Abstract
Despite improved access in expanded postsecondary systems, the great majority of bachelor’s degree graduates are taking considerably longer than the allotted four years to complete their four-year degrees. Taking longer to finish one’s BA has become so pervasive in the United States that it has become the norm for official statistics released by the Department of Education to report graduation rates across a six-year window. While higher education scholars have increasingly explored how social class impacts college dropout, attrition, and completion, they have yet to examine the role social class plays in completing a four-year bachelor’s degree on time. In this paper, we draw on the most recent cohort of the Baccalaureate and Beyond Longitudinal Survey (2008–2009) to examine who completes their bachelor’s degrees on time. Our results indicate that despite controlling for academic performance, educational behaviors, program characteristics, and institutional characteristics, graduates from lower socioeconomic backgrounds do experience difficulties completing their degrees on time. Moreover, our results also reveal that the nature of these relationships vary for traditional and nontraditional students. Our findings highlight another important, albeit less obvious, way where inequality is maintained in expanded postsecondary systems.
Keywords
In the fall of 2013, more than 10.5 million Americans were pursuing a bachelor’s degree at a four-year degree-granting postsecondary institution in the United States (National Center for Education Statistics [NCES] 2015). Of these 10.5 million, approximately 20 percent dropped out after the first year. While many students do carry on and successfully obtain their bachelor’s degree, a sizeable proportion of these graduates are taking considerably longer than the allotted four years to complete their “four-year” degrees. Taking longer to finish one’s BA has become so pervasive that it has become the norm for official statistics released by the NCES to report bachelor’s degree completion data over a six-year window (e.g., NCES 2015). In fact, in 2013, only about 59 percent of students who began seeking their bachelor’s degree at a four-year institution had completed that degree within six years, and only 39.4 percent had done so in four years (Digest of Education Statistics 2015).
Prior scholarship on time to degree has largely approached the issue from an efficiency perspective and been motivated by rationales that seek timely completion rates to avoid wasting student and/or institution dollars (e.g., Carlozo 2012; Knight 2004; Korn 2015; Pitter, LeMon, and Landham 1996). From an institution perspective, decreasing undergraduates’ time to bachelor’s degree attainment is important to ensure the availability of sufficient resources per student (Hakkinen and Uusitalo 2003; Jenkins and Rodriguez 2013). For students, every additional year spent in an undergraduate degree is one less year in the labor market, where graduates could be contributing to their lifetime earnings, savings, and pensionable contributions, not to mention the added burden of tuition and student fees, often resulting in additional student loans (Volkwein and Lorang 1996).
Our approach brings social class background to the forefront and demonstrates how the unequal distribution of parental resources contributes to varied college completion times. To our knowledge, no existing studies have sought to examine the ways in which social class may continue to exert its influence on the timely completion of a college degree. This gap in knowledge is surprising given that the pervasive influence of social background has long been used to explain persistence and school continuation decisions (e.g., Mare 1980, 1981; Stolzenberg 1994). At the same time, research on college retention and attrition has also understood that social background characteristics influence students’ interactions within their institutional environments, their level of integration into the college system and commitment, and ultimately their decision to drop out (e.g., Pascarella and Terenzini 1979; Terenzini and Pascarella 1980; Tinto 1975, 1993, 2012). Moreover, theories of cumulative dis/advantage have explained that not only is the level of education achieved important for determining later life course outcomes but so too is the timing of educational achievements as it further stratifies individuals in the workforce. Life course timing differences in educational attainment, early timing (or “Matthew effect”), or obtaining degrees in a timely fashion, have real implications for wages and one’s likelihood of returning to school for further education (Elman and O’Rand 2004). Yet the nature of the relationship between social class and the timely completion of a degree remains relatively underexplored in the sociology of education.
Additionally, we argue that extended times to completing a degree may also lead to further inequality for traditionally disadvantaged groups. While the ongoing expansion of higher education appears on the surface to be beneficial for a number of traditionally disadvantaged groups, it also tends to generate additional dimensions of inequality within credential tiers (Lucas 2001, 2009, 2017). For instance, previous sociological studies that investigate undergraduate work commitments found that work concurrent with the pursuit of college leads to further social class differentiation in expanded postsecondary systems (Roksa and Velez 2010; Weiss and Roksa 2016). Others have shown how the relationships between social class and postsecondary pathways (e.g., transfers, interrupted patterns) constitute additional layers of stratification (Goldrick-Rab 2006; Goldrick-Rab and Pfeffer 2009). Still others reveal how various forms of horizontal inequality (e.g., field of study, selectivity, public/private sector) within college further reproduce social class inequalities (for a review, see Gerber and Cheung 2008). Timely completion, we argue, constitutes an additional marker of differentiation and stratification, and policymakers and education officials need to be more vigilant in monitoring attrition and time to completion among low-socioeconomic status (SES) youth and perhaps devise specific policies aimed at improving those educational outcomes.
Finally, extended completion times may be particularly salient issues of concern for less selective segments of the higher education system. While expansion has indeed allowed more previously underrepresented populations into the system, many of these new participants are largely confined to accommodating lower ranked institutions, which do not yield comparable returns to elite institutions (Marginson 2016). Rising student demand from new populations in the higher education system has given rise to new avenues of postsecondary education that were simply not options in a previous era (Davies and Mehta 2018). Many students today do not take the “traditional” route to obtain their degrees and often look to balance their education with concurrent employment demands and family responsibilities (Deil-Amen 2015). Rather, these “nontraditional” students typically begin their studies in their mid-20s, are more likely to pursue their degrees on a part-time basis, and have families and dependents (Deil-Amen 2015). These older students who wear multiple hats may be especially prone to lengthy degree times. Less selective, “broad access” institutions have responded to these individuals’ nonconventional pathways through higher education and the irregular rhythm and complexity of their life course transitions (Settersten 2015; Stevens 2015). They offer expansive student supports and flexible degree options to accommodate these students’ “circuit board” style of school-work transitions versus the traditional “pipeline” connection (Davies and Mehta 2018). Therefore, it is interesting to also examine the factors influencing the timely completion of nontraditional students within four-year institutions.
In this article we ask, first, to what extent do some individuals, more than others, take longer to complete their degrees? Specifically, we examine the extent to which low-SES students take longer to finish their undergraduate degrees in an expanded postsecondary system. Second, how pervasive is the influence of SES on one’s likelihood of completing a degree on time after controlling for academic performance, aspirations, educational characteristics, and the type and level of institutional supports? As may be the case with timely completion, the effects of SES on postsecondary entry, entry into particular fields of study, entry into selective institutions, degree and program completion, and/or entry into graduate and professional schools are often direct and indirect, exerting their influence through correlates with SES. Finally, we contribute by specifically examining new college populations who may be particularly at risk of longer completion times. We ask, to what extent do the aforementioned relationships vary across traditional and nontraditional age students?
We draw on the most recent cohort of the Baccalaureate and Beyond Longitudinal Survey (2008–2009) to examine the relationship between SES and the probability of completing a bachelor’s degree on time with nationally representative survey data across multiple sectors of the U.S. higher education system. Existing studies that have looked at time to completion tend to be descriptive in nature, do not attempt to isolate key factors or control statistically for other factors at play, and focus on relationships at only a limited number of institutions or institutional sectors (e.g., Arcidiacono, Aucejo, and Spenner 2012; DesJardins, Ahlburg, and McCall 2002; Knight 2004; Pitter et al. 1996; Volkwein and Lorang 1996).
