A Path Analysis Attitude Model Approach for Degree Persistence Across Student Populations

Brad Kleindl, Ph. D. Posted on
Print Friendly, PDF & Email

DOI: https://doi.org/10.64010/MSVI8391

Abstract

This quantitative exploratory study identifies a retention model based on Ajzen’s (1991, 2001) Theory of Planned Behavior, along with other traditional retention related measures, to develop a comprehensive model of student retention intent across two students populations split by age. Student retention and graduation rates are a major concerns for both universities and society at large. Theories on student retention have used factors such as psychosocial and traditional predictors of high school GPA and rank and ACT/SAT scores. New emerging theories have developed that look at student’s motivations and attitudes impact on retention. For this study, survey data from business students at two universities was merged with student database records. A path analysis model is used to identify relationships between model variables. Pooled results across two different age groups (<27 and =>27) show strong support for the model. Student’s academic attitudes are a stronger predictor of retention intent. Universities can use these attitude measures to identify at risk students.

Introduction

Whether measured as five- or six-year degree attainment or by persistence to the sophomore year of college, student retention continues to be an important measure of the effectiveness of higher education around the world. This issue takes on even greater importance as global competition continues to rise and higher education budgets come under increasing pressure. By 2020, 64% of U.S. jobs are expected to require some postsecondary education (Carnevale, et al., 2013). The U.S. Bureau of Labor Statistics projects that into 2024, occupations that require a bachelor’s degree are projected to grow by 8.2 percent (BLS, 2017). Student attrition is very costly for both the student, society, and educational institutions (Hoverstad, et al., 2001).

Colleges and universities are providing greater access to their institutions serving both traditional and non-traditional students populations. Student retention rates, measured in terms of degree attainment, have remained flat across institution type since the 1970s (Adelman 2004; Barton 2002). Lowest persistence-to-degree rates are two year public programs at 22.1% and BA/BS public programs at 36.9% com-parted to high BA/BS at private programs at 58.1% and PhD private programs at 63.1% (ACT, 2018). The National Center for Educational Statistics tracks completion rates for U.S. institutions. They measure 4 to 6 year (150%) completion rates for students who are first-time, full-time bachelor’s degree-seeking students at 4-year postsecondary institutions. Graduation rates by student profile and institution can vary from a low of 8.6% for black females at for-profit institutions for 4 years to a high of 95.7% for females at highly selective non-profit institutions at the 6 year rate (NCES, 2016). There is no shortage of empirical evidence (Wirt, et al., 2004; ACT, 2016) showing that retention rates in U.S. higher education offer much opportunity for improvement.

Student success, typically considered as degree completion obtained through persistence year-to-year, is viewed as an important measure of the effectiveness of higher education around the world. In the U.S., both Federal and State governments have recognized the need for improvement of postsecondary degree attainment and have adopted a broad range of goals and tactics to address the low completion rates (Collins, 2006). There is also a recognition that adults need degree attainment as well. Economies cannot successfully compete in today’s global and technology-based economy without an increase in the proportion of adults who hold a postsecondary degree or credential (Hoffman & Reindl, 2011). There is still little consensus on the factors that best predict student retention.

Higher education is caught between two competing social goals. The first is the need to provide a college education to a diverse population. This includes not only ethnic diversity but also a varying level of academic skills, college preparation, and commitment. The second goal is to help insure that a broad spectrum of students graduate without lowering the quality of education. Education administrators and faculty are faced with the challenge of finding variables that are managerially actionable in influencing student retention. Although numerous studies have investigated factors related to retention, these typically include traditional predictors (TPs) (HS GPA/rank, ACT/SAT) scores and psychosocial and study skills factors (PSFs). Traditional predictors have been proposed to impact graduation rates (Astin & Oseguera 2005; Robbins, et al., 2004) but are not actionable post admissions. Social factors such as family financial support and perceived social support can be beyond the control of the academic institution. Some psychological factors such as achievement motivation, need-to-belong, self-efficacy, self-worth, and self-concept can be seen as individual-difference characteristics (personality variables). These individual personality characteristics are also not actionable in the short term. They can, however, be used as identifiers of at-risk students. This leaves a limited set of actionable variables. Institutions can implement higher levels of selectivity, using TPs and psychosocial variables, to select only those students who have a high potential for graduation. In fact, institutional selectivity has been shown to be highly correlated to retention (Robbins et. al. 2004). However, this violates the social goal of a broad spectrum of students graduating as outlined above. This exploratory study links existing literature on student retention to Ajzen’s (1991, 2001) Theory of Planned Behavior model in an attempt to model a variety of variables that influence a student’s retention intent.

