First-Year,Nontraditional Student Retention at Four-Year Institutions: How Predictors of Attrition Vary Across Time

Abstract

This study examined the longitudinal persistence behavior of first-time, nontraditionally aged students at four-year institutions by leveraging the 2004 to 2009 Beginning Postsecondary Student data set. A human capital theory framework was employed to assess how student background characteristics, enrollment characteristics, and financial factors influence first-time, nontraditional student retention across multiple years of enrollment. The findings indicated that the highest risk of departure occurred during the first year of enrollment for nontraditional students who were first generation, enrolled full time, and from a low-income background. During the second year, students from a middle-income background were most at risk. Female students were most likely to depart in the third year, and minority students had the greatest risk of departure during the fourth year.

Keywords

nontraditional students first-time students college retention persistence event history modeling

Nontraditional (NT) students, also referred to as adult students or adult learners, represent a substantial portion of postsecondary enrollments nationwide. While the definition of who an NT student is varies from study to study, NT students are often described as undergraduates older than 24 years of age (Shapiro et al., 2017), which would qualify them as financially independent per federal student aid guidelines (Federal Student Aid, 2018). This particular pool of students is often overrepresented within community colleges (e.g., Juszkiewicz, 2017; Ma & Baum, 2016). As Blumenstyk (2018) notes, “While adults make up 27 percent of all undergraduates nationally, only one-third of private-four-year colleges enroll adults at that rate or higher. Among public four-year colleges, less than one-third do” (p. 27). While many NT students who enroll in four-year institutions may have attended previously or have transferred from another institution, some NT students choose to begin their postsecondary studies at a four-year college (e.g., Tennessee Higher Education Commission, 2017). The prominent presence of NT students has not gone unnoticed and has been the center of attention on recent policy initiatives. Efforts have emerged to promote and to support NT students as they enroll (or reenroll) in postsecondary education at the national level (e.g., Lumina Foundation, 2017), the state level (e.g., Drive to 55 Alliance, 2018; Ohio Department of Higher Education, 2015; Washington Student Achievement Council, 2014), and the institution level (e.g., The University of Memphis, 2019).

Because studies of first-year students at four-year institutions often focus on the traditional 18- to 24-year-old students, the challenges and enrollment patterns of first-time NT students remain unclear. NT students, not delimited to only first-time enrollees, are commonly associated with three distinct types of barriers to their progression and success within higher education: situational (those relating to the student’s life circumstances), institutional (those relating to current policies and practices in higher education), and dispositional (student attitudes toward and perceptions of education; Osam, Bergman, & Cumberland, 2017). Compared with NT students attending a community college, those attending a four-year college are exposed to an environment designed for and more intent on younger students who enroll directly after graduating from high school (i.e., institutional barriers; Gilardi & Guglielmetti, 2011). NT undergraduates, in contrast, delayed their initial enrollment or returned to the classroom after having previously dropped or stopped out (which may inform dispositional barriers; Choy, 2002). In addition, NT students are often cited for balancing multiple social roles (e.g., working part- or full-time off campus and supporting children or dependents) concurrent with their enrollment (i.e., situational barriers; Choy, 2002).

Retention studies aimed at a four-year college environment are often more applicable to the prominent “traditional” student body (e.g., Tinto, 1987), are delimited to a single-institution setting (e.g., Markle, 2015), or are cross-sectional (e.g., Roberts & Styron, 2010). Student persistence is a product of longitudinal behavior changes (Ishitani & DesJardins, 2002), yet many existing studies (e.g., Metzner & Bean, 1987; Roberts & Styron, 2010) examined persistence using a student’s aggregated enrollment status. Such methods fail to illustrate how influential factors may vary across time. Furthermore, despite general commonalities (e.g., age, enrollment, and financial independence), NT students should not be considered a homogenous group (Bean & Metzner, 1985; Offerman, 2011). The motivations and circumstances of NT students enrolling for the first time at a four-year institution likely vary from those who have previously enrolled (e.g., Laanan, Starobin, & Eggleston, 2010; Taylor & House, 2010). The purpose of this study, therefore, is to investigate the longitudinal persistence behavior of first-time, NT students within the four-year setting. This will be accomplished through the leveraging of a national data set sponsored by the National Center for Education Statistics (NCES). Through this approach, this study will add to the scholarly literature on NT student retention, will seek to establish a benchmark for future retention research, and will be informative to both policymakers and higher education practitioners as efforts aimed to support NT students continue to emerge.

Theoretical Framework

Retention studies aimed at four-year colleges and at first-time students are commonly rooted in Tinto’s (1987) integration theory, which takes into account how students academically and socially connect with their institution of enrollment. Given that Tinto’s integration theory assumes students will spend large proportions of time on campus participating in organizations and attending events, this commonly held theory does not readily apply to NT students who commute to campus primarily to attend class (Davidson & Wilson, 2013; Tinto, 1982). As such, this study employs a human capital theory framework, which conceptualizes the decision to enroll as an investment and the potential for higher lifetime income as a return on investment (Becker, 1993). Viewed through a human capital theory lens, students ultimately aim to earn a return (i.e., future income, status, and other benefits) for their investment into higher education (Becker, 1993; J. Chen & Hossler, 2017; Rowan-Kenyon, 2007).

Viewing student enrollment and decisions to depart through a human capital theory lens is arguably more appropriate for NT students than for their younger peers, as they would have less time upon completion to seek a return on their investment before reaching retirement age (Brändle, 2017). In addition, NT students are often noted for entering postsecondary education with numerous preexisting obligations and obstacles. Examples of such include being a single parent or supporting dependents, working off-campus part time or full time, being financially independent, and being academically unprepared for college-level coursework (Laing & Robinson, 2003; Wells & Lynch, 2012). With this information in mind, one could argue that an NT student’s decision to enter and to depart postsecondary education is related to his or her perceived return on investment. Considering the higher costs associated with attending a four-year college than a 2-year institution, any potential return would require a greater investment of human capital resources, especially for NT students. Many NT students may be hesitant to make such demanding investment (especially considering other financial obligations they may be managing). Other NT students, however, may commit to the greater investment in hope of a greater return.