Review of the Literature
We begin by reviewing several bodies of work that have largely been investigated independently in the sociology of education literature, but each offer theoretical and substantive findings to situate expectations on how social class might impact timely completion and how this relationship varies across traditional and nontraditional age students.
Expansion and Differentiation within College
As higher education continues to expand and access opens up for more and more traditionally marginalized groups, a growing body of literature on “horizontal inequality” underscores the need to keep a close eye on the ways in which disadvantaged groups may be selected into and out of the various segments of the postsecondary system in large part due to their social class background. Expansion has typically not been accompanied by equal access to elite institutions, and new populations who participate may find themselves in mass higher education institutions or sectors of lesser quality (Marginson 2016). Once inside the halls of higher education, students may be further differentiated by additional stratifying mechanisms such as public/private sectors, fields of study, and hierarchies of institutions (Gerber and Cheung 2008; Goldrick-Rab 2006; Goldrick-Rab and Pfeffer 2009; Marginson 2016; Roksa and Velez 2010; Triventi 2013; Weiss and Roksa 2016). While the underlying processes whereby inequality is either maximally or effectively maintained remain contested (see Lucas 2001, 2009, 2017; Shavit, Arum, and Gamoran 2007), 1 a substantial number of studies find that as higher education expands, social class continues to exert its influence on student choices and outcomes, though often in less obvious ways. Whereas in a previous era, the issue was one of opening up access, in expanded systems of higher education, it has become additionally important to monitor more fine-grained points of selection and differentiation such as entry into selective colleges and lucrative fields of study and as we argue here, timely completion of programs. In expanded systems, the presence of these differences within the system offers more complex challenges to policymakers and reformers than simply relying on expansion as a strategy to reduce inequality (Lucas 2017).
In many countries, SES continues to exert both direct and indirect effects on postsecondary entry (Jackson 2013); college transfer and stopout (Goldrick-Rab 2006; Goldrick-Rab and Pfeffer 2009); entry into particular fields of study (Ayalon and Yogev 2005; Goyette and Mullen 2006; Triventi, Vergolini and Zanini 2017; Zarifa 2012a); entry into highly selective, highly ranked, and well-resourced institutions (Davies and Guppy 1997; Davies, Maldonado, and Zarifa 2014; Zarifa 2012b); degree and program completion (Baker and Vélez 1996; Byun, Irwin, and Meece 2012; Diemer and Li 2012; Doren and Grodsky 2016; Goldrick-Rab 2016; Hamilton 2012; Mendoza 2012); and entry into various levels within the system such as graduate and professional schools (Mullen, Goyette, and Soares 2003; Torche 2011; Zarifa 2012c). Yet, the impact of social class on timely completion has received minimal attention. To what extent do these inequalities at various points of differentiation within the postsecondary journey reflect relationships between social class and completing one’s degree on time? If social class does indeed influence one’s probability of completing a bachelor’s degree in four years, the nature and extent of this relationship is unclear.
On the one hand, individuals from higher parental SES backgrounds may strategically take longer to finish their degrees. Time to degree delays might be the result of conscious decisions made by students from middle- and upper-class families to secure better positions in the labor market. Prior research suggests that some students may intentionally take on reduced course loads, prolonging the length of their degrees as a way of maintaining their competitive edge (i.e., high GPA) for entry into graduate and professional schools (Volkwein and Lorang 1996). Others may intentionally take on lighter course loads to balance work and family responsibilities, maintain their GPA, explore their life options, and/or enjoy college life (Volkwein and Lorang 1996). Still, it is unclear if such action is advantageous, as other research has indicated that it is course difficulty—not course load—that negatively influences GPA and student retention (Szafran 2001).
On the other hand, we may expect to see individuals from low parental SES backgrounds taking longer to complete their degrees. Timely completion by high-SES students might be part and parcel of larger processes of class reproduction. Scholars working within a Bourdieusian framework suggest that parents lacking a familiarity with the ins and outs of the postsecondary system may not be able to help their children navigate the system as smoothly as those from higher SES backgrounds (Bourdieu [1984] 2010; Bourdieu and Passeron [1977] 1990; Lareau and Weininger 2003). One’s habitus and levels of social and cultural capital may exacerbate class differences throughout one’s undergraduate years (Armstrong and Hamilton 2013; Jack 2016; Lareau and Weininger 2008; Lee and Kramer 2013; Lehmann 2014), resulting in longer completion times. Lacking the appropriate resources to “culturally handshake” members of higher-class backgrounds (i.e., professors and peers) may create barriers that hamper academic successes and lead to stopout (Davies and Mehta 2018) and dropout (Lehmann 2007) for those who do not have access to culturally relevant capital.
Parental SES and College Success
Recent higher education scholarship points to a host of mechanisms through which parental advantages might impact student successes during college. Many of the advantages provided by parents go relatively unnoticed, a situation some have referred to as “laundering privilege” (Stevens 2007). Involvement in extracurricular activities and socializing during college may be particularly important for student achievements (Armstrong and Hamilton 2013; Rivera 2015; Stuber 2011). Stuber (2011), for example, argues that the “experiential core” of what happens during college is a key stratifying process that occurs even within college. While working-class students are more likely to see higher education for primarily what it offers by way of instruction in the classroom, middle- and upper-class students acquire additional social and cultural capital by actively socializing and participating in extracurricular activities that in turn influences their acquisition of more social, cultural, and human capital (Stuber 2011). Greater social and cultural capital snowball during college and yield additional benefits that lead to further success (e.g., finding a comfortable residence, connecting with similar peers, off-campus study, internships, and community service). Similarly, Armstrong and Hamilton (2013) show that working-class students may miss out on the valuable social connections and information about how to navigate college that privileged students obtain by engaging in the “party” and social culture. These lower levels of extracurricular engagement among low-SES students may result in feelings of isolation and alienation that ultimately impact students’ grades and their likelihood of completing their degrees (Rivera 2015).
Other scholars point to social class differences in engaging with professors during college. Jack (2016), for example, found that social class background influences one’s likelihood of engaging with one’s professors, which ultimately leads to more positive academic outcomes. Further, even among lower-income undergraduates, significant heterogeneity exists. Lower-income students who had attended boarding schools (i.e., privileged poor) were significantly more likely than other lower-income undergraduates (i.e., doubly disadvantaged) to actively engage and approach their professors.
Other studies have highlighted the necessity of parental involvement throughout the college years for student success, especially as public higher education becomes increasingly privatized in the United States (Hamilton 2016; Hamilton, Roksa, and Nielsen 2018). Extending the theoretical underpinnings of Bourdieu and Lareau at the postsecondary level, their findings suggest that parenting approaches during college have an incredible influence on student outcomes. It is not enough for parents to provide their children with high levels of funding, but savvy parental involvement is also important. In fact, Hamilton (2012) notes that parent income/investments during college increased the likelihood of completing a bachelor’s degree in five years but that greater parental contributions also decreased GPA. Hamilton (2012) concludes that “more” does not always lead to better college outcomes. In fact, students with parental funding may “satisfice”—curb extensive efforts academically in order to get by and complete college without aiming for academic excellence.
