The Correlation of Self-Regulation and Motivation With Retention and Attrition in Distance Education

To provide greater access to students and to meet market demands, institutions of higher education have adopted online delivery of instruction and expanding course and program offerings at a rapid pace (Patterson & McFadden, 2009). The growth rate of online enrollments has been substantially greater than overall higher education enrollments (Allen & Seaman, 2007). The 12.9%growth rate for online enrollment is much greater than the 1.2%growth overall of the higher education student population (Patterson & McFadden, 2009). The focus of this study is to explore the correlation of self-regulation and motivation with retention and attrition in distance education.

A quantitative study conducted at a university in the southeastern United States found that online students were much more likely to drop out than the campus-based students. An MBA program and a communication sciences and disorders (CSD) program were evaluated with the finding that 11%of the campus-based students in the MBA program dropped out as compared to 43%of the online students. In the CSD program, 4% of the campus students dropped out compared to 23.5%of the online students (Patterson & McFadden, 2009). At a community college in Texas, there was an 11%to 15%point difference between course completion rates in the online and on-campus course programs (Carr, 2000). Some of the reasons for this are explored in this study as we look at the defining elements of distance education and the effect those elements have on the motivation and regulation of the learning and teaching process. Motivation and self-regulation issues can be significant factors in academic success when students are separated from each other and from the instructors.

Distance education offers significant benefits to institutions desiring to expand student accessibility as well as significant educational challenges. Both the benefits and the challenges have to do with the nature of distance education. Keegan’s definition of distance education states that distance education is characterized by:

• The separation of the teacher and learner throughout the educational process.

• The influence of an educational organization in the planning and preparation of the learning materials along with student support.

• The use of technical media to connect teacher and learner and deliver the content of the course.

• The provision of two-way communication, and the quasi-permanent absence of a learning group so that instruction is usually to an individual (Keegan, 1996).

The work of Moore (1989) has given specific meaning and submeanings to the term “interaction.” He suggested that distance educators agree on the distinctions between three types of interaction, which he labeled learnercontent interaction, learner-instructor interaction, and learner-learner interaction. Students transform when they interact intellectually with the content, when they are stimulated, directed, and motivated to learn, when they can apply what they have learned, and when they are supported in the learning process.

The importance of student-content interaction in the learning process has also been highlighted in several research studies (Bernard et al., 2009; Morrison & Anglin, 2006; Wittrock, 2010). While the face-to-face resident students may have the advantage of instructional strategies given by the instructor in the classroom, this advantage does not always translate to the online environment. The simple posting of information (e.g., PowerPoint slides) is not teaching. Learners must be actively involved in the learning process and generate their own associations between new information and what they already know. It is through this deep processing of information initiated by the instructional strategy that significant learning occurs and students are integrated, challenged, and motived in the learning process. Distance education students do not always have an external stimulus to engage in this type of deep processing with the content and may fail to do so if the interaction is not part of the instructional design.

Moore (1989) states that the lack of interaction between the learner and teacher leaves “ultimate responsibility for maintaining motivation, for interacting with the presentation, for analyzing the success of application, and for diagnosing difficulties on the learners themselves requiring a high degree of learner autonomy” (p. 3). In distance education, as compared to traditional classroom face-to-face education, additional metacognitive, motivational, and behavioral responsibility is placed on the student to plan, organize, self-instruct, self-monitor, motivate, and self-evaluate their learning process.

Many of life’s endeavors require directed and sustained effort over long periods of time with minimal external regulation and direction. In these situations, students must self-regulate and direct their own motivation and action (Bandura & Schunk, 1981). This concept of self-regulation is seen by a number of researchers as a mediator between computerbased learning environments and academic performance (Winters, Greene, & Costich, 2008). Since no environment ensures learning and even settings with advanced levels of interaction do not guarantee learning, students must activate, alter, or create schema for learning to occur. Self-regulated learning theorists view students as metacognitively, motivationally, and behaviorally active participants in their own learning process (Zimmerman, 1986). There is a growing body of laboratory and field research that highlight the importance of students’ use of self-regulated learning strategies in their academic achievement (Zimmerman, 1990).