Literature Review

Human Capital

This study conceptualizes human capital from the perspective of income, financial aid, and net tuition: the student’s reported family income, the amount the student receives from the Pell grant, the amount the student accepts from student loans, and the amount the student owes to the institution after subtracting all applicable financial aid. This approach differs from other studies leveraging a human capital theory. For example, Perna’s (2005) study on faculty tenure and promotion conceptualized human capital as educational attainment. Rowan-Kenyon (2007) considered human capital characteristics in her study on delayed college enrollment and defined human capital by means of academic factors. Considering human capital through a lens of academic resources (e.g., Perna, 2005; Rowan-Kenyon, 2007) or through financial resources (as in this study), the essence of the theory remains unchanged: investing current capital for the potential of higher returns in the future. As prefaced in the introduction, the financial perspective of human capital theory may be more appropriate for NT students due to the possibility of other financial burdens and their closer proximity to retirement age.

Given the financially independent status of NT students, financial aid is an important factor to consider. In a brief literature review of financial aid for adult learners, Dougherty and Woodland (2009) noted that students between the ages of 24 and 29 represented the greatest proportion of federal aid grant awardees and student loan recipients, with the average grant award amount being smaller and the average loan amount being larger for older students. In a study of financial aid’s effect on student persistence, Bettinger (2004) noted that older students and male students were less likely to persist in relation to their financial aid awards. In another study, Paulsen and St. John (2002) noted that older students from low- and middle-income groups were less likely than their younger peers to persist once variables relating to financial aid and tuition were considered. Likewise, R. Chen and DesJardins (2007), having considered both traditional and NT students in their study, noted that Pell grant awards helped to mitigate attrition of low- and middle-income NT students.

Student Background Characteristics

Prior studies on NT student retention take into consideration select student background characteristics, such as gender, race, first-generation status, and socioeconomic status. Regarding gender, Bergman, Gross, Berry, and Shuck’s (2014) study on adult student persistence at a public university did not find persistence related to differences between males and females. Markle (2015), however, found substantial differences in the persistence behavior of men and women. The results of her study suggested that adult female students, more so than males, demonstrated a will to persist despite the presence of risk factors and multiple priority conflicts (i.e., employment and caring for dependents). Osam et al. (2017) echoed similar sentiments: Noting that gender roles (i.e., those relating to the domestic environment) could present a greater risk factor for females. Especially for those with young children, the added stress of interrole conflict may lead to female student attrition (Spellman, 2007).

Regarding racial background, Metzner and Bean’s (1987) study—one of the most well-known studies on NT student retention—found that race is indirectly related to attrition. The descriptive statistics offered by Shapiro et al. (2017) seem to validate the notion that race alone is not a significant risk factor, as the credential completion gap between racial groups is narrower and less prominent with NT students than it is with younger, traditional-aged college students. Research does support, however, that student retention prediction factors and risks of attrition vary between racial groups (Baker & Robnett, 2012; R. Chen, 2012).

Whether or not an NT student is the first in his or her family to attend college (i.e., a first-generation student) is also a factor worth considering (Bean & Metzner, 1985). First-generation status may represent a variety of challenges to student persistence (Cabrera, Miner, & Milem, 2013). Research and data support that NT students are more likely than their traditional-aged peers to be first-generation students (Ma & Baum, 2016; Rabourn, BrckaLorenz, & Shoup, 2018). Surprisingly, as compared with traditional-aged student groups, scant research is available on the retention of first-generation, NT students. From a traditional-aged student perspective, X. Chen (2005) reported that 43% of first-generation students left their institutions without degrees, whereas 20% of students whose parents were college-educated did so. More recently, Ishitani (2016) discovered that first-generation students were most likely to drop out of college during the second year.

Student Enrollment Characteristics

In addition to accounting for variables relating to a student’s background and finances, this study also takes into consideration characteristics pertaining to student enrollment: college grade point average (GPA), full- versus part-time enrollment, and institutional control. While GPA and enrollment intensity are worthy of a more thorough discussion, institutional control is included in this study to account for differences between public and private universities. Public and private institutions, of course, vary in a number of regards, one of which being their institutional outcomes (Caruth, 2018).

Regarding GPA, Metzner and Bean (1987) found GPA to be the most significant and powerful factor in determining an NT student’s success. While the literature does often associate NT students with being academically underprepared, this is typically in the context of developmental or remedial education at a community college with open-access standards (Senie, 2016). Within a four-year institution with more selective admission criteria, NT students might be more academically prepared for college-level coursework while also being expected to demonstrate higher academic performance. Even though NT students may feel stressed and misplaced in a four-year setting, research supports that NT students exhibit grit and motivation during their enrollment (Chung, Turnbull, & Chur-Hansen, 2017; Markle, 2015).

Part-time enrollment, a characteristic often associated with NT students, has also been observed to act as a risk factor for this student group. NT students are more likely than their traditional-age peers to enroll on a part-time basis (Rabourn et al., 2018). A number of studies have pinpointed the risk of attrition associated with part-time enrollment. For example, Metzner and Bean (1987) found that the number of enrolled hours acts as a significant predictor of NT student retention. Taniguchi and Kaufman (2005) concurred that part-time enrollment increased the risk of departure when accounting for the duration of enrollment (seeing that part-time students take longer to complete degree requirements than full-time students), financial aid factors, and the number of previous enrollments. Seemingly contrary to other studies, Markle (2015) observed that part-time enrollment for NT female students was a significant predictor of retention but not for males.

Methods

Data

The study data were drawn from the 2004 to 2009 Beginning Postsecondary Student (BPS: 04/09) data set sponsored by the NCES. This data set included data collected longitudinally over a period of six years and consisted of first-time, beginning students who enrolled in four-year, 2-year, or less than 2-year institutions in 2004. The total size of the data was 16,648. Of 16,648 students, 7,568 students with intact information attended either public or private four-year institutions for the first time. Of those who matriculated to four-year institutions, 248 students were 24 years or older at the time of matriculation as first-time freshmen. According to Choy (2002), these 248 students were classified as NT due to being older than 24 years and delaying enrollment.