Many tasks have been “outsourced” to parents, and success in today’s colleges requires a “family-university partnership” (Hamilton 2016). Parents have to take on the role of “academic advisor, career counselor, therapist, and life coach” (Hamilton 2016:194). Hamilton et al. (2018) describe the role of upper- and upper-middle-class parents as the “college concierge”—parents who are very involved with their child’s college experience, from academic/social support (college selection, residence selection, academic guidance, extracurricular activity involvement, frequent contact, monitor/intervene with problems) to assistance after college (internship assistance, insider career knowledge, connections). By contrast, parents from middle, lower-middle-, and working-class backgrounds are more likely to act as the “college outsider.” These parents expect the university to provide more supports than that which is actually provided, only have limited contact with their children, are not made aware of problems in their child’s academics, and do not know what support programs are available (Hamilton et al. 2018). The end result is poorer student outcomes coupled with a feeling of betrayal toward the university.
Taken together, the barriers to undergraduate success uncovered in this literature might lead lower-SES students who do manage to persist through their undergraduate years to take longer to do so. Yet, explicit connections between parental actions and resources to timely completion have been largely absent among these works.
Correlates with Parental SES
The impact of parental social class background on completing a degree on time may be largely expressed indirectly via correlates with SES. As mentioned previously, scholars working on issues of “horizontal inequality” have understood the role parental SES plays in predicting educational expectations, academic performance, field of study and college major choice, and college selectivity (Armstrong and Hamilton 2013; Ayalon and Yogev 2005; Davies et al. 2014; Davies and Guppy 1997; Goyette 2008; Goyette and Mullen 2006; Mullen et al. 2003; Triventi et al. 2017; Zarifa 2012a). Not only do these educational avenues further stratify students within credential tiers, but they in turn impact timely completion.
Scholars working on college persistence, attrition, and completion have given considerable attention to the impact of student employment on educational outcomes (Bound, Lovenheim, and Turner 2010; Engle and Tinto 2008; Kouliavtsev and Austin 2013; Mendoza 2012). For many college students, especially those from lower social class backgrounds, working concurrently with one’s degree pursuits has become the norm. The financial costs associated with attaining a college degree have simply outpaced financial aid. While some students do receive financial aid, many students from low-income families are still required to make considerable investments to pursue postsecondary education and must work concurrently with their studies (Engle and Tinto 2008; Mendoza 2012).
This additional time spent working and less time spent on campus by lower-SES students can lead to adverse outcomes. For Engle and Tinto (2008), work commitments can impede integration during college. First-generation, low-income students are less likely to interact with faculty and peers through study groups, extracurricular activities, and support services (academic and social experiences that typically foster successful college completion; Engle and Tinto 2008). The situation becomes particularly problematic for students who work more than 30 hours a week as they may be less likely to graduate in six years (Mendoza 2012). Unfortunately, this is the situation for nearly one-third of dependent students surveyed in the 2007/2008 National Postsecondary Student Aid Survey (Mendoza 2012). Moreover, low-income groups who engage in high-intensity employment and live with their parent/other relatives as a cost-saving strategy were less likely to persist in their postsecondary education (Bozick 2007).
For students from high-SES backgrounds, the strategies are quite different. Roksa and Velez (2010) revealed that students whose parents have a bachelor’s degree or higher are more likely to enter higher education and have limited labor market participation. The authors argue that extensive work commitments concurrent with college lead to differentiation of educational outcomes across social classes and constitute an additional way in which inequality is effectively maintained. These findings are echoed in more recent work that zeroes in on four-year colleges, lending further support to the importance of work as a differentiating mechanism across social classes in expanded systems (Weiss and Roksa 2016).
Institutional Characteristics
Perhaps the obvious solution to reducing work commitments among low-SES students would be to increase the level of financial aid provided by institutions. Quite simply, more financial aid could reduce the time students need to spend working and allow them to devote additional time to their academic pursuits. Goldrick-Rab et al. (2016) revealed, for example, that more student aid increased rates of college completion. But, it was students whose parents had high levels of education in particular who were able to mediate the impact of additional resources by using their cultural and social capital to translate economic boosts into shorter completion times (Goldrick-Rab et al. 2016). Moreover, student aid in the form of grants rather than loans has been shown to be the most effective (Armstrong and Hamilton 2013; Goldrick-Rab 2016; Tinto 2012), increasing the probability of on-time completion at four-year universities by about 28 percent (Goldrick-Rab 2016).
Timely degree completion may also vary by the selectivity and rank of the student’s institution. Students who first enter highly selective institutions (Adelman 2006; Snyder and Dillow 2012), research-intensive universities (Bound et al. 2010), and schools with academic partnerships that offer opportunities to learn abroad (Jean Francois 2016) typically exhibit significantly shorter times spent completing their undergraduate degrees. While both high- and low-SES students with high academic achievements are more likely to progress successfully through the system, high-SES students may have a greater “institutional reach” and enter into more selective institutions (Andrew 2017).
Jenkins and Rodriguez (2013) argue that in a time of scarce resources, broad-access institutions may be sacrificing efficiency and productivity with respect to students’ timely completion to accommodate the rising rate of enrollment. Similarly, Davies and Mehta (2018) theorize that wider calls for expansion may have triggered two new logics in higher education systems. On the one hand, in the most desired higher education sectors, an “intensifying logic” may be emerging, where heightened competition and demand drive up premiums for the most exclusive credentials. Selective schools respond to increased student demands by essentially becoming even more selective and their credentials more exclusive. On the other hand, institutions with low levels of selectivity target new populations entering the system that have unprecedented needs and may trigger a second response, an “accommodating logic.” Among these mass access institutions, boosting enrollments through untapped markets is a common goal as lower ranked institutions adapt and adopt new practices and supports to encourage persistence, reduce attrition, and support timely completion (Davies and Mehta 2018). These accommodating schools, of course, facilitate greater proportions of nontraditional students.
Still, other work suggests that students’ time to degree could actually be longer in selective colleges than less selective institutions (Byun et al. 2012; Martinello 2009). Martinello (2009) attributed differences to more liberal regulation of course selection and program switching in selective colleges that allowed students to “mill around” to find suitable fields of study. Another explanation for varied outcomes might hinge on the culture of the institution. The organizational habitus of some colleges might offer environments conducive to building additional capital (e.g., capital building part-time work, mentorship, integrated residence-extracurricular life; Stuber 2011). The social supports that exist can greatly influence undergraduate success, particularly that of first-generation, working-class students (Stuber 2011).
Finally, time to BA completion rates may vary across the type of institution. Over time, the largest decreases in the proportion of students who completed in four years have been among the public four-year and public two-year institutions (Bound et al. 2007). Moreover, Snyder and Dillow (2012) found that graduates of public two-year and four-year institutions are less likely to complete within 150 percent of the expected time-to-completion window than graduates of comparable private institutions—a finding that might reflect the higher concentrations of lower SES students these public institutions typically serve.
The Nontraditional Student
Nontraditionally aged students represent an important “at risk” group when considering timely completion. In many respects, the odds of completing their degrees and doing so in a timely fashion are disproportionately stacked against them. Many factors impede one’s likelihood of completing an undergraduate degree on time, and nontraditional students are most likely to encounter these obstacles. In addition, older postsecondary students typically also have multiple obligations outside of postsecondary pursuits that limit their ability to fully participate in the requirements of college (Adelman 2006; Deil-Amen 2015; Engle and Tinto 2008; Settersten 2015). 2
Not only do they enter the higher education system at an older age and come from lower SES backgrounds, but these nontraditional students are more likely to continue their formal education well beyond their 20s, have dependents, live off -campus, commute, take their studies part-time, and work full-time (Davies and Mehta 2018; Deil-Amen 2015). The shear juggling of multiple life course events (e.g., work transitions and family transitions) negatively affects college completion (Roksa and Velez 2012). That is, getting married/cohabiting and having children significantly decreases the likelihood of completing a college degree. Moreover, timing also matters; when employment and family transitions occur after college entry, students show significantly lower odds of completion (Roksa and Velez 2012).