Literature Review

Measuring Determinants of Retention

Research into the factors that impact student retention dates back to at least the 1950s. Two early theoretical perspectives emerged, those of Tinto (1973, 1975) and Bean (1980, 1981). These theoretical works formed the basis for a meta-analysis of the factors that best predict higher education student outcomes (Robbins, et al., 2004). A meta-analysis is a study of studies. It combines the analytical power of multiple quantitative studies to produce generalizations regarding the relationships between study variables. The method provides correlation estimates based on the results of multiple studies. In the case of Robbins, et al. (2004), 408 studies were identified and of these, 109 contained usable data. Table 1 presents the study’s results as they pertain to the factors that relate to student retention. In the meta-analysis, retention statistics are taken from university records and are assumed to have a high level of reliability.

Table 1: Meta-Analysis Results

Psychosocial and study skill factors: Correlation Estimate 
Academic-related skills: .366 
Academic self-efficacy: .359 
Academic goals .340 
Institutional commitment: .262 
Social support: .257 
Institutional selectivity: .238 
Social involvement: .216 
Financial support: .188 
Achievement motivation .066 
General self-concept: .050 
Institutional size: -.010 
Traditional Predictors: 
High school GPA: .239 
Socioeconomic status: .212 
ACT/SAT scores: .121 

Robbins, et al. (2004) Meta-Analysis

Traditional predictors (i.e., socioeconomic status, high school GPA, ACT/SAT scores) are important variables and do contribute to the variability in retention as shown in Table 1, but they have lower correlation estimates than academic-related skills, academic self-efficacy, and academic goals. Since these factors are often used in the admissions screening process, it is not surprising that institutional selectivity also has a significant relationship to retention. Adult learners are seen as facing even higher barriers toward persistence to completion. These include socioeconomic status, race/ethnicity, age, gender, marital status, the number of children, prior college credit, and goal commitment (Bergman et al., 2014).

Theory Development in Student Retention

Tinto’s Student Departure Theory began in 1973 in collaboration with Cullen. This theory considered six factors.

  • Pre-entry attributes (prior schooling and family background).
  • Goals/commitment (student aspirations and institutional goals).
  • Institutional experiences (academics, faculty interaction, co-curricular involvement, and peer group interaction).
  • Integration (academic and social).
  • Goals/commitment (intentions and external commitments).
  • Outcome (departure decision— graduate, transfer, dropout).

Another researcher, John Bean, develop a causal model

(Bean, 1981) with five measures.

  • Student background variables.
  • Interaction by students within the institution.
  • Influence of environmental variables (finances, family support).
  • Attitudinal variables (a subjective evaluation of perceived quality and self-satisfaction with the institution).
  • Student intention, such as transfer and degree attainment.

Problems with Early Theory Base

While these models have evolved over time, the early theory development was based on college population in the 1970’s. These were traditionally aged college students with a four year college focus. Student populations, such as community colleges, were at first dismissed because these students didn’t fit the socially integrated model. Other neglected groups in the early research were minorities and non-traditionally aged students (Metz, 2005). Tinto (Tinto & Pusser, 2006) have updated this conceptualization of a model for student retention where external factors, which include abilities, skills, preparation, attributes, and attitude influence the educational process. Tinto identifies four components that are actionable by institutions.

  • Student expectations, influenced by orientation and advising.
  • Student support, influenced by academic, social, and financial support factors.
  • Student feedback, influenced by monitoring, assessment, and early warning systems.
  • ·tudent involvement, influenced by extracurricular activities, curriculum pedagogy, and faculty development.

These items have been practiced by universities in their educational models and they should lead to a quality of effort, learning, and then student success. However, persistence levels to graduation have not increased substantially over time.