Dropout students were operationally defined in this study as students who left their initial four-year institutions and never returned to their original institutions or never enrolled in other institutions before the end of the survey observation period. Students with other departure types were censored at the end of each year. One unique feature in event history modeling is its capability to censor certain types of events. The event of interest in this study was voluntary withdrawal from college (i.e., dropout). Of course, students might have experienced other types of departures such as transferring out, graduating, or stopping out. Event history modeling can focus on the event of interest (in this case, NT students dropping out) while removing students with other departure types from the analysis (Yamaguchi, 1991). Another censoring type includes students who did not experience any type of departure before the survey observation ended (i.e., the end of the fourth year). This is referred to as right-censored observations (Tabachnick & Fidell, 2007). A total of 114 of the 248 students left their initial four-year institutions. Table 1 exhibits a breakdown of 114 students who dropped out by academic year. About 42% of the 114 students dropped out in the second year, which was the highest percentage of dropouts in the study data, followed by 28% in the third year.

Table 1.
Dropouts by Year.

Year Dropout

Count Percentage

1 19 16.7

2 48 42.1

3 32 28.1

4 years and after 15 13.2

Total 114 100.0

Table 2 presents the descriptive summary for explanatory variables included in this study. Over 61% of the sample data were females. About 67% of the students were White, followed by 22% who were Black. Determining who qualifies as a first-generation student is a matter of debate. A recent synthesis of retention studies noted the use of 12 distinct operational definitions of first-generation status (Peralta & Klonowski, 2017). The Federal TRIO program defines first-generation students as those whose parents did not complete a bachelor’s degree. By this definition, about 67% of the study samples were classified as first-generation students. Family income was based on students’ parental income level, and the effect of family income was assessed by using quartiles which were already included in BPS: 04/09.

Table 2.
Descriptive Summary of Study Sample.

Variable Label Count Percentage Mean Standard Deviation

Gender Female 152 61.3

Male^a 96 38.7

Race Asian 10 4.0

Black 55 22.2

Hispanic 16 6.5

White^a 167 67.3

Parent’s education First generation 189 76.2

One or both parents with college degrees^a 59 23.8

Family income Low (under $30,000) 50 20.2

Low middle ($30,000–$59,999) 79 31.8

High middle ($60,000–$89,999) 72 29.0

High ($90,000 and above)^a 47 19.0

College GPA 2003–2004 3.18 0.7

2006–2007 3.01 1.0

Pell amount^b 2003–2004 $2,454 $1,267

2004–2005 $2,859 $1,132

2005–2006 $2,880 $1,231

2006–2007 $3,137 $1,150

Loan amount^b 2003–2004 $5,096 $2,110

2004–2005 $5,532 $2,299

2005–2006 $6,415 $2,311

2006–2007 $7,087 $2,832

Net price 2003–2004 $1,938 $3,289

Intensity: full time 2003–2004^a 117 47.2

2004–2005^a 151 60.9

2006^a 129 52.0

Intensity: part time or mixed 2003–2004 131 52.8

2004–2005 97 39.1

2006 119 48.0

Control Public^a 154 62.1

Private 94 37.9

Note. GPA = grade point average.

^aReference group in event history analysis.

^bStudents with $0 amounts were excluded from mathematical computations.

College GPA was measured twice, in 2004 and 2006. The GPA in 2004 was 3.18 with a standard deviation of 0.7, whereas it was 3.01 with a standard deviation of 1.0 in 2006. The GPA in 2006 was calculated based on students who were still enrolled in their original institutions in 2006.

The applicability of human capital theory to explain NT students’ persistence behavior was tested by the net price of attending college. While many students are eligible for grants to attend college, total amounts of college costs are typically not compensated by the total amount of the grants. As a result, students either have to take loans or independently finance themselves to enroll in college. The net price variable in this study was the total amount of college cost minus the amount of all the grants the student received. This study assumed when students had a higher net price, they were more committed to graduating from college to attain monetary gains to pay off their loans.

The effect of financial aid on persistence was also assessed in the form of average yearly amounts of Pell grants and student loans. The average Pell grant amount ranged from $2,454 to $3,137. The average loan amount varied from $5,096 to $7,087. Loan amounts increased as students advanced in their college careers. This was partially due to upper division courses costing more than lower division class courses (Li, 2010). Finally, approximately 62% of NT students matriculated to public four-year institutions, whereas 38% enrolled in private four-year institutions.

The BPS: 04/09 tracked students’ enrollment intensity status for the first three years. In 2003 to 2004, 47.2% of students were enrolled as full time, and 52.8% enrolled either part time or a combination of full time and part time. About 61% of students enrolled in full time, while 39% of students enrolled less than full time in 2004 to 2005. In the third year, 52% of students were full time, while 48% of students attended full time and part time.

Analysis

This study applied event history modeling, also known as survival analysis, to explore longitudinal student persistence behavior. Event history modeling offers advantages over other statistical approaches by being able to consider the factors related to the outcome (i.e., whether or not the student dropped out) along with when the event (i.e., dropout) occurred (Singer & Willett, 1993). Past retention studies have used aggregated enrollment status over academic years, such as enrollment status at the end of the second or fourth year. In such cases, students who withdrew at different years were aggregated; however, it is reasonable to assume students may have distinct reasons for their departure depending on an academic year. For instance, students may leave their institutions due to a lack of fit to the institution during the first year, whereas students may withdraw because of anticipating cumulative loan-debt amounts in the second year. Using event history modeling, this study was able to unpack the effects of explanatory variables on NT student departure specific to each academic year during the four-year observation period (Yamaguchi, 1991).

A number of modeling techniques are found when one applies event history modeling. The most widely used technique has been Cox regression, also known as semiparametric modeling, which estimates a set of coefficients after taking the effect of time dimension into account. This technique, however, does not offer period-specific effects of certain independent variables. For example, being female is perpetual and time constant, though the effect of being female may vary over time. Because Cox regression typically fails to reveal such detailed effects particular to certain time periods, exponential modeling with period-specific effects was applied in this study to unpack period-specific effects of our explanatory variables unique to certain years (Blossfeld & Rohwer, 1995).

Prior to applying event history modeling, multicollinearity and inflectional outliers were examined. Variance inflation factors (VIFs) were tested to detect any issues related to multicollinearity (i.e., highly correlated independent variables). While various rules of thumb exist for interpreting VIF output, Ethington, Thomas, and Pike (2002) noted that VIF values greater than 10 are evidence of high multicollinearity. VIF values for this study ranged from 1.17 to 4.13, which did not present any alarming issue. Cook’s D and Mahalanobis tests were estimated to identify any potential outliers in continuous explanatory variables. The results of these tests did not reveal any existing outliers. Finally, the Taylor series variance estimation procedure was applied to mitigate issues associated with complex survey data.

