What Makes Adult Learners Persist in College? An Analysis Using the Nontraditional Undergraduate Student Attrition Model

Abstract

This research examines the factors influencing drop out among adult college students. As the traditional-age student population (ages 19–24) declines, the older, part-time, adult learners have emerged as a critical enrollment demographic for higher education institutions. These learners often pursue higher education for career advancement, re-skilling, or re-employment. However, many encounter difficulties in sustaining their academic engagement due to low motivation, limited basic learning skills, or external constraints. Despite the growing presence of adult learners in Korean universities, limited research has analyzed drop-out factors within this specific context. To address this gap, this study applies Bean and Metzner’s nontraditional undergraduate student attrition model, using data from the Korean Educational Longitudinal Study (KELS). It investigates how background characteristics, academic variables, environmental factors, and academic and psychological outcomes influence the drop out of adult learners. The findings reveal that academic variables significantly impact drop-out intentions, while student engagement and social integration show minimal effects. These results offer valuable theoretical insights and practical implications for enhancing adult learner retention in higher education.

1. Introduction

Traditionally, university students in East Asian countries such as Korea and Japan have predominantly been within the traditional age range of 19 to 24. However, in recent years, a new demographic is emerging in higher education: adult learners. Also referred to as nontraditional students, adult learners are typically defined as individuals over the age of 24, part-time students, those with regular employment or dependents to care for, or a combination of these factors (Bean & Metzner, 1985).

The rising demand for higher education among adult learners is driven by various factors. Since 1980, the demand for workers with postsecondary education degrees has grown by 3% per year, while the supply of talent has increased by only 1% (Carnevale & Rose, 2015). Additionally, the introduction of new technologies is transforming the industrial structure, and labor market flexibility is increasing globally. This shift necessitates re-skilling or upskilling through postsecondary education (Gardner et al., 2022). Moreover, as life expectancy increases, individuals need to prepare for new careers after retirement. Consequently, some pursue higher education degrees to secure higher pay, promotions, or career changes, while others engage in higher education for personal development (Bowers & Bergman, 2016; U.-S. Choi, 2006; Lyz & Gladkaya, 2024; Sogunro, 2015).

The growth of adult learners is not limited to Korea; it is a global phenomenon. For example, in fall 2022, approximately 2.9 million students over the age of 25 were enrolled in undergraduate programs in the U.S. (U.S. Department of Commerce, 2022). In 2021, the adult participation rate in lifelong learning was 11% in the EU (Eurostat, 2024). In Korea, the number of adult learners is rising due to government policy initiatives and growing demand from adult learners.

Recently, most Korean universities have struggled to fill their admission quotas. Due to the low birth rate, the school-aged population is declining, and Korea is projected to become the world’s oldest population in the near future. Figure 1 shows that the admission quotas of higher education institutions exceeded the 18-year-old population in 2020. In 2024, 169 universities in Korea failed to meet their admission quotas, representing 85% of all four-year institutions in the country (Seo & Kim, 2024).

In response to these circumstances, developing specialized curricula and learning environments for adult learners has been considered a strategic approach for universities to fill unmet admission quotas. The Korean government has promoted establishing degree programs for adult learners under the “Lifelong Education at Universities for the Future of Education (LiFE)” policy since the early 2010s (Na et al., 2024). In 2024, a total of 30 universities—approximately 15% of all four-year institutions in Korea—officially participated in this project, with many other universities also offering degree programs for adult learners.

Today, there is a consensus among researchers that establishing a learning system to meet the needs of adult learners has become a critical task for most Korean universities (Y. J. Choi & Park, 2018; Rhee, 2021). Although many universities are developing and offering degree programs for adult learners, a significant number of these students drop out for various reasons. However, to date, limited research has examined adult learners enrolled in regular degree programs within the Korean context. While previous studies have focused on adult learners in distance education institutions (S. A. Kwon et al., 2020; Y. M. Kim, 2017; Yang & Jung, 2019; Shin et al., 2020; Joung et al., 2018), there is a notable gap in research addressing student retention among adult learners.

Considering these situations, the purpose of this study is to analyze the factors influencing drop-out intentions among adult learners in Korea, with a specific focus on Bean & Metzner’s nontraditional undergraduate student attrition model (Bean & Metzner, 1985).

2. Literature Reviews

2.1. Characteristics of Adult Learners

According to previous research, it is known that adult learners have different characteristics than traditional-aged students. Usually, adult learners have clear purposes for entering degree programs than traditional-aged students. They are typically more goal-oriented, often entering higher education with clearly defined academic or career-related purposes (Ku et al., 2015). It is also known that adult learners usually demonstrate higher levels of intrinsic motivation for learning and pursuing a higher education degree than their traditional-aged counterparts, with intrinsic motivation being more strongly associated with positive academic experiences and emotional engagement (Bye et al., 2007; Fuhrmann, 1997). In sum, adult learners are self-directed, voluntary, and responsible for their learning, which enhanced learning engagement and learner desired outcomes (Conlan et al., 2003). These characteristics are known to positively influence the academic persistence of adult learners.

