, and the anonymous reviewers for their helpful comments. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-0813962 to the first author. Previous versions of this article were presented at the 2013 American Sociological Association Meeting in New York and the 2014 American Educational Research Association Meeting in Philadelphia. Direct all correspondence to Natasha Yurk Quadlin, Indiana University Department of Sociology, Ballantine Hall 744, 1020 E. Kirkwood Avenue, Bloomington, IN 47405-7103, USA; E-mail:

Student Loan Debt and Time Use in College page1image450136528 page1image450136880 589













© The Author 2015. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail:

Social Forces 94(2) 589–614, December 2015 doi: 10.1093/sf/sov053 Advance Access publication on 5 March 2015


590 page2image458453168 page2image458296304 Social Forces 94(2)

2013; Dwyer, McCloud, and Hodson 2012; Kim 2007). But to date, few have assessed how debt is associated with students’ everyday experiences, or the extent to which they study, take part in extracurricular activities, and socialize with others. These inquiries into the college experience are becoming increasingly important as more Americans pursue higher education (Baird, Burge, and Reynolds 2008).

In this paper, we examine how educational debt is associated with students’ experiences at a group of selective universities in the United States. We use a numerical measure—time use—to understand how debt is related to students’ involvement within and across types of life activities. Although there are a finite number of hours in a day or week, students use their time in vastly different ways. We contend that accruing debt may constrain students’ time use, with implica- tions for how they experience higher education. With these ideas in mind, we ask the following questions:

  1. How is student loan debt associated with time spent on academics, working for pay, partying, and other activities?

  2. What lifestyles (or time use patterns) exist among undergraduates?

  3. How is debt associated with the lifestyles that students exhibit?

Looking at basic time use categories over a one-week period, we find that student loan debt is related to time spent in several categories of time use. Indebted stu- dents spend more time working and less time on athletic activities than their debt-free peers. At the same time, students with more debt spend more time watching television and listening to music, compared to those with less debt. These differences persist net of socioeconomic status (SES) and employment sta- tus. We then focus on how activities cluster into coherent time use patterns. Using a latent variable model, we find that college activities coalesce into three distinct lifestyles: “Serious Students,” who prioritize both academics and working for pay; “Inactive” students, who are relatively uninvolved in campus life; and stu- dents who are “Socially Engaged” by spending time partying, participating in student group activities, playing sports, and attending athletic events.

Finally, we examine how student loan debt is associated with these lifestyles. We find that many debt-free students have time to develop active non-academic agendas as part of the Socially Engaged lifestyle. Indebted students, however, typically belong to one of two divergent categories. Some exhibit the Serious Student lifestyle—a sign that many indebted students may embrace their financial responsibilities, and spend their time making money and amassing the credentials needed to earn gainful employment. Other indebted students exhibit the Inactive lifestyle. For these students, indebtedness may represent a liability that adversely affects their college experience. We argue that educational debt may stratify col- lege students in a way that has received little attention in the literature. Aside from this possibility of a direct effect of student loan debt on time use, we also discuss whether debt acts as a proxy for students’ selection into particular time use patterns. Students who take on debt may have multiple characteristics that also predict their time use. These equally plausible alternative mechanisms are discussed further below.

On top of identifying how indebted students spend their time, our findings are also useful for understanding how they do not spend their time. That is, student


Student Loan Debt and Time Use in College page3image458302352 page3image452786816 591

loan debtors have often been accused of using their loans to bankroll bad habits and bad decisions (Best and Best 2014). Although we cannot refute media accounts of students going into debt to pay for Spring Break (see Sheehy 2013), we find that most indebted students are quite diligent in their time use. They may concentrate their time in activities such as going to class, studying, and working for pay, while eschewing the parties and other activities that one would expect from popular rhetoric. We contend that indebted students’ time use patterns devi- ate from what many have been led to believe.

More broadly speaking, this paper contributes to a growing literature on the sociology of debt (Conley 1999; Dwyer, Hodson, and McCloud 2013; Dwyer, McCloud, and Hodson 2012; Killewald 2013). Debt is a powerful, but risky, financial tool. It allows consumers to access goods and services that they would not otherwise be able to afford, but it also carries substantial costs, including one’s physical and mental health (Drentea and Reynolds 2012; Sweet et al. 2013). Student loan debt is an especially important case in this regard because it is often thought of as “good” debt, facilitating human capital investments that allow students to improve their labor market prospects (Becker 1964). Yet, when stu- dents take on debt to invest in their education, they may incur social and cultural costs that come due long before their monetary costs. Thus, indebted students may have qualitatively different college experiences than others because they are making unique sacrifices for their human capital. We focus on how these differ- ences manifest during the undergraduate years, but also discuss implications for the life course.

Student Loans and the College Experience

Past research has focused on two individual-level consequences of student loan debt. First, debt can help and hinder students’ chances of graduation. Accruing a modest amount of debt may facilitate completion, but students who owe more than $10,000 are less likely to finish their degrees (Dwyer, McCloud, and Hodson 2012). Second, the relationship between debt and college completion is not uni- form within the population. Of those students who drop out without a degree, men do so with less debt than women (Dwyer, Hodson, and McCloud 2013). Black and low-income students are also less likely to complete college if they accrue heavy debt in the first year (Kim 2007). Although these studies help deepen our understanding of college completion for indebted students, they often over- look the everyday occurrences that shape students’ lives.

This omission is not uncommon in the sociology of higher education, as few sociologists have examined the “experiential core of college life” (Stevens, Armstrong, and Arum 2008, 131). But the exceptions, both classic and contem- porary, have shown that college facilitates both personal exploration and the maintenance and emergence of inequality. Formative scholars described student traditions (Cowley and Waller 1935) and identified subcultures that differen- tially engaged with ideas and institutions (Clark and Trow 1966; Terenzini and Pascarella 1977). Recent work corroborates the idea that students can be more or less involved both academically and socially, placing them on different trajec- tories for life after college. Armstrong and Hamilton (2013) describe women’s


592 page4image462314576 page4image462312464 Social Forces 94(2)

pathways through a public university—professional, mobility, and party—and show that time spent in class, studying, working, and partying aggregates over time, shaping outcomes such as college completion and occupational attain- ment. In other words, the everyday choices students make, and the constraints they face, have major implications for the life course (also see Chambliss and Takacs 2014; Grigsby 2009; Moffatt 1989; Nathan 2005). As student popula- tions become even more heterogeneous, additional research is needed to under- stand the various pathways through college life, particularly in the context of rising student debt.

College Inequalities and Consequences of Student Loan Debt: Two Perspectives

Larger traditions of research on college, debt, and emerging adulthood suggest two perspectives on debt and time use. These perspectives are contrasting, but not incompatible with each other, and may apply to different subgroups of indebted students. First, college students may embrace a greater sense of responsibility when they accrue debt. They may spend more time on academics, work, or other pursuits in order to earn gainful employment at graduation (Bodvarsson and Walker 2004; Dwyer, McCloud, and Hodson 2011). We refer to this perspective as “Debt as Responsibility.” A second perspective, “Debt as Liability,” suggests that the financial constraints associated with debt may limit students’ opportuni- ties, acting as a stressor (Drentea 2000) and excluding them from certain aspects of college life (Espenshade and Radford 2009).