At the same time, many older students have “messy vitaes” characterized by multiple attempts to complete higher education (Davies and Mehta 2018). These nontraditional attendance patterns add an additional layer of stratification for low-SES individuals to overcome in their postsecondary pursuits (Goldrick-Rab 2006). Specifically, low-SES students are more likely to have interrupted movement patterns. High SES students, by contrast, are more likely to have fluid movement (attend multiple schools without stopping out; Goldrick-Rab 2006). The influence of social class on college transfer patterns highlights yet another aspect of tertiary-level differentiation facing nontraditional students. Even those from low-SES backgrounds who enter a four-year college may find themselves in a two-year college due to weaker academic performance in their freshmen year (Goldrick-Rab and Pfeffer 2009).
Yet, for these older students, timely completion may be particularly important as they are likely to have little time and money to “waste” and comparatively less time in the workforce to reap the benefits of their education (Settersten 2015). From a life course perspective, the timing of educational achievements also impacts later life outcomes in many respects. Elman and O’Rand (2004), for example, suggest that individuals from advantaged social origins are more likely to complete their postsecondary degrees earlier in life, which in turn leads to higher wages by midlife. That is, obtaining valuable resources early (i.e., education) leads to higher cumulative returns over time (Elman and O’Rand 2004). In contrast, individuals from disadvantaged backgrounds may exit school early in life, earn lower wages, and not return to upgrade their schooling later on. Moreover, even those from advantaged origins who return to school do not reap as much reward as those who did not delay their schooling (i.e., fast trackers).
While not framed in the discourse of the nontraditional student, empirical work from the dropouts and completions literature does point to age as an important predictor. On the one hand, studies suggest older, mature students are more likely to resist dropping out and complete their programs. For example, Tumen, Shulruf, and Hattie’s (2008) exploration of the pathways of university students in New Zealand focused on identifying students who were at risk of not completing their bachelor’s degrees. Ultimately, the age of students was shown to be the only factor that systematically predicted a student’s likelihood of completing a program, while gender, ethnicity, and SES showed little direct effects when accounting for measures of student achievement and educational experience (Tumen et al. 2008). More recent research also points to the importance of age effects, highlighting that the first-term credit completion ratio among adult students may be a stronger predictor of academic persistence than either student characteristics or even environmental factors (e.g., hours of employment and familial responsibilities; Davidson and Holbrook 2014).
However, graduates’ persistence may come at the expense of additional years in the program. Adelman (2006), for example, examined the time to degree among recipients of bachelor’s degrees in 1999–2000 using the NELS:88/2000 cohort. His results revealed that older students were significantly more likely to take longer to finish their degrees. Specifically, 62.8 percent of bachelor’s degree recipients who were over 40 in 2000 started out in postsecondary education before 1980. The elapsed time to degree for these people was 20 years (the average was over 27 years). For those who were between 30 and 39 in 2000, 73.3 percent entered postsecondary education prior to 1990, suggesting an elapsed time to degree for these people was 10 years (the average was nearly 15 years; Adelman 2006).
As four-year colleges continue to structure their classes, living arrangements, and extracurricular activities around the traditional college student with the expectation that parents play a more salient supporting role (Hamilton 2016), we expect that those nontraditional students who do make it into the four-year colleges would have significant difficulty completing their degrees on time. But, empirically it remains to be seen which of these obligations have the greatest impact on the timely completion of their bachelor’s degrees. As such, our study addresses this knowledge gap by examining not only the effect of age on timely degree completion but also the extent to which the predictors of timely completion vary across traditional and nontraditional age students.
Methods
Data
To examine how social class influences the time to complete a bachelor’s degree, this study draws on the restricted data files of the 2008–2009 Baccalaureate and Beyond Longitudinal Study (B&B) from the National Center for Education Statistics (NCES) in the United States. The B&B is a nationally representative sample of recent bachelor’s degree graduates and provides extensive information on the past educational experiences, future education and labor market expectations, and postbaccalaureate activities of respondents (e.g., job search/training, entry into graduate school, and loan repayment/financial status). Of particular importance is the linking of transcript data to survey responses, which eliminates the known inconsistencies of self-reported academic performance data.
Respondents in the B&B 2008 cohort were followed up in 2009 and 2012. 3 The 2008–2009 B&B provides information on the educational experiences of a cohort of recent baccalaureate graduates, who received their degrees during the 2007–2008 academic year. The B&B is a multistage, stratified sample consisting of three stages. First, a sample of 1,960 institutions was derived from 46 institution-level strata from the 2004–2005 IPEDS Institutional Characteristics, Fall Enrollment, and Completions files. From these 1,960 institutions, 1,940 were deemed eligible to participate in the 2008 National Postsecondary Student Aid Study (NPSAS). Second, the NPSAS respondents were limited to only those students who were attending the sampled institutions and were enrolled in a program that led toward a degree or formal award and had an enrollment duration of at least three months. Finally, a subsample of 18,500 baccalaureate degree recipients (the B&B) was selected from the NPSAS respondents, and a final total of 15,090 individuals completed a full interview (88 percent; for details, see Wine et al. 2013). Respondents were asked a series of questions via computer-assisted telephone interviews (CATI), and additional information was obtained from undergraduate transcripts supplied by the students’ postsecondary institution. 4
Restrictions and Subsample
A number of restrictions were placed on the analyses. Specifically, this study includes only students who completed their degree in the 2007–2008 cohort, did not previously obtain a bachelor’s degree prior to this degree, were US citizens 5 , and were paired with institution-level data and transcript data. 6 As such, our final subsample for the analyses was 12,500—11,650 traditional students and 850 nontraditional students. 7
It is important to note that our subsample is comprised of a relatively privileged group, and overall system-wide inequalities for disadvantaged groups may be underestimated in our results here for a couple of reasons. First, the B&B sample includes only those students who successfully graduated. Thus, we do not have information on students who may have entered the higher education system but never graduated (i.e., dropouts). As such, our study is an examination of those who complete on time within a population of baccalaureate graduates. Second, by limiting our scope to only those at four-year colleges, our analyses presented here likely underestimate the effects of both student demographics and institutions since students from disadvantaged backgrounds are more likely to drop out and more heavily concentrated in less selective, two-year institutions (Brint and Karabel 1989; Engle and Tinto 2008; Hermanowicz 2003; Settersten 2015; Stevens 2015).