However, adult learners also exhibit certain characteristics that may negatively affect their academic persistence. Compared with traditional-aged students, adult learners are more likely to be employed full time, have dependents or family obligations, and attend college part time, which makes them more vulnerable to external pressures that can hinder academic persistence and increase the likelihood of drop out (Y. M. Kim & Han, 2012). In addition, because of these characteristics, usually adult learners have trouble securing study hours (U.-S. Choi, 2006; Jeong, 2019). Moreover, since adult learners often resume ‘learning’ and ‘studying’ activities after a while, they are more likely to have difficulty with learning skills and need assistance in learning (I. Kim, 2022).

2.2. Theoretical Background About Student Attrition

For a long decade, student attrition has been one of the major concerns for researchers. Therefore, there are various kinds of theories and models that explain student attrition in higher education in various perspectives (Pascarella & Terenzini, 1980; Tinto, 2012). For example, Tinto’s student drop-out model (Tinto, 1993, 2012) is one of the most well-known student attrition models, with strengths in explaining student stop-out and drop-out. However, including Tinto’s model, most student attrition models were developed focusing on traditional-aged students, who recently graduated high school, are under age 25, are enrolled full time, and are residing at or near their college.

However, considering the characteristics of adult learners, traditional student retention models, which emphasize the influence of social and academic integration, student involvement, student engagement, or the influence of peer groups may not be appropriate for adult learners. In other words, these factors may not be important when we consider the characteristics of adult learners.

In this perspective, it is required to analyze the drop-out factors of adult learners, with theoretical backgrounds focusing on adult learners. There is research that develops a student attrition model for adult learners. Metzner and Bean (1987) developed a student attrition model in 1986 based on the combination of turnover theories of industrial workers and student attrition models. The model consists of six categories of factors: personal backgrounds, academic variables, environmental variables, social integration variables, academic outcomes, psychological outcomes, intent to leave, and drop out (see Figure 2).

Building on the foundations of Bean and Metzner’s (1985) nontraditional student attrition model, subsequent research has sought to refine our understanding of the factors influencing persistence among adult learners. Braxton et al. (2004) extended student departure theories by emphasizing the unique needs of commuter and nontraditional students, many of whom share similar characteristics with adult learners. Their model underscores the importance of academic integration through classroom experiences, institutional commitment, and the role of external environments, such as family support and work responsibilities, which are not fully captured in traditional student retention models. They argue that strategies like active learning, flexible course scheduling, and family-inclusive support structures are key to improving persistence for students who balance multiple life roles.

Further advancing this line of research, M. J. Bergman (2012) and M. J. Bergman et al. (2014a) adapted Braxton et al.’s framework to specifically examine adult learners enrolled in degree-completion programs. Bergman’s model identifies three primary domains influencing persistence: entry characteristics, such as prior academic experiences and education goals; the internal academic environment, including faculty responsiveness and the use of active, relevant learning strategies; and external factors, such as financial stability, family support, and employer encouragement. Empirical findings from Bergman’s studies reveal that educational aspirations (goal orientation), sufficient financial resources, and engaging academic practices are the strongest predictors of adult student persistence.

Beyond these models, other studies have examined the persistence of adult learners, particularly in online education contexts. For example, Park and Choi (2009) and Deschacht and Goeman (2015) identified course relevance and institutional support or responsiveness as critical factors influencing retention. These studies also emphasize the significance of environmental variables, such as support from family and the workplace, which align with the environmental factors outlined in Bean and Metzner’s (1985) model. Moreover, research by Kara et al. (2019) underscores the importance of personal study skills, including time management, self-regulation, and technological proficiency, as essential competencies for persistence in online learning environments. However, since these studies primarily focus on online adult learners, there remains a need for new research exploring the specific challenges faced by adult learners enrolled in traditional degree programs, where institutional integration and academic identity formation may play a greater role.

Recent theoretical perspectives complement these models by highlighting adult learner motivation and re-engagement through frameworks such as self-directed learning (Knowles, 1980), the goal orientation theory (Ames, 1992), and academic identity formation (Kasworm, 2003). The self-directed learning theory emphasizes that adult learners are more likely to persist when they perceive autonomy, relevance, and control over their learning. The goal orientation theory, which resonates with Bergman’s findings on educational aspirations, suggests that adults persist when their academic efforts are aligned with personal mastery and professional advancement. Academic identity formation research further indicates that persistence increases when adults see themselves as capable and legitimate learners, integrating academic success into their broader sense of self. These contemporary theories enrich Bean and Metzner’s original framework by addressing the intrinsic motivational and identity-based factors that drive adult student persistence.