Before outlining our two perspectives, we acknowledge that most scholars have focused on the role of SES in the college experience, as opposed to the role of debt (see Espenshade and Radford 2009; Stuber 2006). SES and student loan debt are negatively correlated, but not perfectly so. In fact, some studies find that middle-income students accrue the most debt because they do not qualify for financial aid, but their parents cannot afford to pay for college out of pocket (Choy and Berker 2003; Houle 2014). Due to the dearth of research on debt and college experiences, we include references to work on SES and financial aid gener- ally, but note that these mechanisms may not seamlessly apply to the case of student loan debt.

Debt as responsibility

In this first perspective, scholars suggest that indebted students take their education more seriously than those without such burdens. By having some of their own “skin in the game” (that is, by spending some of their own money or taking on their own debt), students may be motivated to make the most of their college experience (Hamilton 2010). This sense of purpose may encourage students to actively reject some activities, such as socializing, in favor of those that will give them the greatest chances of success (Stuber 2009). For example, students without financial support are less likely to fail courses and be placed on academic probation (Bodvarsson and Walker 2004). They may also earn better grades than students with parental fund- ing (Bodvarsson and Walker 2004; Hamilton 2013). These patterns indicate that self-supported students, including those with debt, place special emphasis on


Student Loan Debt and Time Use in College page5image454723008 page5image461910720 593

academic activities. They may exert more effort in the classroom because they have invested some of their own resources in their education.

Furthermore, recent research shows that accruing student loan debt may increase young people’s mastery and self-esteem (Dwyer, McCloud, and Hodson 2011). Indebted students could be organized, motivated, and independent in ways that promote achievement. They must maximize their time in college in order to facilitate opportunities that will allow them to pay down or defer their loans. To this end, indebted students may gravitate toward working or interning for pay. Working not only gives students immediate financial returns, but also provides them with valuable credentials for their careers (Bozick 2007; Pascarella and Terenzini 1991).

Debt as liability

A second line of research implies that indebted students are relatively inactive, both in and out of the classroom. Although this characteristic is not unique to the debt as liability paradigm—indeed, the debt as responsibility perspective also suggests that indebted students are unintegrated in campus social life—it manifests in differ- ent ways and occurs for different reasons. Under debt as responsibility, students may actively abstain from social activities in order to fully devote themselves to academics and work. But when debt is a liability, students’ mental health may suffer, leading them to passively withdraw from formal campus activities.

Hamilton (2013) identifies key patterns in this regard; although self-supported students often have high GPAs, they are less likely to graduate than those with parental support. These students may put a great deal of effort into their studies, but could lack the sense of security needed to sustain their enrollment. The litera- ture on credit card debt also suggests that indebted individuals may experience anxiety or depression, and that this relationship holds in populations of both young and older adults (Drentea 2000; Drentea and Reynolds 2012; Hodson, Dwyer, and Neilson 2014). Thus, debt may heighten the usual pressures of col- lege, subjecting indebted students to stressors and threatening their mental health.

Accruing student loans could also have adverse consequences for students’ social lives. Espenshade and Radford (2009) posit that students on financial aid (specifically those with work-study jobs) are isolated at college. They often hold visible positions on campus, such as library clerks and dining hall workers, which mark them as “underprivileged” to onlookers. They may also make accommoda- tions for their lack of resources, spending time in the computer lab or refusing off-campus meals in favor of the dining hall—spaces that are less ubiquitous among wealthier classmates (Armstrong and Hamilton 2013; Espenshade and Radford 2009). These constraints may prevent indebted students from leading active social lives on campus. They might prefer solitary behaviors, such as watch- ing television or browsing the Internet, to extracurricular and social activities.

These two lines of research—Debt as Responsibility and Debt as Liability— suggest contrasting perspectives as to how debt could be associated with student time use. However, they are not necessarily competing, as they may apply to students with different characteristics or attitudes toward college. We adjudicate between these two lines of thinking, while also considering whether they coexist among college students.


594 page6image459919776 page6image459928544 Social Forces 94(2)
The Case of Time Use

Time use measures are well suited for studying individual differences in everyday experiences. Notably, scholars have used time diaries and similar data to assess gender inequalities in household labor (Milkie, Raley, and Bianchi 2009; Sayer, Bianchi, and Robinson 2004; Schneider 2012) and time use among paid workers (Lin 2012). By comparison, variation in college student time use remains rela- tively unexplored. Several surveys have included information on time spent in typical college activities, but most have been limited to students within single colleges (Stinebrickner and Stinebrickner 2004) or state university systems (Brint and Cantwell 2010; Harris and Goldrick-Rab 2012; but see Arum and Roksa 2011; Babcock and Marks 2011).

In an exception, Massey et al. (2006) collected time use data at a sample of selective universities in the United States, focusing on race differences in time use. They find that Black students spend more time on recreation and working for pay, but less time sleeping, than students from other race backgrounds. Time spent on academics and other life-maintenance activities is approximately equal across race groups. Moreover, there are differences in the total number of hours reported. White students report the fewest active hours, suggesting that they spend the least time multitasking and are least often under time pressure (Charles et al. 2009). We use this dataset, the National Longitudinal Survey of Freshmen (NLSF), to examine variations in time use according to a different explanatory variable: student loan debt. Although the original study design centered on minority underperformance in college, the rich measures in the NLSF make it ideal for examining time use across a broad spectrum of student characteristics.

Data, Measures, and Methods


Baseline interviews for the NLSF were conducted in person in the fall of 1999, when students (N = 3,924) were in their first semester of college. Follow-ups were conducted by phone every spring through the fourth year, when most students were preparing to graduate. The NLSF sample has two unique features, both of which have implications for this study. First, the 28 institutions (9 liberal arts, 14 private research, 4 public research, and 1 historically black college) were primarily drawn from the College and Beyond Survey, the basis for Bowen and Bok’s (1998)landmark study of race-based affirmative action. These institutions are described as “selective,” meaning they have low acceptance rates (40 percent average in the base year) and high admission standards (1243 average SAT in the base year) (Massey et al. 2006). As a result, the students in the NLSF are generally high- achieving and privileged, which could translate to elevated time spent on academ- ics, extracurriculars, and other activities. Time spent in these categories is likely lower in the general population.

Second, the sample contains approximately equal numbers of white, Black, Asian, and Hispanic students, so it is not racially representative of the college-going population. The implications of this aspect of the design are more


Student Loan Debt and Time Use in College page7image454736816 page7image454735296 595

difficult to anticipate. In one sense, we know that time use patterns for these groups vary to some extent, as discussed above. But at the same time, little research has addressed race differences in dispositions toward debt, so it is unclear whether the racial composition of the sample affects our results. We do not engage race specifically in this study, but the sample obliges us to consider whether a more typical group of college students would behave differently. Fur- ther implications of the sampling design are included in the discussion.

We use time use data from the first year, for several reasons. First, past research on the college experience has also focused on the first year, citing its importance for college performance and completion (Bozick 2007; Kim 2007). Second, endogeneity concerns are minimized in the first year because students are typically offered financial aid packages before they enroll. Student loan debt is thus determined before time use patterns emerge, as opposed to in later years, when time spent on academics or extracurriculars could influence eligibility for financial aid. Finally, the first-year data present the most conservative estimates of debt’s association with time use. The first year of college is a time of major personal exploration (Nathan 2005), so differences in time use may be mini- mized while students are trying to have a well-rounded experience. Also, most first-year students are at least four years away from entering loan repayment, and the cognitive burdens of student loans should be relatively minor here. Any patterns we find in the first year could be magnified as students progress through college.