Analyses
Our statistical analyses include descriptive statistics as well as binary logistic regression models (Fox 2015; Long 1997; Long and Freese 2014). To aid in the interpretation of the binary logit results, all regression estimates are reported as marginal effects, and predicted probabilities along with 95 percent confidence intervals are calculated and presented in graphical displays for the key explanatory variables in our analyses. Specifically, we estimate three series of models. The first series includes all respondents and predicts whether the graduates completed their four-year bachelor’s degrees on time. To investigate the extent to which the factors influencing timely completion vary across our two key subpopulations, our second and third series of models examine these relationships separately for traditional and nontraditional age graduates. 8 Based on existing studies described previously, a number of key explanatory variables are entered in several stages. To investigate the extent to which social class may have direct or indirect effects on timely completion, initial models include only SES measures, and subsequent models add sociodemographic variables, measures of ability and aspirations, as well as variables that capture program experiences, hours of employment during their undergraduate degree, and the institutional characteristics of their alma mater. 9
Variables
Many of the variables of interest were selected due to their prominence in the sociological studies conducted in the United States, as previously reviewed. Our key response variable across models captures whether the respondents completed their baccalaureate degree on time. As such, we employ a dichotomous response variable that indicates whether graduates from four-year bachelor’s degree programs took longer than four years to complete their degree. 10
A number of key explanatory variables were entered into our models in several stages. Our first model includes measures of SES, and subsequent models add sociodemographic variables including respondent’s sex, racial background or ethnicity, disability status, marital status, dependent children, traditional or nontraditional age at start of degree, 11 as well as theoretically and empirically supported measures of ability, aspirations, program characteristics, and institutional characteristics. As in previous sociological studies using the B&B surveys, the influence of SES on time to completion was measured using both parent’s level of education and parental income (Goyette and Mullen 2006; Zarifa 2012a, 2012b). 12 All analysis variables and categories are shown in Table 1.
Descriptive Statistics for Variables from the 2008–2009 Cohort of the Baccalaureate and Beyond Longitudinal Survey of University Graduates in the United States.
Note: Estimates and standard errors are survey weighted using bootstrap replicate weights. As per National Center for Education Statistics requirements, reported sample sizes have been rounded.
Existing studies also suggest that educational aspirations and academic ability may be important predictors of completing a bachelor’s degree on time. To explore the extent to which those with expectations of continuing beyond the undergraduate level are more or less likely to take longer in their degrees, whether or not the respondents had expectations beyond the BA is included as a dichotomous variable. To measure academic ability, we employ a standardized measure of graduates’ cumulative undergraduate GPA, derived from student transcripts provided by the graduates’ institution.
A number of key educational and program-related factors were also included in the models: field of study (seven categories), attending a co-op/internship (dichotomy), 13 average weekly job hours during 2007–2008, and two derived variables by NCES to indicate whether the graduate had stopped out (i.e., took a break in enrollment of more than four months before the bachelor’s degree) and whether the graduate had formally changed their major since first declaring a major at the 2007–2008 bachelor’s degree–granting institution. Many students simply do not know in advance what they want to do in higher education and have been told that college is the time to explore. However, field switching or not declaring a major until after the first couple of years puts added pressure on students’ timely degree completion (Settersten 2015).
Finally, a number of institutional characteristics were also included in the analyses. Previous research suggests that completing a degree on time may be related to the type of institution and supports that are available (Bound et al. 2007; Davies and Mehta 2018; Doyle 2009; Knight 2004; Settersten 2015; Volkwein and Lorang 1996). To explore the variation in student outcomes across institutions, school selectivity (measured as nonselective, minimum selectivity, moderate selectivity, and high selectivity), 14 enrollment size of the institution (logged), sector and type of the school, and a ratio of institutional to total aid (logged) were included as predictors. 15
Results
Characteristics of the 2008 Cohort
Table 1 shows the distributions of our variables of interest from the 2008–2009 B&B survey. Overall, our summary statistics in Table 1 indicate that the majority (56 percent) of graduates in the B&B took longer than four years to complete their undergraduate degrees. Moreover, the average length of time taken to complete a bachelor’s degree was 6.33 years. On their own, these figures are quite troublesome. Yet for policymakers and education officials to implement effective supports to ensure timely progression through programs, it is important to further identify the extent to which some sociodemographic groups are overrepresented among these graduates. We return to this issue in our logistic regression results that follow.
In terms of sociodemographic characteristics, the estimates in Table 1 indicate that the majority (58 percent) of bachelor’s degree graduates were female. This finding aligns closely with previous work on college enrollment and completions (Buchmann and DiPrete 2006; Buchmann, DiPrete, and McDaniel 2008). For race, 75 percent of respondents were white, approximately 9 percent were Hispanic, 8 percent were black, about 5 percent were Asian, and just over 3 percent belonged to other races and ethnicities. A large majority of graduates were single, as only 15 percent stated being in a marriage or common law situation. About 8 percent of graduates reported having a disability, while 55 percent stated that they had one or more dependent children. In terms of age, about 94 percent of the sample started their degree at a traditional age, and about 6 percent were nontraditional (i.e., 21 or older when they started their baccalaureate). In terms of parental education, roughly 56 percent of graduates had one or more parents who completed at least a baccalaureate degree. For parental income, about 18 percent of graduates came from low-income families. Finally, 75 percent of graduates expected to pursue studies beyond the bachelor’s level.
For program and educational experience characteristics, roughly 21 percent of graduates had attended a co-op program or apprenticeship as part of their undergraduate degree. Nearly one-third of graduates changed majors during their degrees. In terms of field of study, 22 percent graduated with business degree; 16 percent from the social sciences; over 13 percent from math-, science-, and health-related fields; about 13 percent from the humanities; roughly 7 percent from engineering and computer science as well as education; and roughly 21 percent obtained degrees from other fields. Table 1 also indicates that students in the B&B worked an average of 20.75 hours per week during their final year of their undergraduate degree (2007–2008). In terms of stop out, approximately 20 percent of graduates stopped out en route to their bachelor’s degree.
In terms of institutional characteristics, Table 1 reveals that over 30 percent of the sample graduated from a very selective institution, 52 percent from a moderately selective institution, about 10 percent from a minimally selective, and about 9 percent from a nonselective institution. The majority (62 percent) of students attended publicly funded, four-year institutions.
Who Completes on Time?
To investigate the extent to which social class has an impact on completing a degree on time, we turn to our binary logistic regressions. 16 Table 2 provides the logistic regression models for the entire B&B sample. For discrete (categorical) predictor variables, the marginal effects displayed in the tables represent the change in predicted probabilities in comparison to the reference category; for continuous variables (e.g., GPA), the estimates represent the instantaneous rate of change in the probability of completing a degree on time (see Long and Freese 2014). Wald tests are estimated and reported adjacent to variable names for sets of dummy regressors.
Binary Logit Models Predicting the Likelihood of Completing a Four-year Bachelor’s Degree on Time (marginal effects shown), 2008–2009 Baccalaureate and Beyond (B&B) Longitudinal Survey.
Note: Multiple parameter Wald tests are reported adjacent to variable names for sets of dummy regressors. Reported values are based on weighted estimates. As per National Center for Education Statistics requirements, reported sample sizes have been rounded.
p < .10. *p < .05. **p < .01. ***p < .001.
Model 1 includes only our measures of parent SES. As expected from our review of the literature, our multiple parameter Wald tests indicate that both parental income as well as highest level of parental education exhibit strong and statistically significant (p < .001) effects on one’s probability of completing a degree on time. Across both measures, students with higher SES backgrounds were significantly more likely to complete their programs on time.