Results

Because the event history model permitted the researchers to study the longitudinal persistence behavior of NT students, the results discussed in the following paragraphs are separated by each year of the observation period. Within the discussion of each year, the results pertaining to human capital characteristics are noted distinctly from other results. Table 3 presents the results of the event history analysis. Regression coefficients and standard errors, odds ratios (OR), Δ (delta), as well as significance levels were included in the table. An OR greater than 1 indicates the magnitude of positive effect on student’s departure, while an OR less than 1 is associated with lowering student’s dropout rate. Δ indicates the change in the probability of departure. For example, Δ for female students in the first year was −0.654. This is interpreted as NT female students were about 65% less likely (due to the negative sign) than male students to withdraw on average in the first year of enrollment.

Table 3.
Results of Event History Model.

Year 1
Year 2

Variable Label Coefficient Standard error Odds Δ Significance Coefficient Standard error Odds Δ Significance

Intercept −5.866 0.735 * −0.684 0.662

Gender Female −1.062 0.215 0.346 −0.654 * −0.318 0.170 0.728 −0.272 *

Race Asian −2.459 0.265 0.086 −0.914 * −1.476 0.306 0.228 −0.772 *

Black 1.497 0.221 4.468 3.468 * −0.073 0.184 0.929

Hispanic 0.962 0.207 2.617 1.617 * 0.147 0.102 1.158

Parent’s education First generation 2.559 0.774 12.927 11.927 * 0.259 0.423 1.296

Family income Low 1.038 0.215 2.824 1.824 * 0.238 0.217 1.269

Low middle −0.894 0.321 0.409 −0.591 *** 0.469 0.253 1.598 0.598 *

High middle 0.211 0.318 1.235 0.668 0.243 1.950 0.950

College GPA Continuous 0.594 0.110 1.811 0.811 * −0.193 0.041 0.824 −0.176 *

Pell amount Continuous (per $1,000) −1.004 0.173 0.366 −0.634 * −0.226 0.036 0.798 −0.202 *

Loan amount Continuous (per $1,000) −0.322 0.076 0.725 −0.275 * −0.099 0.026 0.905 −0.095 *

Net price Continuous (per $1,000) −0.443 0.058 0.642 −0.358 * 0.074 0.013 1.077 0.077 *

Intensity Full time 0.907 0.271 2.478 1.478 * −0.459 0.124 0.632 −0.368 *

Control Private −0.165 0.237 0.848 −0.249 0.200 0.779

Intercept −2.413 0.388 * −3.629 0.459 *

Gender Female 0.615 0.131 1.849 0.849 * 3.032 0.312 20.746 19.746 *

Race Asian −1.092 0.958 0.336 0.554 0.240 1.740 0.740

Black −0.045 0.235 0.956 0.892 0.314 2.440 1.440 *

Hispanic 0.087 0.278 1.091 1.254 0.319 3.504 2.504 *

Parent’s education First generation −0.012 0.183 0.988 0.219 0.254 1.245

Family income Low 1.738 0.289 5.685 4.685 * 0.319 0.404 1.375

Low middle 1.387 0.275 4.003 3.003 0.501 0.774 1.650

High middle 0.456 0.274 1.578 1.022 0.411 2.778 1.778

College GPA Continuous 0.028 0.099 1.029 −0.583 0.120 0.558 −0.442 *

Pell amount Continuous (per $1,000) −0.219 0.068 0.804 −0.196 * 0.271 0.076 1.311 0.311 *

Loan amount Continuous (per $1,000) −0.025 0.015 0.975 −0.233 0.034 0.792 −0.208 *

Net price Continuous (per $1,000) −0.071 0.020 0.931 −0.069 * −0.023 0.023 0.978

Intensity Full time −0.416 0.162 0.660 −0.340 −0.744 0.299 0.475 −0.525

Control Private −0.062 0.191 0.940 −0.594 0.542 0.552

Note. GPA = grade point average.

p < .10. p < .05. p < .01.

First Year

Beginning with the human capital characteristics in the model, NT students from low-income families were 2.8 times more likely to leave their four-year institutions than their counterparts from high-income families. Conversely, students from low-middle income families were 59% less likely to depart when compared with students from high-income families.

While a number of factors in the first year were related to the likelihood of dropping out, some factors contributed to reducing the odds of departure. For example, the net price variable had a negative effect on student attrition. Every $1,000 increase in student’s net price reduced the likelihood of dropout by 36% in the first year. Similarly, for every $1,000 increase in Pell and loan amounts, dropout rates lowered by 63% and 28%, respectively.

Regarding student background and enrollment characteristics, the most notable finding in the first year was that first-generation NT students were at their highest risk of departure. On average, the chance of first-generation students to drop out was about 13 times (OR) higher than their counterparts. Additional factors associated with increasing the likelihood of departure in the first year were found. In comparison with White students, Black and Hispanic NT students were about 4.5 and 2.6 times more likely to drop out in the first year, respectively. Students who enrolled full time were also about 2.5 times more likely to drop out, whereas every one-point increase in GPA increased the odds of dropout by 81%. On the contrary, female and Asian students were 65% and 91% less likely to depart in the first year.

Second Year

In the second year, a positive impact of high incomes on the odds of dropping out surfaced. In comparison to students from the high-income group, students from low-middle and high-middle income groups were approximately 60% and 95% more likely to leave their four-year institutions during the second year. Interestingly, every $1,000 increase in net price increased the chance of dropout by 8%. In terms of financial aid, Pell grant and loan amounts remained effective in lowering the risk of student departure. Every $1,000 increase in Pell grant and loan amounts reduced the odds of departure by 20% and 10%, respectively.

If the student persisted to the second year of enrollment, the positive effective of first-generation status on the odds of dropping out subsided. Female and Asian students were still 27% and 77% less likely than their counterparts to leave in the second year. Unlike in the first year, a positive effect of college GPA on persistence appeared in the second year. Every one-point increase in college GPA was associated with reducing the chance of departure by about 18%. Also in contrast to the first year of enrollment, NT students enrolled full time during the second year were 37% less likely than those enrolled part time to drop out.