Measures of Dependent and Independent Variables

Time use

For the last full week of classes in the spring semester or quarter, students were asked to estimate the number of hours they spent doing several activities.1 We combined similar activities to yield seven categories of time use: academics (attending class or lab, studying); working for pay; media (watching television, listening to music); athletics (playing or practicing sports, attending a sporting event); attending parties; student groups (doing extracurricular activities, doing volunteer work in the community); and sleeping. These categories are consistent with Charles et al.’s (2009) analysis of race differences in time use.2 Students also made an equivalent estimate for the weekend, and we combined these weekday and weekend measures to generate measures for each time use category for the seven-day period.3 Descriptive statistics for time use and other variables of inter- est are reported in table 1.

Because these time use measures are from a retrospective survey, as opposed to time diaries or an experiential sampling method, they may be subject to recall bias and other inaccuracies. If students could not remember their actual sched- ules, they may have tended toward reporting more socially desirable time use— for example, more time on academics and less time on partying. Accordingly, we have taken care to maximize the robustness of data by bottom-coding at the 1st percentile, top-coding at the 99th percentile, and emphasizing relative time use measures, as discussed below. There is some evidence that the NLSF’s style of questioning, in which students were asked to report activities from a specific


596 page8image458502032 page8image458506128 Social Forces 94(2)
Table 1. Descriptive Statistics for NLSF Analytic Sample, N = 3,676






Dependent variables: Time use (hours)

Sleeping 48.54 11.88 18 80

Academic (class/lab + studying) 46.01 17.73 13 102

Media (watching TV + listening 24.02 21.09 0 119 to music)

Student groups 13.50 11.90 0 58 (extracurriculars + volunteering)

Attending parties 7.24 6.46 0 30

Athletics (playing 6.27 8.51 0 44 sports + attending athletic events)

Working for pay 5.24 8.12 0 39

Key independent variables

Student loan debt (dichotomous)a .45 – 0 1

Student loan debt (continuous, 4,700.46 4,252.68 200 20,000 among indebted)a


Grantsa 9,690.30 10,791.87 0 37,000

Parent contributionsa 14,119.74 12,545.74 0 45,000

Own contributionsa 599.33 1,077.65 0 5,500

Cost of attendance 25,517.60 5,406.60 14,407 30,984


Female .59 – 0 1

Male (reference) .41 – 0 1


Asian .25 – 0 1

Black .27 – 0 1

Hispanic .23 – 0 1

White (reference) .25 – 0 1

Parent SES

Income (median value in category, 82,504.04 52,316.56 1,500 200,000 log reported)a

Parents’ educationa

Less than college degree (reference) .21 – 0 1

At least college degree .79 – 0 1

College type

Liberal arts .09 – 0 1

Private research .59 – 0 1

Public research (reference) .31 – 0 1







Table1. continued

Student Loan Debt and Time Use in College page9image461809584 page9image461896592 597







Did not work for pay during year (reference)




Worked for pay during year





Natural science or pre-medicine




Other or undeclared (reference)




Source: National Longitudinal Study of Freshmen
aMeans and standard deviations (where applicable) have been averaged across imputed data sets.

period, yields more accurate data than questions about “normal” or “typical” time use (Juster, Ono, and Stafford 2003).4 However, there may be biases inherent in the retrospective data for which we cannot fully account.

We examine student time use in two ways. First, we assess whether debt is related to time spent in individual time categories. Second, we focus on our results from a latent class analysis, which suggest that these categories cluster into three distinct classes of student time use. Assessing these classes is important for under- standing how college activities hang together, and may be more revealing than time spent in individual categories.

Student loan debt

The key independent variable is self-reported student loan debt accrued during the first year of college.5 To account for differences within and across levels of indebtedness, we use a spline function, which allows us to include two separate functional forms in the analyses (Marsh and Cormier 2002). The first is a dichot- omous measure of whether students accrued any debt during the year, which distinguishes between indebted and debt-free students. The second is a continu- ous measure of debt, recoded to thousands of dollars and top-coded at the 99th percentile ($20,000), which represents an additional $1,000 in debt among indebted students.

Other funding sources and cost of attendance

As discussed above, the source(s) of students’ funding may shape their academic performance, goals, and overall approach to college (Hamilton 2013). Control- ling for other aspects of the financial aid package also helps isolate debt’s asso- ciation with time use, as opposed to patterns that could be attributable to financial aid generally. We include measures of grants (institutional and external), parent contributions, and own contributions (work-study and other earnings), recoded to thousands of dollars and top-coded at the 99th percentile. We also control for institutional-level cost of attendance (recoded to thousands of dollars) to ensure that these relationships are not due to differences in students’ total financial obligations.6


598 page10image459335424 page10image459920944 Social Forces 94(2)
Basic sociodemographics

Men and women may approach college differently, possibly leading to gender differences in the association between debt and time use. Women may spend more time studying because they are often more academically oriented than men (Buchmann and DiPrete 2006; DiPrete and Buchmann 2013), whereas men spend more time exercising, watching television, and participating in leisure activities than women do (Sax and Harper 2007). We therefore include a dichotomous con- trol for gender (1 = female). Race and ethnicity have also been shown to shape certain categories of time use, including academics (highest for Asians), recreation (highest for Blacks), and working for pay (also highest for Blacks) (Charles et al. 2009). Based on these time use differences, race could influence students’ responses to indebtedness. We include a set of dichotomous variables representing Black, Asian, and Hispanic students, with white students as the omitted category.

Socioeconomic status (SES)

Students from more privileged backgrounds could perceive a stronger safety net, obscuring the way that debt is associated with their time use, so we control for parents’ income and education. Income is measured as the natural log of the median value in an income category.7 Education is included as a dichotomous measure of whether at least one parent or guardian holds a bachelor’s degree.

College type

College type may influence the nature and number of activities available to stu- dents, thus affecting the relationship between debt and time use. Public universi- ties (even highly competitive ones) may facilitate an exclusive party culture, where students spend time socializing within the Greek system and attending sporting events (Armstrong and Hamilton 2013). Conversely, liberal arts colleges may have fewer social and extracurricular options due to smaller enrollments and differences in college culture. We include a set of dichotomous variables repre- senting liberal arts colleges and private research universities, with public research universities as the omitted category.8


Working during the year can affect students in a number of ways. Some have argued that time spent working necessarily restricts time use in other areas, such as studying or socializing with friends, and that students who work more than 20 hours per week may be overwhelmed by the combined demands of coursework, employment, family, and social commitments (Bozick 2007). But modest employ- ment can also help students, keeping them on a regular schedule, introducing them to opportunities such as faculty research, and insulating them from the pressures of debt (Pascarella and Terenzini 1991). We include a dichotomous variable rep- resenting students who worked for pay at any time during the academic year.9


Some majors are more time intensive than others, as students in the natural sci- ences are often required to attend lab sessions that supplement their time in the


Student Loan Debt and Time Use in College page11image454467584 page11image454467936 599

classroom. Students majoring in pre-medicine must also take these courses to fulfill medical school admission requirements. Aside from these academic differ- ences, science and pre-med majors could be relatively unencumbered by their debt because they have strong employment and earnings prospects. We include a dichotomous control representing students who declared a major in biochemis- try, biology, chemistry, physics, pre-medicine, or “other” natural science.