Model 2 introduces several sociodemographic variables. As in Model 1, the SES effects remain statistically significant when sociodemographic variables are included in the model. In addition, women have a .09 higher probability of completing their degree in four years than men (p < .001). At the same time, Hispanics (p < .05) are significantly less likely to finish their degrees on time. Not surprisingly, graduates who are married and have dependent children are also less likely to complete their degrees on time (p < .001). Finally, in an additional model (not shown) with only our variable that distinguishes between traditional and nontraditional age graduates, the effect is highly statistically significant (p < .001) and confirms our expectation that traditional students are much more likely to complete their degrees on time. However, when including these additional factors in the model shown in Table 2, the effect is no longer statistically significant. This is perhaps less surprising since many of the challenges of nontraditional graduates are attributable to family responsibilities (Deil-Amen 2015; Settersten 2015). 17
Measures of educational expectations and academic ability are added in Model 3. Both educational expectations (p < .01) and GPA show strong and significant (p < .001) positive effects on one’s likelihood of completing a degree in four years. As in Models 1 and 2, women remain significantly more likely to finish their bachelor’s degrees on time (p < .001), albeit the effect does weaken slightly with the addition of expectation and ability. In terms of SES, both parent’s education as well as parental income (p < .001) show independent and statistically significant effects on whether the respondents completed their degrees on time even when controlling for the additional measures of educational expectations and academic performance. Finally, the effects of both marital status and the dependent children variables remain highly statistically significant (p < .001).
In Model 4, a number of educational and program characteristics are added to the models. The results indicate that field of study has a strong and significant impact (p < .001) as to whether or not graduates finished their bachelor’s degrees on time. Specifically, graduates from engineering and computer science (p < .01) and education (p < .001) are significantly less likely than business degree graduates to complete their degrees on time. At the same time, social science graduates are more likely (p < .01) than business graduates to complete their degrees on time. School selectivity also shows strong and significant effects on time to completion (p < .001). Those who attended nonselective, minimally selective, and moderately selective schools are significantly less likely (p < .001) than their counterparts who attended highly selective schools to complete their degrees on time. While we cannot determine the extent to which work status impacts integration (e.g., Engle and Tinto 2008), working additional hours concurrent with pursuing one’s degree significantly decreases the odds of completing the bachelor’s degree on time (p < .001). In addition, those who participated in a co-operative education program had a .031 higher probability (p < .05) of completing on time, while those who stopped out during their degrees had a significantly lower probability of finishing on time (p < .001). The inclusion of these educational program and degree-related behaviors does not change the previously noted sociodemographic inequalities. In terms of SES, we observe additional weakening of the parent education and parent income effects, albeit the estimates remain highly statistically significant (p < .001). This resonates with prior research on the impact of SES on entry into particular fields of study and selective colleges (Davies et al. 2014; Goyette and Mullen 2006; Zarifa 2012a, 2012b) but also suggests that part of the delays in time to completion for low-income students may be attributable to increased hours of employment to offset rising college costs (Bound et al. 2010; Roksa and Velez 2010; Weiss and Roksa 2016) or stopping out during their studies (Davies and Mehta 2018; Goldrick-Rab 2006).
Finally, in Model 5, we include a number of key institutional characteristics. The degree to which schools offer financial aid to their students has a strong and positive impact (p < .001) on students’ likelihood of completing their programs on time. As in work with the NELS (e.g., Knight 2004), there is also evidence to suggest that the type of institution matters (p < .01). Graduates from not-for-profit private (p < .01) and religious four-year institutions (p < .01) are significantly more likely to complete their degrees on time compared to their counterparts who attended public institutions.
Interestingly, these saturated models show few differences in the previously noted relationships across demographic and SES variables. The effect of parental education, however, does weaken further (p < .05), which could suggest that highly educated parents play a mediating role in their children’s financial aid (see Goldrick-Rab et al. 2016). To further understand these findings, Figures 1 and 2 display the predicted probabilities of parent’s education and parent income from Models 1 and 5. For both parent education and income, the results in Model 1 show fairly linear, positive relationships between SES and the likelihood of completing on time. However, once we control for all other factors in Model 5, the effects of both SES measures weaken.

Predicted probabilities from binary logit models predicting bachelor’s degree completion on time by parent education.

Predicted probabilities from binary logit models predicting bachelor’s degree completion on time by parent income.
Moreover, it is also clear from Model 5 that beyond these sociodemographic factors, a number of program-related factors also have a significant impact. Academic performance, field of study, school selectivity, and hours worked all continue to have independent effects on completing degrees on time (p < .001). Figure 3 displays the predicted probabilities across fields of study. Interestingly, graduates in the social sciences (.53) show the highest probability of completing on time, followed by a second group of business; math, sciences, and health; and the humanities (all between .43 and .45). Engineering and computer science (.35) and education (.35) majors showed the lowest probabilities of completing on time. This finding is particularly interesting as liberal arts graduates typically experience lower economic returns (Davies and Guppy 1997; Triventi et al. 2017), while graduates of STEM fields typically do quite well in their early employment experiences (see Melguizo and Wolniak 2012). Yet, the higher earnings enjoyed by college graduates of STEM fields, especially those with degrees in engineering, may come with the cost of taking more time than expected to complete their degrees.

Predicted probabilities from binary logit models predicting bachelor’s degree completion on time by field of study.
Finally, Figure 4 displays the predicted probabilities across schools of varying selectivity. In step with prior work (Adelman 2006; Bound et al. 2010; Snyder and Dillow 2012), students in highly selective schools show significantly higher probabilities (.52) of completing on time than their counterparts at less selective institutions.

Predicted probabilities from binary logit models predicting bachelor’s degree completion on time by school selectivity.
One Size Fits All? Comparing Relationships for Traditional and Nontraditional Students
To investigate the extent to which these relationships may differ for traditional students and nontraditional students, we estimate the same sets of models discussed previously separately for traditional and nontraditional students. At the outset, the raw differences in timely completion are striking across these two groups as only 16 percent of nontraditional students finished on time, compared to 46 percent of traditional college students (see Appendix A.1). In Table 3, the results for traditional students are shown. As in the previous analysis, Model 1 includes the two measures for SES, parental education and parental income. For traditional graduates, both of these measures are statistically significant (p < .001). In Model 2, when the remaining sociodemographic variables are added, the effects of both parental education and parental income remain statistically significant at their respective levels (p < .001). Among the new variables in the model, the effects of sex, dependent children, and marital status are also statistically significant (p < .001). The effect of race and ethnicity is also statistically significant (p < .05), as is our age variable (p < .05). That is, even those students who have delayed their schooling and started at age 19 or 20 are also less likely to complete on time compared to those who fast tracked or experienced few delays (age 15–18).
Binary Logit Models for Traditional Age Predicting the Likelihood of Completing a Four-year Bachelor’s Degree on Time (marginal effects shown), 2008–2009 Baccalaureate and Beyond (B&B) Longitudinal Survey.
Note: Multiple parameter Wald tests are reported adjacent to variable names for sets of dummy regressors. Reported values are based on weighted estimates. As per National Center for Education Statistics requirements, reported subsample sizes have been rounded.
p < .10. *p < .05. **p < .01. ***p < .001.
Model 3 includes additional variables for academic performance and educational expectations. Specifically, those who expect to pursue additional schooling are more likely to complete their degrees on time (p < .05), as are students with higher GPAs (p < .001). With the exception of race, which is no longer statistically significant, the effects of the other variables in the model remained statistically significant at their previous levels.