Circumstantial factors play an increasing role as older, nontraditional students enter online programs to further their education while maintaining work and family responsibilities. In a study at the Indira Gandhi National Open University in India (Fozdar, Kumar, & Kannan, 2006), 250 students who had not completed the program completed a 21-item questionnaire listing reasons in relative importance for dropping out. The main reasons students gave for withdrawal were absence of interaction with fellow students (47.06%), high cost of attending to laboratory work (38.24%), lack of time due to changing family circumstances (35.29%), followed by changes in employment status (35.29%). Other factors found to affect withdrawal included marriage obligations (8.82%) and poor health conditions (8.82%). For these older students, circumstances have a major effect on retention. These factors would apply to U.S. students as well, since the same factors appear in our student culture.

Moore highlighted the importance of interaction in education and defined three types of interaction: student-content (S-C), studentteacher (S-T), and student-student (S-S) (Moore, 1989). A meta-analysis by Bernard et al. (2009) of the experimental literature of distance education compares different types of interaction treatments with other distance education instructional treatments. Interaction treatments are the instructional conditions in the courses that are intended to facilitate the three types of interactions. The study supported the importance of the three types of interaction treatments and the strength of the interaction treatments is found to be associated with increased achievement outcomes. A strong association was found between interactive treatment strength and achievement in asynchronous courses and distance education literature gives overwhelming support to the importance of interaction for achievement.

Anderson (2003) has polled distance education students over the years and has concluded that there is a wide range of need and preference for different combinations of course delivery and activity, which has led to the formation of an equivalency theorem:

Theories of retention and attrition in distance education have highlighted the complex and multifaceted personal, circumstantial, and institutional variables and their varying correlation with success in the online academic environment (Bell, 2006). The common factor in all these variables is students and their reactions and interactions to the personal, circumstantial, and institutional environments they are confronted with as they pursue academic success. Since there is no perfect learner or perfect learning environment, self-regulation of reactions and interactions may be an important key to understanding academic success in distance education.

The research of Bandura and Schunk (1981) on self-motivation, efficacy, and intrinsic interest is a recurring theme in academic success studies especially in distance education where students do not have external factors to motivate them, help them to believe in their adequacy, or organize and implement actions for academic success. As described, the lack of adequate interaction in the educational process requires a high degree of learner autonomy to maintain motivation, interact with the content, diagnose difficulties, and get the necessary support. The research of Zimmerman (1989) on active participation in the learning process is especially important for distance education, which requires students to be able to initiate and self-regulate their own learning. Active participation would be defined as any interaction the student has with the content, the instructor, the environment, and the other students as part of the learning process.

By definition, self-regulated students are metacognitively, motivationally, and behaviorally active in their own learning process (Zimmerman, 1989). In self-regulated learning, the acquisition of knowledge is initiated and directed by the students themselves, and they are less dependent on others for direction and motivation. Metacognitively, the self-regulated student plans, organizes, self-instructs, self-monitors, and self-evaluates at various stages during the learning process. Motivationally, self-regulated students view themselves as competent, self-efficacious, autonomous, and perceive their efforts and outcomes as valuable and worthwhile. Behaviorally, the self-regulated students select strategies, structure their learning, and create a learning environment that optimizes their learning. According to Zimmerman (1986), the self-regulated students are constantly aware of the outcomes of their thinking patterns and their actions.

Three important elements in self-regulated learning are learning strategies, self-efficacy, and commitment to academic goals (Zimmerman, 1989). Learning strategies are actions and processes that are used to acquire information or skill and to improve students’ selfregulation of their personal functioning, behavior, or learning environment. Self-efficacy is the student’s perception of performance skill and capabilities. Academic goals can be grades, social standing, self-esteem, employment, obtaining good grades, improving self-esteem, being equipped for a good job, avoiding failure and embarrassment, and pleasing teachers or parents.

Students’ use of strategies to self-regulate their behavior, environment, and personal (covert) functions in a distance learning environment is dependent on a feedback loop consisting of monitoring performance, selfefficacy, and motivation (Zimmerman & Martinez-Pons, 1990). Current theory and research points to the importance of self-regulation in the distance education environment and the strategies that enhance it (Zimmerman & Pons, 1986).