In this regard, based on Bean & Metzner’s research and other research on student attrition or student retention in and outside of Korea, this study suggested an analysis framework (see Figure 3). The framework includes institutional characteristics, student engagement, scholarship, and attendance period in addition to Bean & Metzner’s model, based on the literature review.

3. Materials and Methods

3.1. Description of Data

The data utilized for the analysis was from the KELS (Korean Education Longitudinal Study). The KELS is one of the most representative large scaled, longitudinal national surveys designed for making educational policy. The initial cohort in 2005 consisted of 6908 seventh-grade students sampled from middle schools nationwide, with follow-up surveys conducted biennially through to 2020. For the purposes of this study, data from the 9th follow-up (2018) and 10th follow-up (2020) waves were utilized.

The samples were selected along with the following steps: First, individuals who were currently university students or student–employees were identified. Next, only those who entered their current university after the age of 24 were retained. In other words, students who entered university immediately after high school graduation (between 2011 and 2015) and continued until 2018 were excluded. Then, the individuals who enrolled in universities outside of Korea and had excessive missing values in key variables were excluded. After the final step, 295 adult learners were included for the analysis. The descriptive statistics for the final sample are shown in

Table 1. Descriptive statistics.

Characteristics		N	%
Institution type	Community college	64	21.7
	4-year University	231	78.3
Age	27	210	71.2
	29	85	28.8
Attending periods	1 year	74	25.1
	2 years	112	38.0
	3 years	68	23.1
	4 years	41	13.9
Employment status	Employed	146	49.5
	Not employed	149	50.5
Marital status	Married/living with partner	17	5.8
	Not married	278	94.2
Total		295	100.0

.

3.2. Measurement

The list of the variables and measurements are shown in

Table 2. List of the variables and measurements.

Variable Name	Measurement	Mean	SD
Dependent variables
Drop-out intentions
Stop-out intention	5 Likert scale (1 = strongly disagree~5 = strongly agree) for “I am thinking about leaving college for a while and finishing my degree program later”	2.31	1.162
Drop-out intention	5 Likert scale (1 = strongly disagree~5 = strongly agree) for “I want to drop out”	2.16	1.070
Independent variables
Personal backgrounds
Age	Birth age, 0 = age 27, 1 = age 29	0.29	0.454
Gender	0 = female, 1 = male	0.42	0.495
Educational goal	Desired degree: 0 = undecided, 1 = high school~5 = PhD	2.57	1.602
Institutional characteristics
Type of institution	0 = community college, 1 = 4-year institution	0.78	0.413
Location	0 = not Seoul province, 1 = Seoul province	0.53	0.500
Enrollment state
Grade	Number of years attending current institution	2.26	0.987
Number of credits registered	Number of credits registered this semester: 0 = 0 credits, 1 = 1 to 5 credits, 2 = 6 to 10 credits, 3 = 11 to 15 credits, 4 = 16 to 20 credits, 5 = 21 or more credits	2.77	1.940
Academic variables
Study hours	8 Likert scale (1 = never~9 = more than 21 h per week) for “Use time for Study”	3.58	1.969
Academic engagement	Mean of 7 questions such as “asking questions in the class” measured by 5 Likert scale (1 = never~5 = always)	2.98	0.720
Quality of academic advising	Mean of 2 questions such as “Professors are enthusiastic about the education of students” measured by 5 Likert scale (1 = strongly disagree~5 = strongly agree)	3.43	0.728
Study skills	Mean of 2 recoded questions such as “Classes and assignments felt difficult” measured by 5 Likert scale (1 = strongly disagree~5 = strongly agree)	1.70	0.728
Academic negligence	Mean of 3 questions such as “skip classes for no reason” measured by 5 Likert scale (1 = never~5 = always)	2.79	0.823
Certainty about current university	Mean of 2 questions such as “I am satisfied with the decision to attend my current university” measured on 5 Likert scale (1 = strongly disagree~5 = strongly agree)	3.32	0.766
Course availability	5 Likert scale: “There are various courses that students want” measured by 5 Likert scale (1 = strongly disagree~5 = strongly agree)	3.16	0.929
Social integration
Membership in campus organization	Use time for campus organization measured by 9 Likert scale (1 = “never”~9: “more than 21 h per week”)	1.42	0.948
Faculty contact	Mean of 6 questions such as “How often do you exchange greeting with professors?” measured on 6-point scale (1 = “never”, 2 = “once or twice per semester”, 3 = “once or twice per month”, 4 = “once a week”, 5 = “two or three times a week”, 6 = “almost every day”)	2.69	1.400
Friend contact	Mean of 6 questions such as “How often do you have conversations about personal matters with friends?” measured on 6-point scale (1 = “never”, 2 = “once or twice per semester”, 3 = “once or twice per month”, 4 = “once a week”, 5 = “two or three times a week”, 6 = “almost every day”)	2.63	1.459
Environmental variables
Concern for tuition	“I feel anxiety over tuition fees” measured by 5 Likert scale (1 = “strongly disagree”~5 = “strongly agree”)	2.03	0.951
Scholarship	0 = no scholarship in the previous semester, 1 = earn scholarship in the previous semester	0.27	0.444
Hours of employment	Working hours per week measured on 7-point scale (0 = “none”, 1 = “up to 10 h”, 2 = “up to 20 h”, 3 = “up to 30 h”, 4 = “up to 40 h”, 5 = “up to 50 h”, 6 = “more than 50 h”)	2.19	2.332
Parental involvement	Mean of 4 questions such as “Parental involvement in course selection” measured on a 5-point Likert scale (1 = “hardly involved at all”~5 = “highly involved”)	2.04	0.943
Marital status	0 = “not married”, 1 = “married”	0.03	0.181
Academic outcome
GPA	Average grade level	3.20	0.992
Psychological outcomes
Utility	Mean of 2 questions such as “I found out why I am attending college and what I want to get” measured on 5 Likert scale (1 = “strongly disagree”~5 = “strongly agree”)	3.38	0.718
Satisfaction	Mean of 6 questions such as “satisfaction with the quality of class” measured on 5 Likert scale (1 = “very dissatisfied”~5 = “very satisfied”)	3.28	0.720