Analytic Overview

We organize our analyses in three parts. First, using OLS regression, we exam- ine how student loan debt is associated with time spent in each of our time use categories. Second, we use latent class analysis (LCA) to understand how many and which patterns of time use exist within the NLSF sample. For LCA, we assume that our seven manifest variables (individual time use measures) are related to one another, and that these variables cohere into T latent classes. After fitting T = 1, ..., n models and determining which T best fits the data, students are assigned to a latent class based on their highest posterior member- ship probability. For example, if T = 3 and a student’s time use indicates that the membership probabilities are 0.6, 0.3, and 0.1, the student would be assigned to the first latent class. We provide more detailed information on LCA in a later section. Finally, we use our latent classes as the dependent variable in a multinomial logistic regression to analyze how student loan debt is associated with lifestyle.

Our analytic sample includes 3,676 students. We account for missing data on the independent and control variables using multiple imputation by chained equations. Prior to model estimation, we dropped respondents with missing time use data (n = 248, or approximately 6 percent of the sample).


Student Loan Debt and Categories of Time Use

Table 2 indicates the association between student loan debt and time spent in categories of time use over a one-week period. To reiterate, we used a spline func- tion to examine two functional forms of debt: binary (indebted versus debt-free) and continuous (the effect of an additional $1,000 of debt among indebted stu- dents). Looking at the binary term, we observe two basic differences in the ways that indebted and debt-free students spend their time. Not surprisingly, students with debt spend more time working for pay than debt-free students—approxi- mately 1.2 hours more per week (b = 1.22, p < .01). This difference could emerge because indebted students must hold work-study jobs as part of their financial aid packages. It is also possible that, although indebted students have accrued loans to pay for their tuition, they have not borrowed enough to cover their expenses and must work to meet their everyday needs. Indebted students are also less active in athletic activities than debt-free students (b = –.98, p < .05). This disparity may occur because many varsity athletes receive scholarships and other funding, so they would report high athletic participation and minimal debt. The NLSF did not


Source: National Longitudinal Study of Freshmen
Note: Unstandardized OLS coefficients. Standard errors are adjusted for clusters of students attending the same college. Omitted categories are no loans, men, white, neither parent has degree, public research, did not work for pay, major other than natural science or pre-med.
***p < 0.001 **p < 0.01 *p < 0.05 (two-tailed tests)


602 page14image365172336 page14image454476992 Social Forces 94(2)

collect information on varsity athlete status directly, so we are not certain that it is of consequence, but we venture that it has affected this result in some way.

For one additional category, students’ time use varies based on the amount of debt accrued. Specifically, among indebted students, each additional $1,000 is associated with more time spent watching television and listening to music (b = .41, p < .01). Although this association appears minor, small differences in weekly time use can add up to larger differences across a semester or academic year. Consider those who have accrued the mean amount of debt among indebted students, or approximately $5,000 (see table 1). Over the course of a 32-week academic year, they may spend approximately 66 more hours consuming media than those with the least debt.10,11

As discussed above, our use of the first-year data minimizes endogeneity con- cerns because financial aid packages are typically set before students arrive on campus. However, we also conducted supplementary fixed-effects models with the longitudinal data to rule out unobserved variable bias (see tables S1 and S2 in the online supplementary material). These models suggest that debt’s association with time use extends across the first three years of college, when the NLSF fielded the time use module. On the bivariate level, more debt is associated with more time spent working for pay and consuming media, and less time spent on athletics, partying, and sleeping. When time-varying controls are introduced, more debt is still associated with more time spent working for pay and less time on athletics, which corroborates our findings in table 2.

College Lifestyles as Measured by Time Use

During the first year of college, student debt appears to be unrelated to time spent on academics, student groups, partying, and sleeping. These results are surprising in light of past research on financial status and the college experience. Specifically, we expected some students to embrace the responsibility of debt and to spend more time on academics, yet this pattern does not appear to emerge. Student loan debt is also not significantly related to time spent on student group activities or partying, despite our prediction that debt would create social liabilities for some students. We suspect that the OLS models may be obscuring the association between debt and time use by overlooking the possibility that indebted students follow more than one time use pattern. Accordingly, we turn to our latent class analysis (LCA) results. LCA allows us to identify combinations of activities (or “lifestyles”) in which college students participate, and to examine how debt is related to these lifestyles. If indebted students exhibit divergent time use pat- terns (for example, one that is high on academic time use and another that is low), these lifestyles could help explain why there is no basic link between debt and time spent on several college activities.

To produce the dichotomous variables required for an LCA (Collins and Lanza 2010), we transformed our time use measures into binaries representing above- mean time spent in that category for the week, where the reference category is at or below the mean.12 The exception is the working for pay category. Because many students in the sample worked zero hours (59 percent), the mean is artificially low, and most students who worked for pay did so above the mean


Figure 1. Summary of three-class LCA, N = 3,6760.7

Student Loan Debt and Time Use in College page15image457950288 page15image457950640 603

amount of 5.25 hours. We have thus streamlined this measure so that the binary variable represents students who worked a non-zero amount. Taken together, students who are coded “1” for an activity are said to be concentrating their time in that category. They have made a considerable investment in that activity rela- tive to other students in the sample, and they may sacrifice time spent on other activities to sustain their participation.

We started by generating a model with one latent class, and continued to add classes until the model fit ceased to improve any more, as indicated by the Bayes- ian Information Criterion (BIC). BIC statistics for each of the latent class models are shown in appendix A. We analyzed the three-class model due to its model fit, and because other scholars have identified college lifestyles similar to these in their work, as discussed below. Figure 1 presents a graphical summary of the model, with the x-axis representing each of the activities and the y-axis represent- ing a student’s conditional probability of spending above-mean time on an activ- ity (or participating in the activity at all, in the case of work), given that he or she is in a particular latent class. Below, we label and describe our latent classes based on their relative conditional probabilities of spending above-mean time in certain activities.

Serious students

Students in this latent class may prioritize work- and school-related activities ahead of leisure pursuits. They have the highest probabilities of working for pay and spending above-mean time on academics. They also have a moderate Sleep




Source: National Longitudinal Survey of Freshmen
Note: Probabilities are shown as the conditional probability of participating in the activity above the mean amount, given membership in the latent class; the “work” probability is the conditional probability of working any amount, given membership in the latent class. Percentages do not add to 100 due to rounding.

Pr (above-mean time) | In latent class


604 page16image455938192 page16image455955776 Social Forces 94(2)

probability of spending above-mean time on student group activities, compared to the other two latent classes. In contrast, the Serious Students have the lowest probability of spending above-mean time on the rest of the activities, including media, athletics, partying, and sleep. The Serious Students are the modal latent class, making up approximately 41 percent of the sample.