Emerging Theories

Savage & Smith (2008) used Hope Theory with a student sample from the Community College of the Air Force. Hope Theory is defined as the ability to generate strategies or cognitions toward one’s goals and the motivation and self-efficacy to carry out those strategies (Browning et al., 2018). Savage’s study use a sampling that included non-traditional, non-transfer student population (average age 39.7). Savage found that Hope, as one component of overall goal orientation, was statistically significant predictor of degree attainment. Other researchers have identified that student’s motivation and goals are linked to student retention (Covington, 2000; Alarcon & Edwards, 2013; Morrow & Ackermann, 2012).

  • Attitude Theory as an Organizing Principle

Attitude theories are a foundation of business consumer research. An attitude is an overall evaluation of an object. The Tri-Component Attitude Model consist of three components: cognitions (thoughts), affect (emotions), and conative (behavior intentions). Attitude objects are considered to be broadly construed and can consist of a number of psychological objects including: products, product categories, services, or an attitude toward a behavior. There are a number of advantages in using attitude theory. Attitude theory has been used and validated across product categories, situations, and consumer groups. Attitude theory has strong predictive power related to both intention and future actions. For academic administrators, the ability to tap intention to behave can be very important in intervening in students’ intention to persist. Some aspects of attitude theories have been linked to student persistence. These include cognitions as student satisfaction and dissatisfaction linked to student retention (Mashburn, 2001). Affective commitment, trust, and student loyalty can be significant predictors of university retention (Moore, et al., 2012), (Alarcon & Edwards, 2012). Conative (behavioral) items such as hours earned have been shown to be related to retention (Metz, 2005).

Attitude theory also allows for a theoretical integration of isolated measures. For example, most retention studies use actual retention (1: retained 0: dropped) as a dependent variable. These studies have found relationships between isolated measures and retention. When measured in isolation, R2s may be low because of all the other unmeasured factors related to retention. However, when all of these measures are included in a single regression model, there is likely to be a large amount of collinearity between the various independent variables (as with HS rank, ACT scores, etc.). Attitude theory posits that intention precedes actual behavior. The advantage of measuring intention is that it may be impacted by academic action.

A theoretical underpinning of intention is outlined in the Theory of Planned Behavior (TPB). This is an attitude model based on the premise that behavior is influenced by behavioral intentions that are in turn influenced by an individual’s attitude toward the behavior, subjective norms related to the performance of the behavior, and the individual’s perception of the ease with which the behavior can be performed or controlled (Ajzen 2001). Ajzen’s TPB Model posits that there are interactions between each attitudes construct (cognitions, affect, and conative {behavioral}), which in turn are moderated by subjective norms, and perceived behavioral control, all of which influence behavioral intentions. Ajzen has noted that behavioral control (the belief that people have the volitional control of performance of a behavior) can be linked to other psychometric measures such as self-efficacy. The overall model design used in this study is based on the TPB Model.

Theory of Planned Behavior for Study

The TPB Model allows for analyzing an influence chain for student retention intent. This process moves from pre-admissions factors, through socio-economic support (subjective norms), student individual confidence (perceived behavioral control), then through academic attitudes, and then an impact on retention intent (behavioral intention). As describe in Ajzen’s TPB Model, there can be interactions between each of these constructs and the variables within the constructs. This can lead to problems in both developing measures and testing measures. For this exploratory study, relationships between antecedent constructs and items in the dependent constructs will be measured. This will lead to a path analysis showing a chain of influences between the theoretical constructs. The relationship between these variables and traditional measures used in retention studies and the methodology used to create the variables are described in Figure 1 below. The model shows four major constructs, traditional predictors, socio-economic support (including university support), individual confidence, and academic attitudes. Preceding constructs variables are posited to have an influence on the each of the following constructs variables. The Academic Attitude Construct variables are posited as having the direct influence on the dependent variable of Retention Intent.

Figure 1: Model Constructs and Path

Traditional Predictors Construct

The traditional predictors of college success include high school GPAs, standardized college readiness tests comprehensive scores, comprehensive exam math sub-scores, and high school rank. ACT comprehensive scores (used in this study of Midwestern universities) and high school rank were included in the study. These variables were pulled from the traditional university’s student database and matched to individual student survey results. The traditional predictors of ACT scores and high school rank were included in the study only for the traditional university. The non-traditional university does not collect this data for admissions. For this study, the ACT score and high school rank are posited as an indications of a student’s ability to have perceived behavioral control over their ability to persist in college and therefore influence the individual confidence construct.