* Binary variables (e.g., gender, type of institution, and location) are coded as 0 and 1. Therefore, the mean values represent the proportion of respondents coded as ‘1’ (e.g., a mean of 0.42 for gender indicates 42% of respondents are male). Detailed distributions (frequencies and percentages) for these binary variables are provided separately in Table 1.

. The variables were measured based on the literature reviews. This study is grounded in two well-established frameworks for understanding student persistence and development: Tinto’s (1993) model of student departure and Bean & Metzner’s (1985) nontraditional student attrition model. These frameworks provide a comprehensive lens to analyze how individual characteristics, environmental factors, and institutional experiences shape both academic outcomes and psychological development.

Following Bean’s student attrition model and Tinto’s model, which conceptualizes the intention to leave as a function of student inputs, the educational environment, and academic and psychological outcomes, the independent variables were categorized accordingly. Academic engagement and social integration were treated as environmental (process) variables, reflecting students’ interaction with the learning environment and social context. Academic and psychological outcomes were positioned as distal outputs of the educational experience.

The dependent variables were stop-out intention and drop-out intention, each measured using a 5-point Likert scale. An item for stop-out intention is “I am thinking about leaving college temporarily and finishing my degree program later,” and for drop-out intention is “I want to drop out.” According to previous studies on adult learners’ drop out, independent variables were selected based on factors identified as influential in prior research (M. J. Bergman, 2012; Park & Choi, 2009). These variables included personal backgrounds (age, gender, and educational goal), institutional characteristics (type of institution and location), and enrollment state (grade and number of credits registered in this semester).

Academic engagement was measured using a set of behavioral indicators adapted from the National Assessment of Student Engagement in Learning for Korean Universities (Yu et al., 2012), which itself is based on the U.S. NSSE framework (National Survey of Student Engagement, 2013). These included items on classroom participation (e.g., asking questions and expressing opinions), metacognitive behaviors (e.g., evaluating information quality and seeking feedback), and self-directed learning (e.g., voluntary writing practice and independent study). In addition, the quality of academic advising was measured using items adapted from adult learner retention literature that treated academic advising as an independent dimension of student–faculty support (e.g., Anderson, 2011). The questionnaires for study challenges and academic negligence are adopted from the Motivated Strategies for Learning questionnaire (MSLQ) subscale (Pintrich et al., 1991) for effort and cognitive strategy. Students’ certainty about their current university was measured by questionnaires adopted from the ‘institutional commitment’ items of Bean and Metzner’s (1985) study. Course availability was measured by items adopted from the ‘institutional support’ items of Bean and Metzner’s (1985) study.

Social integration, drawing from Tinto’s emphasis on academic and social involvement as retention predictors, was operationalized through three dimensions (Tinto, 1993; Pascarella & Terenzini, 1980): participation in campus organizations (frequency of weekly involvement), contact with faculty (e.g., greeting or consulting with professors), and peer interaction (e.g., conversations on personal matters with friends).