These students appear to be relatively uninvolved in campus life. They have by far the lowest probabilities of spending above-mean time on academics and stu- dent group activities, and they hover between the other two classes in terms of working for pay, athletics, and partying. Interestingly, the Inactives have a rela- tively high probability of spending above-mean time on media, and they have the highest probability of spending above-mean time on sleep. They comprise approximately 26 percent of the sample.13

Socially Engaged

These students are highly involved on campus, both socially and otherwise. They have the highest probabilities of spending above-mean time on athletics, party- ing, student group activities, and (to a lesser extent) media. They also have the lowest probability of working for pay, and have modest probabilities of spend- ing above-mean time on academics and sleep. It may be that students in the Socially Engaged class spend so much time on extracurricular activities that they must satisfice academically, spending enough time to get by but not enough to excel in the classroom (Hamilton 2013; Simon 1955, 1957).14 Satisficing is not an ideal strategy for learning (Arum and Roksa 2011), but the Socially Engaged students do not neglect their studies, either.15 Approximately 34 percent of the sample is Socially Engaged.

As noted above, in past research on college student typologies and time use, schol- ars have identified “types” of undergraduates that are similar to the Serious Stu- dents (Rau and Durand 2000), Inactives (Brint and Cantwell 2012), and Socially Engaged (Arum and Roksa 2011; Moffatt 1989). In fact, some have described types that approximate all three (Armstrong and Hamilton 2013; Brint and Cantwell 2010; Clark and Trow 1966). Most research in this area is descriptive, and does not attempt to map students’ characteristics onto typologies. Armstrong and Hamilton’s (2013) study is an exception; they show that SES shapes women’s capacity to break into more socially oriented lifestyles, as well as their ability to thrive in whatever pathway they take through higher education. Our emphasis on student loan debt complements this work, while refining the idea that students’ characteristics may bear some relation with their college experiences.16

Student Loan Debt as a Predictor of Lifestyle

How is indebtedness associated with student lifestyles? To answer this question, we conducted a multinomial logistic regression using our latent classes as the out- come variable (see table 3).17 Column 1 indicates that indebted students have higher odds of being Serious Students versus Socially Engaged (eb = 1.35, p < .01),



Student Loan Debt and Time Use in College 605Table 3. Multinomial Logistic Regression of Lifestyle Types, N = 3,676


Column 1: Serious students– Socially engaged

Column 2: Inactive– Socially engaged

Column 3: Serious students– Inactive

Student loan debt 1.35** 1.30* 1.04


(.11) (.11) (.10)

Student loan debt 1.01 1.02 1.00


(.02) (.02) (.01)

Grants 1.00 .99 1.00

(.01) (.01) (.01)

Parent contributions .98** 1.00 .99*

(.01) (.01) (.01)

Own contributions 1.07* 1.03 1.05

(.03) (.05) (.04)

Cost of attendance 1.05** 1.00 1.05*

(.01) (.01) (.02)

Female 2.47*** 1.63*** 1.52***

(.11) (.11) (.10)

Asian 1.68*** 1.23 1.36*

(.12) (.15) (.13)

Black 1.21 .98 1.24

(.12) (.20) (.15)

Hispanic 1.39** 1.09 1.27

(.12) (.17) (.14)

Parents’ income .90 1.00 .90

(.06) (.07) (.07)

Parent has degree .92 .77* 1.20

(.10) (.12) (.11)

Liberal arts .98 1.16 .85

(.23) (.26) (.28)

Private research .69** 1.04 .67

(.11) (.17) (.21)

Natural science 1.37** .92 1.49**

(.11) (.12) (.11)




Source: National Longitudinal Survey of Freshmen
Note: Standard errors are adjusted for clusters of students attending the same college. Omitted categories are no loans, men, white, neither parent has degree, public research, major other than natural science or pre-med.
***p < 0.001 **p < 0.01 *p < 0.05 (two-tailed tests)


606 page18image455957424 page18image455895040 Social Forces 94(2)

and the continuous term is positive but not significant. In column 2, the odds ratio has a lower significance level, which helps us sort out the relative odds of indebted students being in each of the latent classes. Indebted students have higher odds of being Inactive versus Socially Engaged (eb = 1.30, p < .05), and the continuous term is again not significant. Based on these first two columns, we conclude that students with debt clearly have low odds of exhibiting the Socially Engaged life- style. However, column 3 shows that the remaining two classes are less differenti- ated. Indebted students have slightly higher odds of being Serious Students versus Inactive, but the odds ratio is not significant (eb = 1.04, p = .73). Thus, indebted students have elevated odds of being both Serious Students and Inactive. More- over, the fact that a student has accrued any debt (versus no debt) may be more salient for predicting lifestyle than the specific amount of debt accrued.

This apparent duality—the fact that indebtedness is related to both the Serious Student and Inactive lifestyles—also helps clarify some seemingly incongruous findings between our OLS and multinomial results. Specifically, we observed in table 2 that student loan debt is not significantly related to time spent on academ- ics, student groups, or partying, but here we find that there are indeed group differences in these activities. This anomaly arises because indebted students are effectively separated into two very different groups. Most are Serious Students (high on academics, relatively high on student group activities, low on partying), but some are Inactive (low on academics, low on student group activities, rela- tively high on partying). When both of these classes are combined under the cat- egory of “indebted students” in the OLS models, the relationship between debt and time use may effectively wash out.18 Accordingly, the LCA models are impor- tant for capturing two types of indebtedness. Differences in individual time use categories are instructive at a basic level but, ultimately, they do not reveal the extent to which indebted students vary in their time use.

Discussion and Conclusion

Student loans are becoming an increasingly relevant factor in higher education finance. Although scholars have addressed debt’s relationship with persistence and completion, we have little empirical knowledge of how student loans are associated with everyday college experiences. Some scholars contend that debt fosters a sense of responsibility in students, and that they may actively reject socializing in order to devote themselves to academic and work activities. Others envision a very different scenario, in which indebted students are faced with lia- bilities that oblige them to passively withdraw from campus life.

In this paper, we used time use as a proxy for college experiences and examined how student loan debt is associated with time use in college. We find that many indebted students belong to a class of Serious Students, who spend much of their time working for pay, going to class, and studying. These students may have embraced the responsibility of student loan debt. Yet, we also find that indebted students are disproportionately Inactive. These students may have succumbed to the liability of debt, ending up largely detached from the academic and extracur- ricular aspects of college. For them, debt may be disabling in the sense that it encourages withdrawal from campus life. Based on these divergent patterns of


Student Loan Debt and Time Use in College page19image454160720 page19image454161072 607

embracing responsibility and succumbing to liability, we posit that holding debt may polarize the college experience, producing students who are differentially involved on campus. In other words, there may not be just one indebted student experience, but stratification within this subpopulation.

There could be several reasons that students appear to respond to debt in dis- parate ways. First, loan type could be driving these findings. Students with per- sonal loans or high-interest private loans could feel more overwhelmed by their debt, pushing them toward the Inactive category. Conversely, if a student’s parents are assuming their debt, or if they have access to more lenient government loans, they could be motivated to exhibit a Serious Student lifestyle. Our data do not include these loan characteristics, but they could very well be motivating students’ outlooks toward their debt. There may also be demographic differences in indebted student lifestyles. Although table 3 gives us some insight into how stu- dent characteristics are associated with lifestyles, a multinomial logistic regression with just the indebted students is shown in appendix B. Here, we are most con- cerned with the differences between the Serious Students and Inactives (column 1), but we also include comparisons to the Socially Engaged class (columns 2 and 3). Interestingly, women and natural science majors are more often Serious Students than Inactives, but funding sources and other controls do not differentiate between the two groups.19 We posit that students’ characteristics may provide some insight into how indebtedness links with these divergent lifestyles. It is outside the scope of this paper to pinpoint exactly what separates indebted students into the Serious Students and the Inactives, but further research could explain these cleavages.