Socio-Economic Support Construct

Both Tino (1993) and Bean (1981) have recognized that family influence and family support can be related to student intentions. The education level of students has also been recognized as an influence on student intentions (Thayer, 2000). For this study, socio-economic support variables were designed to capture the family’s education level (highest degree held by either parent), perceived levels of family support, family influence, and friends influence. The influence variables were framed in terms of career goals and obtaining a degree. The family support variable was framed as supporting the pursuit of a degree. The Socio-Economic Support construct is posited as an antecedent to a student’s individual confidence and goal setting.

University support is conceptualized as those items that are under the control of the academic institution both at the classroom level and those items that influence a student’s educational and areer goals (Tino & Pusser, 2006).

Individual Confidence Construct

Ajzen’s behavior control concept (the belief that people have the volitional control of performance of a behavior) can be related to achievement goal theory. This is a social-cognitive approach to motivation within the motives-as-goals tradition (Dweck, 1986). It is usually construed in retention literature as one’s motivation to achieve success, enjoyment of surmounting obstacles, and completing tasks undertaken, and the drive to strive for success and excellence (Robbins et al., 2004).

Individual confidence is conceptualized to consist of three variables: self-efficacy, a student’s goal orientation, and degree goals. Academic self-efficacy has been recognized in Robbins’ (2004) meta-analysis as the second strongest influence on student retention. Self-efficacy items were based on self-efficacy literature and resulted in seven items with a Cronbach Alpha of .894 (Kleindl & Cragin, 2006).

As with behavior control in an attitude model, achievement goal theory postulates a causal relationship between an individual’s goal orientation and their behavioral responses. This relationship is mediated by individual cognitions (Covington, 2000), or the cognitive aspect of an attitude mode. As originally presented, the achievement goal model had two classes of goals: performance goals (AKA ego, self-enhancing, ability) and learning goals (AKA, mastery, task), Performance goals relate to an individual’s desire to gain favorable judgments at the expense of others. For this study, this is defined as Goal Orientation. Learning goals refer to an individual’s desire to increase their competence, understanding, and appreciation for what is being learned. For this study, this is defined as Degree Goals.

Research in achievement goal theory has demonstrated that learning goals favor an emphasis on deep understanding and the strategic processing of obtaining information which facilitates persistence and mastery-oriented behaviors even in the face of set-backs or low perceived ability (Grant & Dweck, 2003; Covington, 2000). Goals should relate positively to both school achievement and retention. Robbins, et al. (2004) were unable to examine the motivational constructs of performance and learning goals in their meta-analysis because there were not enough studies found, but they did note that “researchers need to test the role of performance and mastery [i.e., learning] goals within the college adjustment process” (p. 276). The measure of behavior control/learning goal orientation in this study is adapted from Roedel, et al. (1994).

Academic Attitude Construct

Attitude concepts are beginning to be integrated into retention theories (Mashburn, 2001; Moore, et al., 2012; Alarcon & Edwards, 2012; Metz, 2005; Morrow, 2012). The Academic Attitude Construct in this study uses the tri-component theory items: cognitive, affective, and conative. The affective and cognitive components of the TPB focus on the actions that are controllable by academic administrators. The cognitive component consists of two variables, the perceived academic reputation of the academic institution and programs and the likelihood to recommend the academic institution and programs (satisfaction).

In the current study, affective variables are framed in a negative valence. This is due to the fact that negative information can have a stronger impact on evaluations than positive information (Ajzen, 2001). Attitude measures can be compared to the institutional commitment variables used in other retention studies. Institutional commitment is usually defined as the students’ confidence in and satisfaction with their institutional choice and the extent that they feel committed to the college in which they are currently enrolled (Robbins et al., 2004).

For this study, six items were used to measure negative affect and seven items were used for academic reputation. When these variables were analyzed using factor analysis (Varimax rotation), the thirteen items separated into two unique factors. Cronbach’s Alphas also show strong internal reliability. These variables were validated against a single item measure of satisfaction with the school programs. Academic reputation had a correlation of .783 (p<.000) with satisfaction. Negative affect had a correlation of -.703 (p<.000) with satisfaction. The conative (behavioral) component was a measure of the student’s earned hours. The more earned hours a student has, the more likely they are to be committed to persist towards graduation.