Environmental variables were conceptualized using Bean and Metzner’s student attrition model (1985), which posits that external factors significantly affect persistence decisions. These included concern for tuition (measured on a Likert scale), scholarship status (binary-coded), weekly hours of employment, parental involvement (outside encouragement), and marital status.

Finally, outcome variables were grouped into academic and psychological domains. Academic outcome was measured by students’ GPA. Psychological outcomes included utility, indicating clarity of educational purpose (e.g., “I found out why I am attending college”), and satisfaction, reflecting the perceived quality of learning experiences (e.g., “satisfaction with the quality of class”).

For constructs measured using multiple items (e.g., academic engagement, social integration, and psychological outcomes), confirmatory factor analysis (CFA) was conducted to assess the construct validity of the measurement models. In addition, internal consistency reliability was evaluated using Cronbach’s alpha coefficients, with values above 0.70 considered acceptable for research purposes (Nunnally & Bernstein, 1994). The range of reliability was α = 0.702~0.962. The results of the confirmatory factor analyses and internal consistency reliability (Cronbach’s alpha) for each construct are provided in Appendix A.

3.3. Analysis

Descriptive statistics were conducted to examine the characteristics of the sample. To validate the construct structure of the measurement variables, confirmatory factor analysis (CFA) was performed, and the internal consistency of each scale was assessed using Cronbach’s alpha. To address the research questions, multiple regression analyses were conducted to identify factors influencing the two dependent variables: stop-out intention and drop-out intention. All analyses were conducted using IBM SPSS Statistics version 26.0.

4. Results

The results of the analysis were shown in

Table 3. Factors affecting stop-out intention of adult learners.

		DV: Stop-Out
Category	Variables	B		s.e	β
	(Constant)		3.376 ***	0.989
Personal backgrounds	Age		0.095	0.204		0.037
	Gender	-	−0.313	0.167	-	−0.129
	Educational goal		0.001	0.053		0.002
Institutional characteristics	Type of institution		0.351	0.208		0.119
	Location	-	−0.056	0.176	-	−0.023
Enrollment state	Grade		0.008	0.091		0.006
	Number of credits registered	-	−0.210 ***	0.057	-	−0.253
Academic variables	Study hours	-	−0.090	0.048	-	−0.147
	Academic engagement		0.070	0.138		0.042
	Quality of academic advising	-	−0.336 *	0.148	-	−0.206
	Academic negligence		0.020	0.113		0.013
	Study skills	-	−0.334 **	0.108	-	−0.225
	Certainty about current university		0.329 *	0.143		0.215
	Course availability		0.171	0.099		0.140
Social integration	Membership in campus organization	-	−0.035	0.086	-	−0.030
	Faculty contact		0.167	0.100		0.184
	Friend contact		0.007	0.093		0.009
Environmental variables	Scholarship	-	−0.177	0.170	-	−0.072
	Concern for tuition		0.144	0.088		0.115
	Hours of employment		0.066	0.043		0.128
	Parental involvement		0.149	0.090		0.119
	Marital status		0.662	0.449		0.097
Academic outcome	GPA	-	−0.081	0.100	-	−0.071
Psychological outcome	Utility of college education	-	−0.112	0.156	-	−0.069
	Satisfaction about college education	-	−0.092	0.174	-	−0.059
R2 = 0.379, adjusted R2 = 0.283

* p < 0.05, ** p < 0.01, *** p < 0.001.

and

Table 4. Factors affecting drop-out intention of adult learners.

		DV: Drop out
Category	Variables	B		s.e	β
	(Constant)		3.981 ***	0.816
Personal backgrounds	Age		0.222	0.168		0.098
	Gender		0.006	0.138		0.003
	Educational goal		0.008	0.043		0.011
Institutional characteristics	Type of institution		0.063	0.172		0.024
	Location	-	−0.212	0.145	-	−0.102
Enrollment state	Grade		0.029	0.075		0.028
	Number of credits registered	-	−0.103 *	0.047	-	−0.143
Academic variables	Study hours	-	−0.128 **	0.039	-	−0.239
	Academic engagement		0.108	0.113		0.075
	Quality of academic advising	-	−0.184	0.122	-	−0.129
	Academic negligence		0.060	0.093		0.044
	Study skills	-	−0.272 **	0.089	-	−0.210
	Certainty about current university		0.010	0.118		0.007
	Course availability		0.036	0.081		0.034
Social integration	Membership in campus organization	-	−0.064	0.071	-	−0.063
	Faculty contact		0.117	0.082		0.147
	Friend contact		0.062	0.077		0.085
Environmental variables	Scholarship	-	−0.260	0.140	-	−0.122
	Concern for tuition		0.055	0.072		0.050
	Hours of employment		0.008	0.035		0.018
	Parental involvement		0.194 **	0.074		0.179
	Marital status		0.257	0.371		0.043
Academic outcome	GPA	-	−0.201 *	0.083	-	−0.201
Psychological outcome	Utility of college education	-	−0.129	0.129	-	−0.092
	Satisfaction about college education		0.053	0.144		0.039
R2 = 0.442, adjusted R2 = 0.356

* p < 0.05, ** p < 0.01, *** p < 0.001.