We must also assess alternative explanations aside from a direct casual effect of student loan debt. It may be that students who take on debt have multiple charac- teristics that also predict their time use; in other words, debt may act as a proxy for students’ selection into particular time use patterns. For example, past research shows that middle-income students are often indebted (Choy and Berker 2003;Houle 2014), so it may be that middle-income students have unique college expe- riences—they may struggle to relate to wealthier students, but are also not included in programming meant for students from low-income families. Indebted students could also come from families with relatively little savings (and, accordingly, less of a financial safety net), which could shape their approach to college life and time use generally. Additionally, if students with debt were involved in financial plan- ning for college, they may simply be more cognizant of the costs and potential benefits of their education. To reiterate, holding debt itself could be influential for some students, but these selection effects are also plausible. Future research should further examine these factors to clarify the relationship between debt and time use.

Due to the NLSF’s unique sampling design, our results capture the experiences of a particular subset of undergraduates. The racial composition of the sample (approximately equal numbers of white, Black, Asian, and Hispanic students) indi- cates that our findings are most applicable to minority students at selective colleges. Supplementary analyses suggest that our overall patterns hold when the sample is stratified by race—that is, indebted students within each race group are more likely to be either Serious Students or Inactive, compared to Socially Engaged—but fur- ther research is needed to extend our findings to more typical samples of college students.20 The nature of the NLSF institutions could also be influencing our results


608 page20image407327312 page20image444877280 Social Forces 94(2)

in multiple ways. First, most of these students were quite accomplished in high school and college, both academically and in their chosen activities (Massey et al. 2006; Charles et al. 2009). At less prestigious schools, we would expect smaller proportions of Serious Students, and larger proportions in the Inactive and Socially Engaged classes. The overall construction of the lifestyle patterns could also change. Serious Students elsewhere may be as involved in extracurricular activities; the Socially Engaged may only make time for parties and sleep; and the Inactives may be even less involved. Finally, the socioeconomic backdrop at selective colleges could be altering our findings. The indebted experience may be less burdensome at elite schools because students anticipate a bigger payoff in the long term. Addi- tional data collection is needed to understand how a broader swath of indebted students do college, and to what extent the NLSF students are atypical.

Our selective college backdrop also suggests implications for the life course. Specifically, there may be negative consequences for indebted students who are not socially active at elite schools. Although indebted students are polarized between the Serious Student and Inactive lifestyles, it is clear that they are least often Socially Engaged. Yet, the activities associated with a Socially Engaged life- style—playing sports, attending sporting events, partying, and extracurricular activities—may comprise the dominant script of college life (Grigsby 2009;Hamilton 2010). Even employers may believe that a selective college degree car- ries expectations about a person’s cultural experiences. Rivera (2011) shows that many elite students aim for jobs in law, investment banking, and consulting, and that these firms value involvement even more than grades. If indebted students’ resumes are less competitive in this regard, they may be excluded from opportu- nities to vault themselves into the upper-middle or upper classes.

Even if they do not aspire to professional service work, we venture that indebted students may not enjoy the benefits that come with an active social life. To be sure, there are obvious drawbacks to partying in college, and we do not argue that indebted students are directly disadvantaged by not exhibiting this lifestyle. But it is possible that students with debt are leaving college with fewer network ties and less familiarity with elite-status culture—resources that can be especially valuable when they are stemming from prestigious schools (Karabel 2005). Indebted stu- dents may thus be making a distinct trade-off when it comes to the forms of capital (Becker 1964; Bourdieu 1986). By accruing student loans, they have been able to develop human capital beyond what their financial circumstances would dictate. But in the process of earning their degrees, indebted students may be losing out on opportunities to develop social and cultural capital. A similar process has been documented among low-income college students (Lee and Kramer 2013; Stuber 2006, 2009), but debt may be uniquely related to inequalities beyond those of SES.

In light of our findings, we contend that a popular account of indebted stu- dents may be unfounded. As noted above, various figures in the media, politics, and elsewhere have asserted that indebted students are irresponsible youths who are using their loans to finance their indiscretions. To the contrary, we find that most students with debt are less involved in partying and other non-academic activities than those without debt. As discussed above, we recognize that our sample shapes the extent to which indebted students engage in certain college activities, but there appears to be little basis for this generalization overall.


Student Loan Debt and Time Use in College page21image455398432 page21image455398784 609

By most accounts, the college experience is changing for many Americans. Students who would never have considered furthering their education are now enrolling in degree programs. Many are choosing institutions outside the tradi- tional university model. But just as college demographics and structures have changed, students are increasingly using loans to cover some or all of their college costs, implying that there are differences not just across institutions, but also within institutions. This paper represents one of the first inquiries into how finan- cial status in general, and student loan debt in particular, is associated with stu- dents’ experiences in college. With more data and research in this area, we may more fully understand the responsibilities and liabilities of college life.

Supplementary Material

Supplementary material is available at Social Forces online, http://sf.oxfordjourn-


  1. Time use at the end of the term could be more academic than at other points in the year because students are preparing for final exams. However, we focus on how stu- dents compare to each other in their time use, and there is little reason to suspect that relative measures would be altered during this period.

  2. We excluded a measure of time spent “socializing with others” because our latent class analyses with this variable produced solutions with local maxima, suggesting that the model is poorly defined for the data (Muthén and Muthén 2010). We suspect this occurred because “socializing” is a broad concept that elicited a higher mean and range than other activities in the survey.

  3. In supplementary analyses, coefficients for debt were similar for the week alone, the weekend alone, and the week and weekend combined.

  4. Students also completed a time diary for the most recent Tuesday that school was in session, and these data were comparable to their weekday estimates. We suspect that this exercise improved the internal validity of students’ time use.

  5. Our data do not enable us to distinguish between students who accrued their own loans versus those whose parents accrued loans for them, but this distinction could shape time use. We address this issue in the discussion.

  6. The data do not allow us to discern between in-state and out-of-state students, so we use out-of-state cost of attendance for public research institutions.

  7. The NLSF asked respondents to approximate their parents’ income in the first and second years. Although the first-year measure is detailed for lower incomes, the uppermost category includes all incomes $75,000+—a modest amount for elite stu- dents. The second-year measure is more complete for the higher incomes. If students reported $75,000+ in both years, we imputed the second-year measure. If students reported $75,000+ in the first year and were missing in the second year, we imputed the modal income for students reporting $75,000+ in the second year—$112,500. For students reporting $75,000+ in the first year and $200,000+ in the second year, we imputed $200,000.

  8. Students at HBCUs may exhibit unique time use, but based on the way the data were reported, we cannot distinguish the HBCU from other public universities.

  9. Supplementary analyses with various specifications of the employment measure (high/ low, continuous, quadratic, etc.) were consistent with those presented.


610 page22image453790560 page22image453790848 Social Forces 94(2)

  1. Wetestedforinteractionsbetweenstudentloans(binaryandcontinuous)andgender, race/ethnicity, and family income. Of these 70 coefficients, only three were significant, which is well within the range of randomness.

  2. In supplementary analyses, higher SAT scores were associated with more time spent on academics and less time spent on athletics and partying. The effect of debt was unchanged with this control.

  3. We conducted an LCA instead of a latent profile analysis (a similar method using continuous variables) because the large ranges in our time use data produced unac- ceptably small cell sizes.