Third Year

The results pertaining to human capital characteristics were the most notable during the third year of the observation period. The most prominent finding in the third year was the substantial and positive effect of family income on NT student departure. The odds of students from low and low-middle income families to drop out were 5.7 and 4 times higher than students from high-income families. Every $1,000 increase in net price lowered the likelihood of dropout by about 7% in the third year of college. A positive effect of Pell grants on retaining NT students was still statistically significant in the third year. Every $1,000 increase in a Pell amount reduced the odds of departure by 20%.

As for other notable results, gender emerged as a significant predictor in the third year. Female students were 85% more likely than male students to leave during the third year. In addition, being enrolled full time was found to reduce the likelihood of departure by 34%, relative to part-time enrolled students.

Fourth Year

Regarding human capital characteristics in the fourth year, higher loan amounts were associated with lowering NT student attrition. A $1,000 increase in a student’s loan amount reduced the likelihood of departure by 21%. Students from the high-middle income group were at the highest risk of dropout during the final year of the observation period. The likelihood of these students dropping out was approximately 2.8 times higher than the odds of students from the highest income group.

Concerning student and enrollment characteristics, strong and negative effects pertaining to race on departure were observed. Relative to White students, the likelihoods of departure for Hispanic, Black, and Asian students were approximately 3.5, 2.4, and 1.7 times higher in the fourth year, respectively. Conversely, higher GPAs and enrollment intensities were associated with lowering NT student attrition. Every one-point increase in college GPA was related to reducing the likelihood of dropping out by about 44%. The effect of being full time on persistence was stronger than in the third year, and students enrolled full time were about 53% less likely than their part-time counterparts to depart.

Limitations

There are a number of limitations pertaining to this study. Enrollment intensity, Pell grant, and loan amounts were measured every year in the BPS: 04/09, whereas college GPA was collected twice (in 2004 and 2006). As a result, college GPAs in 2004 served as substitutions for GPAs in 2005. In a similar vein, college GPAs in 2006 functioned as proxies for the same variables in 2007 and thereafter. While it is reasonable to believe that GPAs may change over time, this study was not able to estimate true period-specific effects of college GPAs.

In the U.S. population, a larger group of NT students enrolled in four-year institutions are not first-time freshmen. The cohort in the BPS: 04/09 was, however, limited to first-time, beginning students, which resulted in a smaller study sample of NT students who initially enrolled in four-year institutions. Moreover, it was not clear how many years elapsed before the sampled NT students decided to pursue their college career. With this in mind, the sample analyzed in this study could be more indicative of a pool of younger NT students than of a wider range of ages.

Finally, readers need to interpret findings for the fourth year and beyond with caution. The number of students in the fourth year became smaller due to students who dropped out in previous years being removed from the fourth-year analysis. As a result, certain student characteristics (e.g., females) in the fourth year might have been overestimated. Thus, readers are advised to treat findings associated with the fourth year and years afterward as noteworthy references rather than robust findings.

Discussion

NT students are distinct from their younger peers. They enter postsecondary education at a later stage in life and balance responsibilities and multiple roles not common with traditionally aged students. Wyatt (2011) noted that “nontraditional students are always in transition” (p. 14). Postsecondary administrators must, therefore, understand the needs of these students throughout their enrollment. The findings of this study posit that institutions need to coordinate their retention efforts according to characteristics unique to NT students. Following a discussion on the relevance of human capital theory and other salient findings from the event history analysis, implications and recommendations for practice are specifically outlined.

Relevance of Human Capital Theory

This study conceptualized human capital through the lens of family income, financial aid, and net price. As such, the results presented a different perspective on the applicability of human capital theory than can be found in other studies (e.g., Perna, 2005; Rowan-Kenyon, 2007). Generally, the findings presented herein support human capital theory as a theoretical framework to explain NT students’ persistence behavior at four-year institutions. Throughout the four years of the observation period, the factors pertaining to human capital demonstrated significant influence on the likelihood of an NT student dropping out. Some of the most dramatic effects these factors exhibited occurred during the first year of enrollment. For example, the net price demonstrated the strongest effect on student persistence (reducing the likelihood of departure) during the first year. Net price retained its significant effect through the third year of enrollment, though it acted as a positive predictor of dropping out in the second year. Furthermore, this study underpins the importance of financial aid for NT students to persist in college. The amount of financial aid awarded (Pell Grant and student loan amounts) consistently exhibited a negative impact on the likelihood of dropping out (i.e., higher amounts were positively associated with retention). Similar to the net price variable, both Pell grant and loan amounts exhibited the strongest impact on the first-year persistence.

Relevance of Other Factors

Higher college GPA and full-time enrollment showed the negative effect of first-year persistence. Given multiple life responsibilities NT students carry, one plausible explanation for such negative effects was that these students might have spent more time on academic work during the first year while they fell behind on fulfilling family and employment responsibilities. Echoing back to the literature, family support is a critical factor pertaining to NT student retention (Bergman et al., 2014), Assuming this, NT students might have decided to withdraw from college to focus on familial and other responsibilities neglected during the first year. This may be more pressing for female students than for male students. Markle (2015) noted that women often cited issues relating to family obligations when considering to withdraw. Considering that female students were significantly more likely than males to dropout during the third year, the reasons may be attributable to prolonged familial obligations taking precedence over the continuation of their studies.

First-generation, traditional students were more likely to leave their four-year institution during the second year in college (Ishitani, 2016). In this study, NT, first-generation students were at the highest risk of dropping out in the first year. Negative effects of other variables such as race and low income were also prominent during the first year. This study posits that institutional efforts to retain NT students must be intensified during the first year. Institutional administrators may also consider offering summer programs prior to matriculation specifically targeting NT students. Coupled with an NT student’s multiple obligations, institutions must realize the varying needs and complex challenges these students face. Further investigation on first-year college experiences among NT students in the four-year setting is pressingly needed to support their college access and success.