. First, Table 3 shows the result of factors affecting the stop-out intention of adult learners. With an adjusted R² = 0.283, the model demonstrates moderate explanatory power—indicating that nearly one-third of the variance in stop-out intentions can be explained by the included predictors.

The results identified several key factors that significantly influenced stop-out intention among adult learners. Among the predictors, the number of credits registered (β = −0.253, p < 0.001) was negatively associated with stop-out, indicating that students who enrolled in more academic credits were less likely to discontinue their studies temporarily. This suggests that heavier course loads may reflect higher academic commitment and engagement.

Additionally, the quality of academic advising (β = −0.206, p < 0.05) emerged as a significant protective factor against stop-out. This finding highlights the critical role of supportive and effective academic guidance in helping adult learners navigate their educational pathways. Similarly, study skills (β = −0.225, p < 0.01) were negatively associated with stop-out behavior, underscoring the importance of academic preparedness and self-regulated learning in adult student persistence.

Interestingly, the certainty about attending the current university was positively associated with stop-out (β = 0.215, p < 0.05). While this may appear counterintuitive, it could reflect a unique characteristic of adult learners who, despite feeling secure in their institutional choice, may pause their studies with the intent of returning at a later time due to competing life demands.

Other variables such as GPA, satisfaction with college education, hours of employment, and social integration factors (e.g., faculty and peer contact) were not statistically significant in this model. However, some of these, particularly faculty contact and concern for tuition, exhibited moderate effect sizes and may warrant further investigation in future research.

Table 4 presents the results of the regression analysis examining factors influencing adult learners’ drop-out intentions. The adjusted R² value was 0.356, indicating a moderate to strong level of goodness-of-fit.

Among the academic variables, study hours (β = −0.239, p < 0.01) and study skills (β = −0.210, p < 0.01) were both negatively associated with drop-out intention. This suggests that students who spend more time studying and possess stronger learning strategies are less likely to consider dropping out. Additionally, the number of credits registered (β = −0.143, p < 0.05) was a significant negative predictor, reflecting that heavier course enrollment may be indicative of greater academic commitment and persistence.

In terms of environmental factors, outside encouragement—particularly from parents or close supporters—was unexpectedly found to be positively associated with drop-out intention (β = 0.179, p < 0.01). This result contrasts with previous research, which suggests that when adult learners perceive that someone cares about their education, they are more likely to persist in their degree programs (Ritter-Williams & Rouse, 2011). One possible explanation for this counterintuitive finding is that adult learners who receive strong external support may also be facing complex life circumstances—such as work–family conflicts or health issues—that prompt them to temporarily withdraw, often with reassurance or approval from their support system. Additionally, for adult learners in their 30s, parental involvement may inadvertently conflict with their developmental need for autonomy and self-direction, potentially leading to tension or resistance that manifests as increased drop-out intention. GPA (β = −0.201, p < 0.05) also significantly predicted lower drop-out intention, reinforcing the importance of academic performance in adult student retention.

Other variables, such as faculty contact, scholarship status, and utility of college education, displayed moderate effect sizes but did not reach statistical significance. These findings highlight that while academic engagement and environmental pressures are relevant, the most impactful predictors center around academic behaviors (study habits and credit load) and perceived support structures. Overall, the results underscore the importance of fostering academic readiness and providing targeted advising and learning support to mitigate drop-out intention among adult learners.

5. Discussion and Conclusions

5.1. Summary of Findings

This study analyzed the factors affecting college persistence of adult learners. Specifically, this study investigated the factors influencing two types of student departures among adult learners: stop-out (temporary withdrawal with intent to return) and drop-out (permanent departure without degree completion). The analysis incorporated a broad set of predictors spanning personal, academic, institutional, social, environmental, and psychological domains based on Bean & Metzner’s student attrition model (1985). While some predictors were common to both outcomes, their effect sizes and significance varied, suggesting meaningful differences between temporary and permanent forms of disengagement.