Model 4 includes the school-related variables, field of study, school selectivity, whether the respondents participated in a co-op program, hours worked, and whether the students switched majors. Similar to the estimates for this model in the pooled analysis, the effect of field of study is highly statistically significant (p < .001), as are the effects for the school selectivity, job hours (p < .001), and stopping out (p < .001). Even when these variables are added to the model, the effects of sex, marital status, the presence of children, academic performance, and parental income remain statistically significant at their prior levels (p < .001), while the parental education effect weakens slightly (p < .01).
In Model 5, the institutional characteristics are included. As with the pooled model, the effect of enrollment size is not statistically significant. Likewise, the extent to which schools offer financial aid to their students has a positive and statistically significant impact (p < .001) on whether students will complete their programs on time. In terms of institutional sector, students who attend private not-for-profit and religious institutions are more likely to graduate on time than students who attend public institutions (p < .001). The effects of the other variables in the model remained statistically significant at their respective levels, with the exception of parent education. Similar to the pooled models described previously, the effect of parental education weakens to p < .05 in Model 5. Overall, it appears that the marginal effects are very similar in terms of both size and direction when comparing the estimates for the pooled models in Table 2 with the estimates for traditionally aged graduates in Table 3.
In Table 4, the logistic regressions for nontraditional students are presented. Once again, Model 1 includes only estimates for parental education and parental income. Interestingly, while the effects of both SES measures were strong and highly statistically significant in this model for the traditional group, parental education is not statistically significant for the nontraditional group, and the effect of parental income is modestly statistically significant (p < .05).
Binary Logit Models for Nontraditional Age Predicting the Likelihood of Completing a Four-year Bachelor’s Degree on Time (marginal effects shown), 2008–2009 Baccalaureate and Beyond (B&B) Longitudinal Survey.
Note: Multiple parameter Wald tests are reported adjacent to variable names for sets of dummy regressors. Reported values are based on weighted estimates. As per National Center for Education Statisticrequirements, reported subsample sizes have been rounded.
p < .10. *p < .05. **p < .01.
In Model 2, measures for sex, race, disability status, marital status, dependent children, and age are added. Of these variables, only the estimate for the dependent children is statistically significant (p < .05). Among nontraditional graduates, having dependent children reduced the predicted probability of completing a degree on time by .095 compared to graduates without dependent children. When the sociodemographic variables are added to the model, the effect of parental income is no longer statistically significant. In general, the magnitude of the marginal effects for the nontraditional group is much smaller than those of the traditional group.
In Model 3, we include the variables that capture student expectations and academic performance. Only student expectations has a statistically significant effect on whether the respondents complete their degrees within four years, but unlike the traditional group, those who expect to complete additional schooling are less likely to complete on time (p < .05). The weak and nonsignificant impact of GPA is interesting as the effect of this variable was much stronger and highly statistically significant for the traditional group. The effect of the dependent child variable remained statistically significant (p < .05) when these variables are added to the model, while the other variables remained nonsignificant.
In Model 4, among the educational program characteristics added, only the effects of school selectivity (p < .10), field of study (p < .10), whether the respondents had changed their major (p < .05), and whether the respondents had stopped out (p < .05) are statistically significant. In addition, the dependent child variable remained statistically significant (p < .05), and the expectation variable increased slightly to p < .01.
Finally, none of the institution variables added in Model 5 are statistically significant. The most noteworthy changes to the existing variables in the model are the significance level for the dependent child variable, which changed from p < .05 in Model 4 to p < .01 in Model 5, and the significance level of the field of study variable, which increased from p < .10 to p < .05. Lastly, the level of statistical significance for the effect of the school selectivity remained the same (p < .10).
Overall, the results from Model 5 suggest that many of the factors that predict traditional age students’ likelihood of completing their degrees on time are not statistically significant for nontraditional students. Most notably, for nontraditional students, social class may be relatively less important in determining the likelihood of completing a degree on time. In Figure 5, for parental income, the same linear pattern noticed for the pooled models previously (in Figure 2) can be seen for the traditional students but not the nontraditional students. In terms of selectivity, Figure 6 indicates that two sets of relationships exist. For traditional students, the story appears to be quite linear—the more selective the school, the higher the probability of completing on time. However, for nontraditional students, the situation is a bit different. Nontraditional students in minimally selective schools show the highest probability of completing on time. It is possible that these minimally selective or “broad access” schools provide the greatest school accommodations and flexibility to support the life course balancing act facing nontraditional students and encourage their students to complete on time (Davies and Mehta 2018; Settersten 2015; Stevens 2015).

Predicted probabilities from binary logits predicting bachelor’s degree completion on time by parent income, traditional and nontraditional.

Predicted probabilities from binary logits predicting bachelor’s degree completion on time by selectivity, traditional and nontraditional.
Discussion and Conclusions
This paper examines the extent to which social class background influences undergraduates’ ability to complete their four-year bachelor’s degrees in four years. Indeed, despite widened access to postsecondary education in the United States, our results reveal significant inequalities even among the B&B sample of a relatively privileged group of individuals who have already successfully transitioned from secondary schooling to postsecondary education. That is, even those disadvantaged groups who have “made it,” so to speak, experience further differentiation in their completion times. Moreover, without examining completion time, system-wide levels of inequality will be underestimates as the additional inequalities revealed here compound with other established difficulties disadvantaged groups experience through their attrition from four-year colleges and relegation to less selective sectors of the higher education system.
Our results offer three key contributions to the existing literature. First, in terms of social class, graduates from low-SES backgrounds were significantly more likely to require extra time to complete their bachelor’s degrees. Moreover, our results do not support the previous literature that has attributed disparities in degree completion to employment characteristics between less affluent and affluent students (Bound et al. 2010; Engle and Tinto 2008; Kouliavtsev and Austin 2013; Mendoza 2012); instead, our results suggest that these differences cannot be explained away by low-income students spending more time working concurrently with their undergraduate studies.
At the same time, we also did not find support for the suggestion that middle- and upper-class students are strategically utilizing their times to degree as a means of remaining competitive for professional school and graduate school entry (Volkwein and Lorang 1996). Rather, our findings align more closely with Bourdieusian approaches of social reproduction and parental strategies, processes of horizontal inequality, and frameworks on maintaining inequality. The situation appears to be one that is consistent with the other postsecondary disadvantages of low-SES students as SES effects materialize both directly and indirectly through educational choices. Taken together, low-SES students are less likely to complete their degrees on time. Yet SES in the form of parental income showed strong direct effects, whereas parental education showed a combination of direct and indirect effects operating through fields of study, college selectivity, and academic performance. The stronger parental financial influence resembles that of the college completion literature and might suggest that once youth are “out on their own,” there is a greater dependence on financial support to not only complete their degrees but also to finish their degrees on time (Doren and Grodsky 2016).
Second, we observe very strong and robust field of study and college selectivity effects on completing a bachelor’s degree on time. It is important to note that these differences hold even when controlling for SES as well as academic performance. These findings underscore the importance of continued research on access to various fields as well as types of institutions. Previous work has shown these points of differentiation to be significantly related to social class background characteristics and that they represent critical junctures across the postsecondary landscape where inequality is maintained in expansive higher education systems (Ayalon and Yogev 2005; Davies et al. 2014; Davies and Guppy 1997; Goyette and Mullen 2006; Zarifa, 2012a, 2012b). Our findings show these points of selection are additionally important in determining timely degree completion.