Implications for Practice

At least two implications for practice can be drawn from this study. First, for those who have families to support, NT students are more concerned with financing their education. This notion manifested in the significant finding on financial aid, especially the effect of Pell grants in the first year. Across all years in the observation period, however, financial aid-related variables demonstrated significant influence in retaining NT students. As the total Pell grant amount depends on individual financial situations, administrators should consider subsidizing NT students with institutional grants. For full-time, NT students who may be underemployed or unemployed, financial aid workshops should be offered to inform them of other federal-based financial aid, such as work-study programs. The second implication for practice is that institutional administrators should consider the results of this study as evidence for timely, targeted intervention strategies. For example, this study found that NT students from lower income families were more likely to drop in the first year. Thus, financial aid workshops on student loans and stress management seminars specifically tailored for first-year NT students from lower income families should be offered during the orientation.

Conclusion

All institutions should consider best practices in serving and supporting success among NT students. The findings of this study posit that institutions should approach issues on NT students and their persistence behavior differently. Future research on this student group should investigate the motivations for first-time NT students to enroll. Such research may also assess the differences between NT students who enroll for the first time versus those who are reenrolling or transferring at a later age. To allow for more robust research on nationally representative data, NCES should consider surveys which target or oversample NT undergraduates. As both policy and practice become more sensitive to the NT undergraduate body, future research in this area will prove timely, informative, and transformational.

Year	Dropout
1	19	16.7
2	48	42.1
3	32	28.1
4 years and after	15	13.2
Total	114	100.0

Variable	Label	Count	Percentage	Mean	Standard Deviation
Gender	Female	152	61.3
Male^a	96	38.7
Race	Asian	10	4.0
Black	55	22.2
Hispanic	16	6.5
White^a	167	67.3
Parent’s education	First generation	189	76.2
One or both parents with college degrees^a	59	23.8
Family income	Low (under $30,000)	50	20.2
Low middle ($30,000–$59,999)	79	31.8
High middle ($60,000–$89,999)	72	29.0
High ($90,000 and above)^a	47	19.0
College GPA	2003–2004			3.18	0.7
2006–2007			3.01	1.0
Pell amount^b	2003–2004			$2,454	$1,267
2004–2005			$2,859	$1,132
2005–2006			$2,880	$1,231
2006–2007			$3,137	$1,150
Loan amount^b	2003–2004			$5,096	$2,110
2004–2005			$5,532	$2,299
2005–2006			$6,415	$2,311
2006–2007			$7,087	$2,832
Net price	2003–2004			$1,938	$3,289
Intensity: full time	2003–2004^a	117	47.2
2004–2005^a	151	60.9
2006^a	129	52.0
Intensity: part time or mixed	2003–2004	131	52.8
2004–2005	97	39.1
2006	119	48.0
Control	Public^a	154	62.1
Private	94	37.9

		Year 1	Year 2
Intercept		−5.866	0.735			***	−0.684	0.662
Gender	Female	−1.062	0.215	0.346	−0.654	***	−0.318	0.170	0.728	−0.272	*
Race	Asian	−2.459	0.265	0.086	−0.914	***	−1.476	0.306	0.228	−0.772	***
Black	1.497	0.221	4.468	3.468	***	−0.073	0.184	0.929
Hispanic	0.962	0.207	2.617	1.617	***	0.147	0.102	1.158
Parent’s education	First generation	2.559	0.774	12.927	11.927	***	0.259	0.423	1.296
Family income	Low	1.038	0.215	2.824	1.824	***	0.238	0.217	1.269
Low middle	−0.894	0.321	0.409	−0.591	***	0.469	0.253	1.598	0.598	*
High middle	0.211	0.318	1.235			0.668	0.243	1.950	0.950	**
College GPA	Continuous	0.594	0.110	1.811	0.811	***	−0.193	0.041	0.824	−0.176	***
Pell amount	Continuous (per $1,000)	−1.004	0.173	0.366	−0.634	***	−0.226	0.036	0.798	−0.202	***
Loan amount	Continuous (per $1,000)	−0.322	0.076	0.725	−0.275	***	−0.099	0.026	0.905	−0.095	***
Net price	Continuous (per $1,000)	−0.443	0.058	0.642	−0.358	***	0.074	0.013	1.077	0.077	***
Intensity	Full time	0.907	0.271	2.478	1.478	***	−0.459	0.124	0.632	−0.368	***
Control	Private	−0.165	0.237	0.848			−0.249	0.200	0.779
Intercept		−2.413	0.388			***	−3.629	0.459			***
Gender	Female	0.615	0.131	1.849	0.849	***	3.032	0.312	20.746	19.746	***
Race	Asian	−1.092	0.958	0.336			0.554	0.240	1.740	0.740	**
Black	−0.045	0.235	0.956			0.892	0.314	2.440	1.440	***
Hispanic	0.087	0.278	1.091			1.254	0.319	3.504	2.504	***
Parent’s education	First generation	−0.012	0.183	0.988			0.219	0.254	1.245
Family income	Low	1.738	0.289	5.685	4.685	***	0.319	0.404	1.375
Low middle	1.387	0.275	4.003	3.003	**	0.501	0.774	1.650
High middle	0.456	0.274	1.578			1.022	0.411	2.778	1.778	**
College GPA	Continuous	0.028	0.099	1.029			−0.583	0.120	0.558	−0.442	***
Pell amount	Continuous (per $1,000)	−0.219	0.068	0.804	−0.196	***	0.271	0.076	1.311	0.311	***
Loan amount	Continuous (per $1,000)	−0.025	0.015	0.975			−0.233	0.034	0.792	−0.208	***
Net price	Continuous (per $1,000)	−0.071	0.020	0.931	−0.069	***	−0.023	0.023	0.978
Intensity	Full time	−0.416	0.162	0.660	−0.340	**	−0.744	0.299	0.475	−0.525	**
Control	Private	−0.062	0.191	0.940			−0.594	0.542	0.552

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Jacob A. Kamer is a PhD candidate in the Higher Education Administration program at the University of Tennessee, Knoxville.

Terry T. Ishitani is an associate professor in the Educational Leadership & Policy Studies department at the University of Tennessee, Knoxville. He received his doctoral degree from the University of Iowa.

References

Baker

C. N.

Robnett

(2012). Race, social support and college student retention: A case study. Journal of College Student Development, 53(2), 325–335.

Bean

J. P.

Metzner

B. S.

(1985). A conceptual model of nontraditional undergraduate student retention. Review of Educational Research, 55(4), 485–540.