  4. The Inactives were more likely to transfer in subsequent years than students in other latent classes. Latent class assignment does not significantly predict dropout.

  5. Accordingly, these students’ first-year grades ranked in the middle of the other two classes. The Serious Students earned the best grades, marginally higher than the Socially Engaged and significantly higher than the Inactives.

  6. In supplementary analyses, we dichotomized the time use variables at the median instead of the mean (with work still dichotomized as none/any) and found that the LCA model and relationships with debt were robust. The means provided slightly better separation of the latent classes.

  7. In supplementary analyses, we fit separate LCAs for the second and third years. In all three years, the modal latent class approximates the Serious Students. In the second year, the Socially Engaged are replaced by an all-around engaged class, with high academics, work, and other activities. In the third year, many students straddle the 50 percent conditional probability mark for most activities, and a small number spend above-mean time on partying and sleep only.

  8. Weremovedtheemploymentcontrolfromthismodelbecausetimespentworkingfor pay is incorporated into the outcome.

  9. IndebtedSeriousStudentsspentapproximately20morehoursperweekonacademics (54 hours versus 34 hours, p < .001) and 8 more hours per week on student groups (14 hours versus 6 hours, p < .001) than indebted Inactives. Conversely, indebted Inactives spent approximately 3 more hours partying than indebted Serious Students (7.3 hours versus 4.7 hours, p < .001). These groups obscure each other when com- bined in the OLS models.

  10. In supplementary analyses, interaction terms for (1) debt and race/ethnicity and (2) debt and family income did not significantly differentiate between the latent classes.

  11. In an exception, Asian indebted students are not more likely to be Inactive than Socially Engaged.

Appendix A. BIC Statistics for Optimal Number of Classes in the LCA, N = 3,676
















Source: National Longitudinal Survey of Freshmen



Student Loan Debt and Time Use in College 611Appendix B. Multinomial Logistic Regression of Lifestyle Types for Indebted Respondents,

N = 1,662


Column 1: Serious students– Inactive

Column 2: Serious students– Socially engaged

Column 3: Inactive– Socially engaged

Grants 1.01 1.00 .99

(.01) (.01) (.01)

Parent contributions 1.00 .99 .99

(.01) (.01) (.01)

Own contributions 1.10 1.10* 1.00

(.05) (.05) (.07)

Cost of attendance 1.05 1.05** 1.01

(.03) (.02) (.02)

Female 1.42* 2.62*** 1.84***

(.15) (.14) (.13)

Asian 1.35 1.23 .91

(.21) (.22) (.22)

Black 1.17 1.10 .94

(.19) (.18) (.24)

Hispanic 1.09 1.31 1.20

(.18) (.19) (.27)

Parents’ income .89 .86 .96

(.08) (.08) (.08)

Parent has degree 1.02 1.00 .98

(.15) (.14) (.16)

Liberal arts .76 .88 1.16

(.33) (.24) (.37)

Private research .64 .73 1.13

(.26) (.20) (.24)

Natural science 1.71** 1.56* .91

(.18) (.17) (.18)




Source: National Longitudinal Survey of Freshmen
Note: Standard errors are adjusted for clusters of students attending the same college. Omitted categories are no loans, men, white, neither parent has degree, public research, major other than natural science or pre-med.
***p < 0.001 **p < 0.01 *p < 0.05 (two-tailed tests)


612 page24image451791712 page24image451809088 Social Forces 94(2)
About the Authors

Natasha Yurk Quadlin is a PhD candidate in the Department of Sociology at Indiana University. She is broadly interested in social stratification, particularly in the contexts of families and schools. In her current work, she is using data from a national survey experiment to examine norms for intra-family resource distri- bution based on siblings’ relative achievement and gender.

Daniel Rudel is a PhD candidate in the Department of Sociology at Indiana University. His primary areas of interest are in the sociology of education and stratification, with a special focus on higher education. His recent research has examined the consequences of student loan debt, inequalities in the college expe- rience, and gender differences in educational expectations.


Armstrong, Elizabeth A., and Laura T. Hamilton. 2013. Paying for the Party: How College Maintains Inequal- ity. Cambridge, MA: Harvard University Press.

Arum, Richard, and Josipa Roksa. 2011. Academically Adrift: Limited Learning on College Campuses. Chicago: University of Chicago Press.

Babcock, Philip, and Mindy Marks. 2011. “The Falling Time Cost of College: Evidence from Half a Century of Time Use Data.” Review of Economics and Statistics 93:468–78.

Baird, Charlie L., Stephanie W. Burge, and John R. Reynolds. 2008. “Absurdly Ambitious? Teenagers’ Expec- tations for the Future and the Realities of Social Structure.” Sociology Compass 2:944–62.

Becker, Gary. 1964. Human Capital. New York: National Bureau of Economic Research.
Best, Joel, and Eric Best. 2014. The Student Loan Mess: How Good Intentions Created a Trillion-Dollar

Problem. Berkeley: University of California Press.
Bodvarsson, Orn B., and Rosemary L. Walker. 2004. “Do Parental Cash Transfers Weaken Performance in

College?” Economics of Education Review 23:483–95.
Bourdieu, Pierre. 1986. “The Forms of Capital.” In Handbook of Theory and Research for the Sociology of

Education, edited by Richardson, John G., 241–58. New York: Greenwood.
Bowen, William G., and Derek Bok. 1998. The Shape of the River: Long-Term Consequences of Considering

Race in College and University Admissions. Princeton, NJ: Princeton University Press.
Bozick, Robert. 2007. “Making It Through the First Year of College: The Role of Students’ Economic

Resources, Employment, and Living Arrangements.” Sociology of Education 80:261–85.
Brint, Steven, and Allison M. Cantwell. 2010. “Undergraduate Time Use and Academic Outcomes: Results from the University of California Undergraduate Experience Survey 2006.” Teachers College Record

Brint, Steven, and Allison M. Cantwell. 2012. Portrait of the Disengaged. Berkeley, CA: Center for Studies

in Higher Education.
Buchmann, Claudia, and Thomas A. DiPrete. 2006. “The Growing Female Advantage in College Comple-

tion: The Role of Family Background and Academic Achievement.” American Sociological Review

Chambliss, Daniel F., and Christopher G. Takacs. 2014. How College Works. Cambridge, MA: Harvard

University Press.
Charles, Camille Z., Mary J. Fischer, Margarita A. Mooney, and Douglas S. Massey. 2009. Taming the River:

Negotiating the Academic, Financial, and Social Currents in Selective Colleges and Universities.

Princeton, NJ: Princeton University Press.

Choy, Susan P., and Ali M. Berker. 2003. “How Families of Low -and Middle-Income Undergraduates Pay for College: Full-Time Dependent Students in 1999–2000.” National Center for Education Statistics.


Student Loan Debt and Time Use in College page25image453699776 page25image453700128 613

Clark, Burton R., and Martin Trow. 1966. “The Organizational Context.” In College Peer Groups: Problems and Prospects for Research, edited by Newcomb, Theodore, and Everett K. Wilson, 17–70. Chicago: Aldine.

Collins, Linda M., and Stephanie T. Lanza. 2010. Latent Class and Latent Transition Analysis: With Applica- tions in the Social, Behavioral, and Health Sciences. Hoboken, NJ: Wiley.