5.2. Discussions for Both Stop-Out and Drop-Out

5.2.1. Academic Preparedness and Enrollment Status as a Common Factor

First, for both stop-out and drop-out intention, the regression analysis revealed that academic preparedness and enrollment status were the common and most important factors for persistence of adult learner’s learning. Across both models, study skills emerged as a significant negative predictor of both stop-out (β = −0.225, p < 0.01) and drop-out intention (β = −0.210, p < 0.01). According to previous research, most adult learners have trouble with basic study skills and knowledge for higher education. Therefore, effective study strategies are one of the critical factors predicting greater academic self-efficacy and planning abilities, which reduce academic stress and drop-out risk. This is supported by literature emphasizing that academic skills—particularly in self-regulation—mediate drop-out intentions through enhanced self-efficacy and integration (Nemtcan et al., 2020).

Similarly, study hours were negatively associated with drop out (β = −0.239, p < 0.01), and the number of credits registered predicted both lower stop-out and drop-out intentions. This implies that learners who are academically active—taking on more coursework—are more engaged and likely to persist. These findings are consistent with studies showing that academic momentum, such as accumulating credits steadily, and academic self-regulation and behavioral engagement are critical to adult learner persistence (M. Bergman et al., 2014b; Kasworm, 2003; Wlodkowski, 2003). In sum, these findings align with prior research. This suggests that adult learners who are better prepared academically are less likely to withdraw, whether temporarily or permanently.

5.2.2. Role of Academic Advising and Institutional Support

In addition, the quality of academic advising and institutional support have an important influence on the persistence of adult learners. The quality of academic advising significantly reduced stop-out intention (β = −0.206, p < 0.05) but had no significant effect on drop-out intention. This implies that strong and supportive advising may help adult learners stay engaged during difficult periods or consider temporary leave rather than a permanent exit. This supports literature emphasizing the importance of advising as a retention strategy for adult learners (Donaldson & Graham, 1999; Ross-Gordon, 2011). Prior research indicated that advising relationships are especially critical for adult learners who may lack access to broader institutional support structures (Ross-Gordon, 2011). Previous research also suggested that advising quality significantly affects adult learner retention by clarifying educational pathways and enhancing institutional trust (Preuss et al., 2023). This finding implies that reinforcing proactive and personalized academic support is essential, particularly for adult learners who experience difficulties in sustaining their studies.

5.2.3. Unique Patterns in Student Attrition Among Adult Learners

Outside encouragement (parent’s care) was unexpectedly positively associated with drop-out intention (β = 0.179, p < 0.01). While previous studies suggest that social support promotes persistence (Ritter-Williams & Rouse, 2011), this finding may indicate that external encouragement from parents—especially for adult learners in their 30s—may conflict with their need for autonomy and agency (Bye et al., 2007). Alternatively, it may reflect supportive exit scenarios, where learners receive encouragement to leave temporarily or permanently to manage external responsibilities (Taniguchi & Kaufman, 2005).

A particularly noteworthy finding was the positive association between certainty about remaining in the current university and stop-out intention (β = 0.215, p< 0.05). This counterintuitive result may suggest that learners who are highly certain about returning to the same institution feel safer pausing their studies temporarily—viewing a stop-out as a planned intermission rather than academic failure. Alternatively, it might reflect an internal conflict masked by outward expressions of commitment, a phenomenon needing deeper qualitative exploration (Bulotaitė et al., 2025). Bean and Metzner’s (1985) model of nontraditional attrition supports this view, highlighting the importance of environmental factors and life-course flexibility in adult learners’ academic decisions.

5.2.4. Limited Influence of Social Integration and Psychological Outcomes

Academic engagement (defined here in terms of affective or motivational engagement), social integration (e.g., faculty or peer contact and campus involvement), and psychological outcomes (e.g., satisfaction and perceived utility of college) had no significant effect on either stop-out or drop-out intentions. Social integration factors such as faculty and peer contact, scholarships, and membership in campus organizations, especially, were not statistically significant in either model.

According to previous literature about traditional-aged students’ retention, academic and social engagement, along with satisfaction, are one of the most powerful predictors of retention (Tinto, 1993). However, since adult learners often have more practical and well-defined goals for attending college (Bye et al., 2007; Fuhrmann, 1997), it can be inferred that social engagement and psychological status may be less relevant. As previous studies have noted, adult learners often pursue higher education with clearly defined and practical goals (Bye et al., 2007; Taniguchi & Kaufman, 2005). Therefore, it is reasonable to infer that social engagement and psychological satisfaction, while valuable for traditional students, may not play a central role in the retention decisions of adult learners. Instead, task-relevant academic variables and the perceived instrumental value of coursework may be more critical to sustaining motivation and enrollment among this population.