Retention Intent (Model Dependent Variable)

Intent precedes actual behavior. Measuring the factors leading to intent can allow for measuring intentions rather than waiting until the actual behavior (leaving college) occurs. The measure of persistence intent in this study is similar to the academic goals construct from the Robbins, et al. (2004) meta-analysis. Retention Intent consists of three items. These items were designed to measure the reasons that an individual may leave their current institution. These relate to switching to another university, dropping before graduating, and getting a job rather than graduating. The Cronbach’s Alpha for these three items was .713, acceptable for exploratory research.

Method

Data Collection:

An online survey system was used to collect data from student populations at two universities. One university served a primarily traditional (80% under the age of 27) undergraduate students. This university sample was 67% while and 56% female. The other university served a primarily non-traditional (82% at or over the age of 27) undergraduate students. This university sample was 46% while and 58% female. Students could volunteer their student ID numbers so their results could be matched with TPs and other data in the university database. A total of 804 usable surveys were obtained, 404 from university serving mostly traditional students and 400 from university serving mostly non-traditional students.

Scale Items

Factor analysis were run on the variables within each construct, not across constructs due to the likelihood of strong item inter-correlations. Cronbach Alphas were run on the identified variables. The survey item results are shown in Table2. Results shown variables have alphas of at least .7, considered to be acceptable for internal reliability. A six point Likert Scale with Strongly Disagree (1) to Strongly Agree (6) was used. For Family Education Level, the level of education was used as a metric variable (High School:

1, Associate: 2, Bachelors: 3, Graduate: 4).

Results

Data Analysis

Stepwise regressions were run separately between model construct items against each of the following construct variables used as dependent variables (individual variables treated as separate DVs). Stepwise regression runs multiple regression a number of times, keeping the strongest correlated IVs and dropping the weakest correlated IV. At the end of the procedure the process leaves the variables that best explain the model.

Data was analyzed using a path analysis technique. Path analysis allows for simultaneous analysis across complex models (Jeon, 2015). Path analysis creates causal models by examining the relationships between two or more independent variables (IV) and a dependent variable (DV). Path coefficient models are developed using standardized betas (β) from stepwise regression. The standardized regression beta coefficients allow for identifying the direct effect of an independent variable on a dependent variable and standardized betas can be compared across regression models. This allows for a comparison of the individual IV items through a complex model to the final model DV, Retention Intent. The path analysis technique allows for following the beta coefficients across a complex model. This also allows for comparisons across population models. This process helps to eliminate the collinearity issues within complex models. For this study, path models were was split on age data between traditionally age students (<27) and non-traditionally age students (>=27).

Figure 2: Path Model Standardized Regression Β s Models Age < 27

Figures 2 shows results for the less than 27 age group. Significance is indicated with <. 000 as *** and < .05 with **. Each construct, Socio-Economic Support, Individual Confidence, and Academic Attitudes consists of a number of variables. Only those paths that have significant betas (β) are included. If no significant relationship exists, such as between Friends Influence and Goal Orientation, then no path is indicated.

For the under 27 age population, the Traditional Predictors were included in the same stepwise regression model with the Socio-Economic variables because of the posited influence on the Self-Confidence Construct. For example, Family Support (β .288, p<.000), University Support (β .334, p<.000), Family Education Level (β .146, p<.001) along with the students High School Rank (β .161, p<.000) emerge from the stepwise regression to show a direct influence on Self-Efficacy (R2: .232, p<.000). Family Support (β .166, p<.000) and University Support (β .384, p<.000) also show a direct influence on Goal Orientation (R2: .197, p<.000). University Support (β .384, p<.000) and Family Influence (β .384, p<.000) show a direct, but weak, influence on Degree Goals (R2: .085, p<.000). These results show the relationships between the Socio-Economic Support and Individual Confidence constructs. Family Support influence, Family Education Level, and University Support all have an influence on a student’s Self-Efficacy and the student’s Goal Orientation. The IV of Friend’s Influence has no significant direct influence on the Individual Confidence variables.