Conley, Dalton. 1999. Being Black, Living in the Red: Race, Wealth, and Social Policy in America. Berkeley: University of California Press.

Cowley, W. H., and Willard Waller. 1935. “A Study of Student Life.” Journal of Higher Education6:132–42.

DiPrete, Thomas A., and Claudia Buchmann. 2013. The Rise of Women: The Growing Gender Gap in Educa- tion and What It Means for American Schools. New York: Russell Sage Foundation.

Drentea, Patricia. 2000. “Age, Debt, and Anxiety.” Journal of Health and Social Behavior 41:437–50. Drentea, Patricia, and John R. Reynolds. 2012. “Neither a Borrower Nor a Lender Be: The Relative Importance of Debt and SES for Mental Health Among Older Adults.” Journal of Health and Aging

Dwyer, Rachel E., Randy Hodson, and Laura McCloud. 2013. “Gender, Debt, and Dropping Out of College.”

Gender & Society 27:30–55.
Dwyer, Rachel E., Laura McCloud, and Randy Hodson. 2011. “Youth Debt, Mastery, and Self-Esteem: Class-

Stratified Effects of Indebtedness on Self-Concept.” Social Science Research 40:727–41.
______. 2012. “Debt and Graduation from American Universities.” Social Forces 90:1133–55. Espenshade, Thomas J., and Alexandria Walton Radford. 2009. No Longer Separate, Not Yet Equal: Race

and Class in Elite College Admission and Campus Life. Princeton, NJ: Princeton University Press. Grigsby, Mary. 2009. College Life Through the Eyes of Students. Albany, NY: SUNY Press.
Hamilton, Laura T. 2010. “Strategies for Success: Parental Funding, College Achievement, and the Transition

to Adulthood.” PhD dissertation, Department of Sociology, Indiana University, Bloomington.
______. 2013. “More Is More or More Is Less? Parental Financial Investments during College.” American

Sociological Review 78:70–95.
Harris, Douglas N., and Sara Goldrick-Rab. 2012. “Improving the Productivity of Education Experiments:

Lessons from a Randomized Study of Need-Based Financial Aid.” Education Finance and Policy

Hodson, Randy, Rachel E. Dwyer, and Lisa A. Neilson. 2014. “Credit Card Blues: The Middle Class and the

Hidden Costs of Easy Credit.” Sociological Quarterly 55:315–40.
Houle, Jason. 2014. “Disparities in Debt: Parents’ Socioeconomic Status and Young Adult Student Loan

Debt.” Sociology of Education 87:53–69.
Juster, F. Thomas, Hiromi Ono, and Frank P. Stafford. 2003. “An Assessment of Alternative Measures of Time

Use.” Sociological Methodology 33:19–54.
Karabel, Jerome. 2005. The Chosen: The Hidden History of Admission and Exclusion at Harvard, Yale, and

Princeton. Boston: Houghton Mifflin.
Killewald, Alexandra. 2013. “Return to Being Black, Living in the Red: A Race Gap in Wealth That Goes

Beyond Social Origins.” Demography 50:1177–95.
Kim, Dongbin. 2007. “The Effect of Loans on Students’ Degree Attainment: Differences by Student and

Institutional Characteristics.” Harvard Educational Review 77:64–100.
Lee, Donghoon. 2013. “Household Debt and Credit: Student Debt.” Federal Reserve Bank of New York.
Lee, Elizabeth M., and Rory Kramer. 2013. “Out with the Old, In with the New? Habitus and Social Mobility

at Selective Colleges.” Sociology of Education 86:18–35.
Lin, Ken-Hou. 2012. “Revisiting the Gap Between Stylized and Diary Estimates of Market Work Time.” Social

Science Research 41:380–91.
Marsh, Lawrence C., and David R. Cormier. 2002. Spline Regression Models. Thousand Oaks, CA: Sage.


614 page26image454073312 page26image368615664 Social Forces 94(2)

Massey, Douglas S., Camille Z. Charles, Garvey Lundy, and Mary J. Fischer. 2006. The Source of the River: The Social Origins of Freshmen at America‘s Selective Colleges and Universities. Princeton, NJ: Princeton University Press.

Milkie, Melissa A., Sara B. Raley, and Suzanne M. Bianchi. 2009. “Taking on the Second Shift: Time Alloca- tions and Time Pressures of US Parents with Preschoolers.” Social Forces 88:487–517.

Moffatt, Michael. 1989. Coming of Age in New Jersey: College and American Culture. New Brunswick, NJ: Rutgers University Press.

Muthén, Linda K., and Bengt O. Muthén. 2010. Mplus User‘s Guide, 6th ed. Los Angeles: Muthén & Muthén. Nathan, Rebekah. 2005. My Freshman Year: What a Professor Learned by Becoming a Student. Ithaca, NY:

Cornell University Press.
Pascarella, Ernest T., and Patrick T. Terenzini. 1991. How College Affects Students. San Francisco: Jossey-

Rau, William, and Ann Durand. 2000. “The Academic Ethic and College Grades: Does Hard Work Help

Students to ‘Make the Grade’?” Sociology of Education 73:19–38.
Rivera, Lauren A. 2011. “Ivies, Extracurriculars, and Exclusion: Elite Employers’ Use of Educational Creden-

tials.” Research in Social Stratification and Mobility 29:71–90.
Sax, Linda J., and Casandra E. Harper. 2007. “Origins of the Gender Gap: Pre-College and College Influences

on Differences Between Men and Women.” Research in Higher Education 48:669–94.
Sayer, Liana J., Suzanne M. Bianchi, and John P. Robinson. 2004. “Are Parents Investing Less in Children?

Trends in Mothers’ and Fathers’ Time with Children.” American Journal of Sociology 110:1–43. Schneider, Daniel. 2012. “Gender Deviance and Household Work: The Role of Occupation.” American Jour-

nal of Sociology 117:1029–72.
Sheehy, Kelsey. 2013. “Undergrads Blow It with Student Loan Refunds.” US News & World Report. http://

Simon, Herbert A. 1955. “A Behavioral Model of Rational Choice.” Quarterly Journal of Economics 59:

______. 1957. Models of Man, Social and Rational: Mathematical Essays on Rational Human Behavior.

New York: Wiley.
Stevens, Mitchell L., Elizabeth A. Armstrong, and Richard Arum. 2008. “Sieve, Incubator, Temple, Hub:

Empirical and Theoretical Advances in the Sociology of Higher Education.” Annual Review of Sociology

Stinebrickner, Ralph, and Todd R. Stinebrickner. 2004. “Time-Use and College Outcomes.” Journal of Econo-

metrics 121:243–69.
Stuber, Jenny. 2006. “Talk of Class: The Discursive Repertoires of White Working- and Upper Middle-Class

College Students.” Journal of Contemporary Ethnography 35:285–318.
______. 2009. “Class, Culture, and Participation in the Collegiate Extra-Curriculum.” Sociological Forum

Sweet, Elizabeth, Arijit Nandi, Emma K. Adam, and Thomas W. McDade. 2013. “The High Price of Debt:

Household Financial Debt and Its Impact on Mental and Physical Health.” Social Science and Medicine

Terenzini, Patrick T., and Ernest T. Pascarella. 1977. “An Assessment of the Construct Validity of the Clark-

Trow Typology of College Student Subcultures.” American Educational Research Journal 14:225–48.

DMU Timestamp: February 03, 2020 23:30