Finally, our findings also speak to recent work on new populations in higher education (Davies and Mehta 2018; Deil-Amen 2015; Stevens and Kirst 2015; Settersten 2015). Our comparisons across traditional and nontraditional age groups suggest that the nontraditional students are far less likely to complete their programs on time. Moreover, the relationships between completion and the individual and institutional factors described previously may largely apply to traditional students who make a fairly linear transition from secondary to postsecondary education with only minimal delays and not to nontraditional students. For traditional students, social class indeed exhibited strong and robust effects across all models, but for nontraditional students, many of these relationships either did not emerge or looked characteristically different. Indeed, as we might expect, having dependent children, changing one’s major, stopping out, and the selectivity of the institution were consistent factors influencing nontraditional students’ completion times. Yet it is important to note that both in terms of parental income and selectivity, the story for nontraditional students was much more complex and nonlinear, much like their educational careers (Settersten 2015; Stevens 2015). Moreover, even within the four-year degree-granting sector examined here, our findings suggest that less selective, “broad access” institutions may be finding ways to accommodate the needs of nontraditional students and encourage timely completion (Davies and Mehta 2018; Jenkins and Rodriguez 2013).
While studies are increasingly concerned with improving student retention, reducing dropouts and attrition, and increasing degree completion, comparatively fewer studies have examined the difficulties some students face in completing their bachelor’s degrees in four years, and fewer still have examined the issue by focusing on the barriers to timely completion faced by low-SES students and new populations in higher education. The inequalities in completing one’s degree on time that we uncovered underscore the need to fold time to completion into a growing envelope of research that seeks to understand persisting inequalities in college attainment in expanded postsecondary systems. Our results do in fact mirror the well-documented difficulties faced by low-SES individuals in reaching parity in accessing postsecondary education, selective institutions, particular majors, and graduate and professional schools. But as higher education systems continue to evolve and seek ways to accommodate new populations of nontraditional students, policymakers and education officials interested in reducing completion times need to be aware that one-size-fits-all policies may largely be ineffective for these nontraditional students.
In this study, we brought SES background to the forefront of timely completion scholarship. It is also important for future research to explore the mechanisms through which gender and racial inequalities influence timely completion. Scholars have documented the difficulties males experience in terms of their educational attainment, dropout rates, and degree completion (Buchmann and DiPrete 2006; Buchmann et al. 2008; DiPrete and Buchmann 2013; Tumen et al. 2008), but the inequalities identified in this study suggest that timely completion of programs may be another important complication for males. Females in the B&B were significantly more likely than their male counterparts to finish their degrees on time. While this effect did weaken slightly when accounting for educational factors, it remained statistically significant even when additionally controlling for institutional differences. In terms of race, our findings showed significantly lower rates of completion among Hispanic graduates (Byun et al. 2012; Desjardins et al. 2002; Freeman 2004). Unlike Adelman’s (2006) work with the NELS:88/2000 and Volkwein and Lorang’s (1996) study at a PhD-granting, research-intensive college, the Hispanic effect (albeit weakening across models) persisted even when accounting for program characteristics, work commitments, academic performance, and institutional characteristics.
Future work would also benefit from shifting the focus to understanding the impact of timely degree completion on students’ postgraduate experiences. First, we know very little about both the immediate employment experiences and the success of pursuing additional educational programs of those individuals who do not complete their bachelor’s degrees on time. That is, how much of a difference does one or two or three extra years make in terms of one’s likelihood and the timing of securing full-time, stable employment? Alternatively, how much does additional time in one’s undergraduate degree reduce the likelihood of pursuing or delaying entry into first professional or graduate degrees? Certainly, Elman and O’Rand (2004) suggest that earlier and timely completion may snowball into future education and workforce advantages, but further exploration is needed. Second, we know equally as little about the interaction of such experiences with other immediate or longer-term life course outcomes. That is, how might additional years at the undergraduate level alter/impact other life course transitions? As nontraditional students increasingly balance multiple roles, it would be beneficial to assess empirically the extent to which alterations in the educational trajectory impact other life course outcomes. Finally, our understanding of the extent to which on-time completion may interact with school or degree characteristics (e.g., selectivity or field of study) in its influence on workforce or continuing education experiences is equally limited. Further exploration of the nature and extent of disadvantage faced by these individuals as they transition from school to work or continue their education is highly warranted.
Footnotes
Appendix
Descriptive Statistics by Degree Time for Variables from the 2008–09 Cohort of the Baccalaureate and Beyond Longitudinal Survey of University Graduates in the United States.
| More Than Four Years | On Time (Four Years) | ||
|---|---|---|---|
| Proportion/Mean | Proportion/Mean | ||
| Sex | *** | ||
| Female | 53 | 47 | |
| Male | 61 | 39 | |
| Race/ethnicity | *** | ||
| White | 53 | 47 | |
| Black | 68 | 32 | |
| Nonwhite Hispanic | 69 | 31 | |
| Asian | 51 | 49 | |
| Other or mixed | 63 | 37 | |
| Marital Status | *** | ||
| Single, Never Married | 50 | 50 | |
| Married | 89 | 11 | |
| Disability | ** | ||
| No | 56 | 44 | |
| Yes | 62 | 38 | |
| Dependents | *** | ||
| No | 51 | 49 | |
| Yes | 94 | 6 | |
| Age | *** | ||
| Traditional (15–20) | 54 | 46 | |
| Nontraditional (21+) | 84 | 16 | |
| Parent education | *** | ||
| High school or less | 72 | 28 | |
| Some college | 64 | 36 | |
| Baccalaureate or higher | 47 | 53 | |
| Parent income | *** | ||
| Low income | 82 | 18 | |
| Low mid-income | 69 | 31 | |
| Upper mid-income | 53 | 47 | |
| High income | 37 | 63 | |
| Student expects more than BA | *** | ||
| No | 64 | 36 | |
| Yes | 54 | 46 | |
| Standardized GPA | 316.82 | *** | 336.41 |
| Field of study | *** | ||
| Business | 61 | 39 | |
| Engineer/computer science | 64 | 36 | |
| Math/science/health | 53 | 47 | |
| Social science | 44 | 56 | |
| Humanities | 47 | 53 | |
| Education | 65 | 35 | |
| Other field | 63 | 37 | |
| Institution selectivity | *** | ||
| Nonselective | 80 | 20 | |
| Minimally selective | 69 | 31 | |
| Moderately selective | 61 | 39 | |
| Very selective | 37 | 63 | |
| Ever enrolled in co-op/internship | *** | ||
| No | 58 | 42 | |
| Yes | 51 | 49 | |
| Hours worked during school | 24.91 | *** | 15.44 |
| Changed major | |||
| No | 55 | 45 | |
| Yes | 58 | 42 | |
| Stopped out | *** | ||
| No | 49 | 51 | |
| Yes | 85 | 15 | |
| Ratio of institutional aid to total aid in 2007–2008 | 8.89 | *** | 26.21 |
| Institution type | *** | ||
| Public | 62 | 38 | |
| Private for-profit | 83 | 17 | |
| Private | 39 | 61 | |
| Religious | 44 | 56 | |
| Enrollment size | 18,352 | *** | 16,274 |
| n | 7,020 | 5,490 | |
Note: Estimates are survey weighted using bootstrap replicate weights. Design-based F-tests are reported. As per National Center for Education Statistics requirements, reported subsample sizes have been rounded.
p < .01. ***p < .001.
Acknowledgements
This research was supported by funding from the Canada Research Chairs program and the Social Sciences and Humanities Research Council (Grant No. 430-2012-0376).