Becker

G. S.

(1993). Human capital: A theoretical and empirical analysis with special reference to education (3rd ed.). Chicago, IL: The University of Chicago Press.

Bergman

Gross

J. P. K.

Berry

Shuck

(2014). If life happened but a degree didn’t: Examining factors that impact adult student persistence. The Journal of Continuing Higher Education, 62(2), 90–101.

Bettinger

(2004). How financial aid affects persistence. In Hoxby

C. M.

(Ed.), College choices: The economics of where to go, when to go, and how to pay for it (pp. 207–237). Chicago, IL: University of Chicago Press.

Blossfeld

H. -P.

Rohwer

(1995). Techniques of event history modeling: New approaches to causal analysis. Mahwah, NJ: Lawrence Erlbaum Associates.

Blumenstyk

(2018). The adult student: The population colleges—and the nation—can’t afford to ignore. Washington, DC: The Chronicle of Higher Education.

Brändle

(2017). How availability of capital affects the timing of enrollment: The routs to university of traditional and non-traditional students. Studies in Higher Education, 42(12), 2229–2249.

Cabrera

N. L.

Miner

D. D.

Milem

J. F.

(2013). Can a summer bridge program impact first-year persistence and performance? A case study of the New Start Summer Program. Research in Higher Education, 54(5), 481–498.

10.

Caruth

G. D.

(2018). Student engagement, retention, and motivation: Assessing academic success in today’s college students. Participatory Educational Research, 5(1), 17–30.

11.

Chen

Hossler

(2017). The effects of financial aid on college success of two-year beginning nontraditional students. Research in Higher Education, 58(1), 40–76.

12.

Chen

(2012). Institutional characteristics and college student dropout risks: A multilevel event history analysis. Research in Higher Education, 53, 487–505.

13.

Chen

DesJardins

S. L.

(2007). Exploring the effects of financial aid on the gap in student dropout risks by income level. Research in Higher Education, 49(1), 1–18.

14.

Chen

(2005). First generation students in postsecondary education: A look at their college transcripts (NCES 2005-171). Washington, DC: U.S. Department of Education.

15.

Choy

(2002). Nontraditional undergraduates (NCES 2002–012). Washington, DC: U.S. Department of Education.

16.

Chung

Turnbull

Chur-Hansen

(2017). Differences in resilience between ‘traditional’ and ‘non-traditional’ university students. Active Learning in Higher Education, 18(1), 77–87.

17.

Davidson

Wilson

(2013). Reassessing Tinto’s concepts of social and academic integration in student retention. Journal of College Student Retention: Research, Theory & Practice, 15(3), 329–346.

18.

Dougherty

B. C.

Woodland

(2009). Understanding sources of financial support for adult learners. Journal of Continuing Higher Education, 57(3), 181–186.

19.

Drive to 55 Alliance. (2018). Drive to 55. Retrieved from http://driveto55.org/

20.

Ethington

C. A.

Thomas

S. L.

Pike

G. R.

(2002). Back to the basics: Regression as it should be. In Smart

J. C.

(Ed.), Higher education: Handbook of theory and research ( Vol. 17, pp. 263–293). New York, NY: Agathon

21.

Federal Student Aid. (2018). FAFSA. Retrieved from https://studentaid.ed.gov/sa/fafsa

22.

Gilardi

Guglielmetti

(2011). University life of non-traditional students: Engagement styles and impacts on attrition. The Journal of Higher Education, 82(1), 33–53.

23.

Ishitani

T. T.

(2016). First-generation students’ persistence at four-year institutions. College and University, 91(22),

24.

Ishitani

T. T.

DesJardins

S. L.

(2002). A longitudinal investigation of dropout from college in the United States. Journal of College Student Retention: Research, Theory & Practice, 4(2), 173–201.

25.

Juszkiewicz

(2017). Trends in community college enrollment and completion data, 2017. Washington, DC: American Association of Community Colleges.

26.

Laanan

F. S.

Starobin

S. S.

Eggleston

L. E.

(2010). Adjustment of community college students at a four-year university: Role and relevance of transfer student capital for student retention. Journal of College Student Retention: Research, Theory & Practice, 12(2), 175–209.

27.

Laing

Robinson

(2003). The withdrawal of non-traditional students: Developing an explanatory model. Journal of Further and Higher Education, 27(2), 175–185.

28.

(2010). They need help: Transfer students from four-year to four-year institutions. The Review of Higher Education, 33(2), 207–238.

29.

Lumina Foundation (2017, November 30). Increasing support for adult student success goal of Lumina Foundation adult promise initiative. Today’s Student. Retrieved from https://www.luminafoundation.org/news-and-views/increasing-support-for-adult-student-success-goal-of-lumina-foundation-adult-promise-initiative

30.

Baum

(2016). Trends in community colleges: Enrollment, prices, student debt, and completion. New York, NY: College Board.

31.

Markle

(2015). Factors influencing persistence among nontraditional university students. Adult Education Quarterly, 65(3), 267–285.

32.

Metzner

B. S.

Bean

J. P.

(1987). The estimation of a conceptual model of nontraditional undergraduate student attrition. Research in Higher Education, 27(1), 15–38.

33.

Offerman

(2011). Profile of nontraditional doctoral degree students. New Directions for Adult and Continuing Education, 129, 21–30.

34.

Ohio Department of Higher Education. (2015). Student success for adult learners: 8th report on the condition of higher education in Ohio [PDF file]. Retrieved from https://www.ohiohighered.org/sites/ohiohighered.org/files/uploads/board/condition-report/2015-Conditions-Report_FINAL_web.pdf

35.

Osam

E. K.

Bergman

Cumberland

D. M.

(2017). An integrative literature review on the barriers impacting adult learners’ return to college. Adult Learning, 28(2), 54–60.

36.

Paulsen

M. B.

St. John

E. P.

(2002). Social class and college costs: Examining the financial nexus between college choice and persistence. Journal of Higher Education, 73(2), 189–236.

37.

Peralta

K. J.

Klonowski

(2017). Examining conceptual and operational definitions of ‘first-generation college student’ in research on retention. Journal of College Student Development, 58(4), 630–636.

38.

Perna

L. W.

(2005). Sex differences in faculty tenure and promotion: The contribution of family ties. Research in Higher Education, 46(3), 277–307.

39.