Individual Confidence constructs have nomological relationships to Negative Affect and Academic Dread. Self-Efficacy has the strongest relationships with the Academic Attitudes Construct. Self-Efficacy (β -.292, p<.000) is negatively related to Negative Affect (R2: .060, p<.000). Self-Efficacy (β -.335, p<.000) and Degree Goals (β -.173, p<.000) are negatively related to Academic Dread (R2: .181, p<.000). The negative relationship with negatively valanced variables indicates that the stronger the Self-Confidence Construct IVs, the weaker the students’ Negative Affect and Academic Dread. In turn, Negative Affect (β .395, p<.000) and Academic Dread (β .241, p<.000) show a direct positive influence on Retention Intent (negatively valenced) (R2: of .282, p<.000). This implies, that for the traditionally aged population Negative Affect and Academic Dread account for over 28% of a student’s intent to remain to leave the institution. Implications for the traditionally aged student population for this model it that the student’s past high school academic experience (rank), a number of Socio-Economic Support variables, and University Support have an influence on the student’s Self-Efficacy, which in turn, can lower a student’s Negative Affect and Academic Dread leading to greater retention.

Figure 3: Path Model Standardized Regression Β s Models Age >=27

For the non-traditional student population, Family Support (β .430, p<.000) and University Support (β .347, p<.000), along with the Family Influence variable (β -.113, p<.028) emerge from the stepwise regression to show a direct influence on Self Efficacy (R2: .330, p<.000). The Family Influence variable (family has strongly influenced my career goals and family has strongly influenced my decision to obtain a college degree) has a negative relationship with Self Efficacy. It is possible that individuals with a lower self-efficacy rely upon family influences for support. University Support has a positive relationship with a student’s Self-Efficacy and Goal Orientation, but not degree goals. It is likely that these non-traditional students have returned to the institution to pursue self-identified goals.

All three Individual Confidence variables influence Academic Attitude variables. Self-Efficacy (β .325, p<.000) influences Recommendation (R2: .168, p<.000). Self-Efficacy (β -.369, p<.000) and Degree Goals (β -.188, p<.000) are negatively related to Negative Affect (R2: .205, p<.000). Self-Efficacy (β -.386, p<.000) and Degree Goals (β -.199, p<.000) are negatively related to Academic Dread (R2: .227, p<.000). Academic Attitude constructs have nomological relationships to the model DV of Retention Intent. The stronger a student is to Recommend the university (β .-135, p<.024), the less likely they are to leave. Students with stronger Negative Affect (β .319, p<.000) and stronger Academic Dread (β .329, p<.000) are more likely to indicate they would leave. The more Earned Hours (β -.131, p<.000) the less likely a student is to leave. These four variables have a model R2 of .449 (p<.000) on Retention Intent. However; the affect variables of Negative Affect and Academic Dread accounting for an R2 of .424 (p<.000), or a majority of the influence on the DV. Implications for the non-traditionally aged student population for this model is that the student’s family and the university have an influence on the student’s Self-Efficacy, which in turn, along with the students strong degree goals, can lower a student’s Negative Affect and Academic Dread leading to greater retention. Figure 6 below overlays both path models and an analysis will be related back to the literature with implications.

When comparing across student populations, patterns emerge. Self-Efficacy is the intermediate variable in the path model. Self-Efficacy has the strongest standardized beta linkages to the Academic Attitude variables. Self-Efficacy, in turn, can be influenced through family Socio-Economic Support and by University Support. Degree goals also influence Academic Attitudes and these Degree Goals are influences by family and the university, but only for traditionally aged students. The Academic Reputation of the institution has no influence on Retention Intent for either population. The non-traditional students factor more Academic Attitude Construct items in their Retention Intent, including the likelihood to Recommend and Earned Hours.

These findings are supported by the literature. Alar-con & Edwards (2013) found that increased negativity leads to a decreased probability of finishing school and first-generation students tend to have lower retention rates than non-first-generation students. Baier’s (2014) dissertation study found no relationship between SAT and persistence and that self-efficacy is the most important predictor for intent to persist. Bergman et. al. (2014) found that entry characteristics such as high school class rank, standardized test scores, college prep curriculum, and high school friends are included in studies of traditional students, but are less relevant to adult students’ persistence.