6. Conclusions

6.1. Theoretical Implications and Suggestions for Furture Studeis

Some theoretical implications could be suggested based on the results of this study. First, it is required to develop a student attrition model for adult learners in college, considering their characteristics. In most cases, adult learners go to college for practical and clear purposes compared with traditionally aged students. Therefore, it could be referred to that the quality of the curriculum and lectures and the quality of academic advising are more important than social integration on campus. In addition, adult learners usually take courses while they have a regular job or take care of dependents, so they have limited time and interest for socializing on campus. Some of the literature argued that the adult learner’s motivation of postsecondary education could be explained by pursuing credentials (Gardner et al., 2022). Therefore, more research is required to figure out the retention factors of adult learners, considering the characteristics of adult learners and their motivation.

Second, future studies need to measure ‘drop-out decisions’, rather than drop-out intentions. It is well known that drop-out intention is the most powerful predictor of drop-out decision (Sheeran, 2002; Tinto, 1993). Therefore, it is meaningful to measure drop-out intention because it is possible to prevent drop-out behaviors before the students decide to leave college. However, for developing a more precise student attrition model, future studies need to include actual drop-out decisions.

6.2. Practical Implications for Universities

For practical implications, the findings of this study point to the need for institutions to rethink retention strategies for adult learners. Traditional models that focus heavily on social integration and campus engagement may be insufficient. Understanding the distinct motivational and behavioral profiles of adult learners is essential for designing policies and interventions that reduce both temporary stop-outs and permanent drop-outs. Universities that offer curricula for adult learners should provide targeted support to facilitate their learning.

For example, institutions should invest in strengthening academic advising that is responsive to adult learner needs, offering flexible course scheduling and credit pacing options and enhancing access to study skills and academic support resources. Especially since many adult learners are motivated and have goal-oriented purposes for higher education degree. However, many of them face academic challenges because they have often been away from formal education for an extended period (Lee & Hong, 2022). Traditional student support services, such as learning communities or peer mentoring programs, typically require additional time commitments and extended on-campus presence. However, adult learners often lack the time to participate in such programs due to work, family responsibilities, or other external commitments (Urban & Jirsáková, 2022).

To address this issue, it is crucial to develop and implement learning assistance programs specifically designed for adult learners. These programs should accommodate their unique needs, offering flexible and practical strategies for academic success (Bellare et al., 2023). For example, a tailored learning support program could include workshops on “How to Manage Reading Assignments Efficiently While Caring for Dependents” or provide resources on effective time management strategies. In addition, it is required to establish an educational program from the perspective of adult learners (Moges et al., 2023). Such initiatives would help adult learners balance their educational pursuits with their personal and professional responsibilities, ultimately enhancing their academic persistence and reducing permanent drop-out rates.

Most importantly, practitioners should recognize that stop-out behavior does not necessarily indicate failure but may instead represent a strategic response to life complexities. Differentiating between temporary withdrawal and permanent drop-out is essential for designing effective interventions that support eventual re-engagement. Additionally, institutions should consider offering flexible enrollment options that enable adult learners to reduce or temporarily pause their course loads without stigma or penalty.

These findings also have implications beyond the Korean context. Many countries around the world, including Japan, Germany, Italy, and several Northern and Eastern European nations, are experiencing similar demographic changes—namely, declining traditional-aged student populations and an increasing demand for lifelong learning due to rapid technological shifts and labor market volatility (OECD, 2021). As a result, adult learners have become a strategic demographic for higher education institutions globally.

Accordingly, universities in other aging or skill-transforming societies can benefit from rethinking their retention strategies by prioritizing academic readiness, flexible learning structures, and advice tailored to adult students. These approaches align with the global policy trend of promoting inclusive, lifelong access to higher education as a means of addressing both personal development and national workforce needs (International Commission on the Futures of Education, 2021).

6.3. Suggestions for Future Study

This study employed stop-out intention and drop-out intention as proxies for actual student withdrawal behavior. This approach has practical value, as intention is known to be a strong predictor of future behavior (Ajzen, 1991; Tinto, 1993), and it allows institutions to identify students who may be at risk before they formally leave. By capturing students’ psychological readiness to depart, intention-based data offer valuable insights for early intervention.

However, intentions do not always translate into action. Students may plan to leave but remain enrolled due to external constraints (e.g., financial and familial) or may unexpectedly drop out despite reporting low intention. This gap between intention and behavior suggests that relying solely on intention can limit the precision of predictive models. To address this, future studies should complement intention-based measures with longitudinal tracking of actual drop-out behavior, using institutional enrollment data across semesters. Mixed-method designs that combine surveys, interviews, and administrative data could offer a more robust and nuanced understanding of adult learner persistence.