Rabourn

K. E.

BrckaLorenz

Shoup

(2018). Reimagining student engagement: How nontraditional adult learners engage in traditional postsecondary environments. The Journal of Continuing Higher Education, 66(1), 22–33.

40.

Roberts

Styron

(2010). Student satisfaction and persistence: Factors vital to student retention. Research in Higher Education, 6(3), 1–18.

41.

Rowan-Kenyon

H. T.

(2007). Predictors of delayed college enrollment and the impact of socioeconomic status. The Journal of Higher Education, 78(2), 188–214.

42.

Senie

K. C.

(2016). Implementing transfer and articulation: A case study of community colleges and state universities. Community College Journal of Research & Practice, 40(4), 269–284.

43.

Shapiro

Dundar

Huie

Wakhungu

Yuan

Nathan

Hwang

Y. A.

(2017, April). A national view of student attainment rates by race and ethnicity–Fall 2010 cohort. Herndon, VA: National Student Clearinghouse Research Center.

44.

Singer

J. D.

Willett

J. B.

(1993). It’s about time: Using discrete-time survival analysis to study duration and the timing of events. Journal of Educational Statistics, 18(2), 155–195.

45.

Spellman

(2007). Enrollment and retention barriers adult students encounter. Community College Enterprise, 13(1), 63–80.

46.

Tabachnick

B. G.

Fidell

L. S.

(2007). Using multivariate statistics (5th ed.). New York, NY: Pearson.

47.

Taniguchi

Kaufman

(2005). Degree completion among nontraditional college students. Social Science Quarterly, 86(4), 912–927.

48.

Taylor

House

(2010). An exploration of identity, motivations and concerns of non-traditional students at different stages of higher education. Psychology Teaching Review, 16(1), 46–57.

49.

Tennessee Higher Education Commission. (2017). 2016-2017 Tennessee higher education adult student fact book [PDF file]. Retrieved from https://www.tn.gov/content/dam/tn/thec/bureau/research/other-research/adult_fb/2016-17_Adult_Student_FB_SUPPRESSED.pdf

50.

The University of Memphis. (2019). Adult student services center. Retrieved from https://www.memphis.edu/adultstudents/

51.

Tinto

(1982). Limits of theory and practice in student attrition. The Journal of Higher Education, 53(6), 687–700.

52.

Tinto

(1987). Leaving college. Chicago, IL: University of Chicago Press.

53.

Washington Student Achievement Council (2014). Adult reengagement. Retrieved from https://www.wsac.wa.gov/adult-reengagement

54.

Wells

R. S.

Lynch

C. M.

(2012). Delayed college entry and the socioeconomic gap: Examining the roles of student plans, family income, parental education, and parental occupation. The Journal of Higher Education, 83(5), 671–697.

55.

Wyatt

L. G.

(2011). Nontraditional student engagement: Increasing adult student success and retention. The Journal of Continuing Higher Education, 59(1), 10–20.

56.

Yamaguchi

(1991). Event history analysis. Newbury Park, CA: SAGE.

		Year 1					Year 2
Variable	Label	Coefficient	Standard error	Odds	Δ	Significance	Coefficient	Standard error	Odds	Δ	Significance
Intercept		−5.866	0.735			***	−0.684	0.662
Gender	Female	−1.062	0.215	0.346	−0.654	***	−0.318	0.170	0.728	−0.272	*
Race	Asian	−2.459	0.265	0.086	−0.914	***	−1.476	0.306	0.228	−0.772	***
	Black	1.497	0.221	4.468	3.468	***	−0.073	0.184	0.929
	Hispanic	0.962	0.207	2.617	1.617	***	0.147	0.102	1.158
Parent’s education	First generation	2.559	0.774	12.927	11.927	***	0.259	0.423	1.296
Family income	Low	1.038	0.215	2.824	1.824	***	0.238	0.217	1.269
	Low middle	−0.894	0.321	0.409	−0.591	***	0.469	0.253	1.598	0.598	*
	High middle	0.211	0.318	1.235			0.668	0.243	1.950	0.950	**
College GPA	Continuous	0.594	0.110	1.811	0.811	***	−0.193	0.041	0.824	−0.176	***
Pell amount	Continuous (per $1,000)	−1.004	0.173	0.366	−0.634	***	−0.226	0.036	0.798	−0.202	***
Loan amount	Continuous (per $1,000)	−0.322	0.076	0.725	−0.275	***	−0.099	0.026	0.905	−0.095	***
Net price	Continuous (per $1,000)	−0.443	0.058	0.642	−0.358	***	0.074	0.013	1.077	0.077	***
Intensity	Full time	0.907	0.271	2.478	1.478	***	−0.459	0.124	0.632	−0.368	***
Control	Private	−0.165	0.237	0.848			−0.249	0.200	0.779
Intercept		−2.413	0.388			***	−3.629	0.459			***
Gender	Female	0.615	0.131	1.849	0.849	***	3.032	0.312	20.746	19.746	***
Race	Asian	−1.092	0.958	0.336			0.554	0.240	1.740	0.740	**
	Black	−0.045	0.235	0.956			0.892	0.314	2.440	1.440	***
	Hispanic	0.087	0.278	1.091			1.254	0.319	3.504	2.504	***
Parent’s education	First generation	−0.012	0.183	0.988			0.219	0.254	1.245
Family income	Low	1.738	0.289	5.685	4.685	***	0.319	0.404	1.375
	Low middle	1.387	0.275	4.003	3.003	**	0.501	0.774	1.650
	High middle	0.456	0.274	1.578			1.022	0.411	2.778	1.778	**
College GPA	Continuous	0.028	0.099	1.029			−0.583	0.120	0.558	−0.442	***
Pell amount	Continuous (per $1,000)	−0.219	0.068	0.804	−0.196	***	0.271	0.076	1.311	0.311	***
Loan amount	Continuous (per $1,000)	−0.025	0.015	0.975			−0.233	0.034	0.792	−0.208	***
Net price	Continuous (per $1,000)	−0.071	0.020	0.931	−0.069	***	−0.023	0.023	0.978
Intensity	Full time	−0.416	0.162	0.660	−0.340	**	−0.744	0.299	0.475	−0.525	**
Control	Private	−0.062	0.191	0.940			−0.594	0.542	0.552