Discussion and Future Research

There are a number of factors that have an influence on student retention. This study shows a causal path that links TPs, Socio-Economic Support, Individual Confidence, Academic Attitudes, and Retention Intent. These paths differ by age cohort. Negative Affect is the strongest driver of Retention Intent. Self-Efficacy has both most paths and the strongest linkages to Academic Attitudes. An academic institution’s support can have an influence on the student’s Self-Efficacy. Academics should realize that different student age groups follow different paths toward their retention intent.

The path analysis models show logical consistency with both past retention theory and new emerging theories. HS Rank does impact a subject’s Self Efficacy for those under 27. Socio-Economic Support items influence Individual Confidence with Family Educational Level and Family Influence impacting Self Efficacy for those under 27. Family Support influences Self Efficacy and Goal Orientation for both populations. Importantly, the university does influence both Self Efficacy and Degree Goals.

Academic Attitudes play an intermediary role between the student’s individual confidence level and the student’s Retention Intent. For those under 27, the affect items have a direct impact on Retention Intent (R2: .282, p<000). For those over 27, a cognitive attitude item (Recommend) and a conative attitude item (Earned Hours) are included in the model with affect items to have a direct impact on the likelihood to drop (R2: .449, p<000). For those over 27, the standardized β s for affect items were considerably stronger (β >.300) over the one cognitive item (β = -.135) and the conative item (β = -.131). The over 27 student population affect only items account for the majority of the influence on Retention Intent (R2: .424, p<000).

This study has limitation in that only student age splits were used and not other demographic factors such as gender and ethic category. In addition, there are issues with both step-wise regression path models. Structural equation modeling is a more powerful technique to analyze complex models using a combination of factor analysis and multiple regression analysis.

In future research, the approach followed in this study can be used to identify differences in decision paths across student segments such as ethnic background, socioeconomic status, or other measurable items. Linking student’s survey results to student databases can track actual student completion. The strength of the relationship between the Academic Attitude affect items and intention to drop is the strongest predictor. This can be a focus for schools to use to help in identifying possible drop outs. The affect items can be influenced by the student’s Self-Efficacy which in turn can be influenced by the school support.

References

  • ACT. (2018). National Collegiate Retention and Persistence to Degree Rates, 2018. https://www.act.org/content/dam/act/unsecured/doc-uments/MS2807rev1-retention-persistence-2018-07.pdf
  • Adelman, C. (2004). Principal Indicators of Student Academic Histories in Postsecondary Education. 1972–2000. Washington, DC: U.S. Department of Education.
  • Alarcon, G., & Edwards, E. (2013). Ability and Motivation: Assessing Individual Factors That Contribute to University Retention. Journal of Educational Psychology, Vol. 105, No. 1, 129–137.
  • Astin, A., & Oseguera, L. (2005). Degree Attainment Rates at American Colleges and Universities (Revised Edition). Higher Education
  • Research Institute, Graduate School of Education. University of California, Los Angeles.
  • Ajzen, I. (1991). Theory of Planned Behavior. Organizational Behavior and Human Decision Processes. 50, 179-211.
  • Ajzen, I. (2001). Nature and Operations of Attitudes. Annual Review of Psychology. 52, 27-58.
  • Alarcon, G., & Edwards, J. (2013). Ability and Motivation: Assessing Individual Factors That Contribute to University Retention. Journal of Educational Psychology. 105(1), 129–137.
  • Baier, S. (2014). The Role of Academic Factors, Self-Efficacy, Mentoring Relationships, And Learning Communities In Persistence And Academic Success Of Freshmen College Students. Wayne State University Dissertations. Paper 870.
  • Barton, P. (2002). The Closing of the Education Frontier? Princeton, NJ: Educational Testing Service, Policy Information Center.
  • Bean, J. (1980). Dropouts and Turnover: The Synthesis and Test of a Causal Model of Student Attrition. Research in Higher Education, 12(2), 155-187.
  • Bean, J. (1981). The synthesis of a theoretical model of student attrition. American Education Research Association, Los Angeles, CA.
  • Bergman. M., Gross, J., Berry, M., & Shuck, B. (2014). If Life Happened but a Degree Didn’t: Examining Factors That Impact Adult Student Persistence. The Journal of Continuing Higher Education, 62, 90–101.
  • Bureau of Labor Statistics. (2017, June 28) 37 percent of May 2016 employment in occupations typically requiring postsecondary education. https://www.bls.gov/opub/ted/2017/37-percent-of-may-2016-employment-in-occupations-typically-requiring-postsec-ondary-education.htm.
  • Browning, B., McDermott, R., Scaffa, M., Booth, N., & Carr, N. (2018). Character Strengths and First-Year College Students’ Academic Persistence Attitudes: An Integrative Model. The Counseling Psychologist, 46(5) 608–631.
  • Carnevale, A., Smith, N., & Strohl, J. (2013). Recovery: Job Growth and Education Requirements Through 2020. Georgetown University, Center on Education and the Workforce, McCout School of Public Policy.
  • Covington, M. (2000). Goal Theory, Motivation, and School Achievement: An Integrative Review. Annual Review of Psychology, 51, 171-200.
  • Dweck, C. (1986). Motivational Processes Affecting Learning. American Psychologist, 41, 1040–1048.
  • Grant, H., & Dweck, C. (2003). Clarifying Achievement Goals and Their Impact. Journal of Personality and Social Psychology, 85(3), 541-553.
  • Hoffman, L., & Reindl, T. (2011). Improving Postsecondary Attainment Among Adults. NGA Center for Best Practices.
  • Hoverstad, R., Sylvester, R., & Voss, K. (2001). The Expected Monetary Value of a Student: A Model and Example. Journal of Marketing for Higher Education, 10(4), 51-62.
  • Jeon, J. (2015). The Strengths and Limitations of the Statistical Modeling of Complex Social Phenomenon: Focusing on SEM, Path
  • Analysis, or Multiple Regression Models. World Academy of Science, Engineering and Technology, International Journal of Social, Behavioral, Educational, Economic, Business and Industrial Engineering (9)5, 1634-1642.
  • Kleindl, B., & Cragin, S. (2006). Attitude Toward Persistence in Business Majors: Phase One of A Longitudinal Study. National Symposium for Student Retention.
  • Markle, G. (2015). Factors Influencing Persistence Among Nontraditional University Students. Adult Education Quarterly, 65(3) 267-285.
  • Mashburn, A. (2001). A Psychological Process of College Student Dropout. Journal of College Student Retention, 2(3) 173-190.
  • Metz, G. (2005). Challenge And Changes to Tinto’s Persistence Theory: A Historical Review. Journal of College Student Retention, 6(2) 191-207.
  • Moore, D., & Bowden-Everson, L. (2012). An Appealing Connection–The Role of Relationship Marketing in the Attraction and Retention of Students in an Australian Tertiary Context. Asian Social Science, 8(14), 65-80.
  • Morrow, J., & Ackermann, M. (2012). Intention To Persist And Retention Of First-Year Students: The Importance Of Motivation And Sense Of Belonging. College Student Journal, 46(3). 483-491.
  • National Center for Educational Statistics. (2016). Digest of Educational Statistics. https://nces.ed.gov/programs/digest/d16/tables/ dt16_326.10.asp.
  • Robbins, S., Lauver, K., Le, H., Davis, D., Langley, R., & Carlstrom, A. (2004). Do Psychosocial and Study Skill Factors Predict College Outcomes? A Meta-Analysis. Psychological Bulletin, 130(2), 261-288.
  • Roedel, T., Schraw, G., & Plake, B. (1994). Validation of a Measure of Learning and Performance Goal Orientation. Educational and Psychological Measurement, 54(4), 1013-1021.
  • Savage, J., & Smith, A. (2008). General and Specific Goal Orientations as Correlates of Adult Student Degree Completion: Lessons From The Community College Of The Air Force, Journal of College Student Retention, 9(4), 461-485.
  • Tinto, V. (1975). Dropout from Higher Education: A Theoretical Synthesis of Recent Research. Review of Educational Research, 45(1), 89-125.
  • Tinto, V., & Cullen, J. (1973). Dropout in higher education. A review and theoretical synthesis of recent research. New York: Teachers College.
  • Tinto, V., & Pusser, B. (June 2006). Moving from Theory to Action: A Model of Institutional Action for Student Success. National Postsecondary Education Cooperative.
  • Wirt, J., Choy, S., Rooney, P., Provasnik, S., Sen, A., & Tobin, R. (2004). The Condition of Education 2004 (NCES 2004-077). U.S. Department of Education, National Center for Education Statistics. Washington, DC: U.S. Government Printing Office.