A study of college-age students involved in an asynchronous web-based course was conducted by Bell (2006) to explore whether factors related to self-regulated learning and epistemological beliefs could predict learning achievement. The study included 201 undergraduate students enrolled in a variety of asynchronous web-based courses at a university in the southeastern United States. Data were collected via a web-based questionnaire and subjected to a factor analysis of self-regulated learning using 24 questions from the Motivated Strategies for Learning Questionnaire (MSLQ). This study’s findings suggest that individuals with the greatest expectancy for learning were the most successful asynchronous learners.

Two key components of the learning and teaching process are motivation and attitude (Morrison, Ross, Kalman, & Kemp, 2013). Morrison et al. (2013) state:

Learner attitude, while it can affect the learning process, is different than motivation. A student may be motivated to learn a subject but may feel inadequate to do so. This lack of confidence will result in less than optimum learning even though the motivation may be present. As we look at self-regulation in the learning process, both motivation and attitude play a large role in the successful regulation of the process. While both students and teachers bring their levels of motivation and attitude into the learning context, the design and conduct of the instructional process can have a direct affect on both motivation and attitude. (Anderman & Dawson, 2011; Driscoll, 2005; Mayer, 2011; Schunk & Zimmerman, 2006).

Based on the literature, there seems to be a paucity of research on how motivation and self-regulation influence student retention in a higher education setting. This study extends the current research of student retention by examining the role of motivation and self-regulation in online course attrition and retention and is unique in that it measures levels of motivation and self-regulation using the MSLQ with students who are or have been enrolled in a distance education program at two different schools of higher education. While motivation and self-regulation are known to be important factors in education success, this study measures those two factors with students who are currently enrolled in a distance education program or have recently dropped out of one. The primary purpose of the study was to attempt to correlate levels of motivation and regulation with attrition or retention in two higher education distance education programs and two different student groups.

This study was guided by the following research questions:

1. To what extent is there a correlation between the student’s score on the motivation scales and retention?

2. To what extent is there a correlation between the student’s score on the learning strategies scales and retention?

3. What are the expressed reasons given by students for either dropping a course beyond the allowable drop date, dropping out of the program completely, or for continuing with a course until completion?

In this study, two groups of students were administered the MSLQ survey. The first group consisted of students who had dropped out of a distance education class or a particular distance education program. The second group consisted of students who were still enrolled in the distance program and have not dropped any of their classes. An 81-item MSLQ was used to determine the level of student motivation and use of learning strategies. The results were analyzed for determining differences in average responses between the two groups. In addition to the 81 MSLQ questions, participants were asked to provide information regarding their enrollment status in the program (i.e., dropped or not dropped) by selecting from the list a statement that best describes their status in the program. An open-ended question asked participants about their felt reasons for attrition or retention in the program.

This study targeted distance education students at two universities in the Midwest region of the United States. Both universities have a residential face-to-face program and an equivalent distance education program. Both schools are accredited private liberal arts universities offering 4-year degrees and a graduate program. One university has an enrollment of 1,300 undergraduate and 230 graduate students. The other university has 2,200 undergraduate and 600 graduate students. Approval to administer the survey was granted by the individual school’s institutional review boards.

The study focused on students in the undergraduate and graduate distance education programs at the selected schools and two groups of students were evaluated. A total of 347 students in the enrolled category were contacted and a total of 162 in the dropped category were contacted. Of the 509 students contacted, 22 from the dropped category and 91 from the not dropped category participated in the study by completing the survey. Using a database of students provided by the schools, students were contacted via e-mail and regular mail inviting them to participate in the study. The first group of students consisted of those who dropped a class beyond the allowable drop date while enrolled in the program or had dropped out of the program entirely. The second group consisted of those who had not dropped courses beyond the allowable drop date and continue to be enrolled in the program.

The MSLQ is an 81 item, self-report instrument ( Appendix A) designed to measure college students’ motivational orientations and their use of different learning strategies (Pintrich, Smith, Garcia, & Mckeachie, 1993). The MSLQ uses a social-cognitive view of motivation and self-regulation and directly links students’ motivation to their ability to selfregulate their learning activities. The MSLQ items are divided into two broad categories of motivation scales and learning strategies scales. The motivation scales consist of 31 items that assess value components (i.e., intrinsic and extrinsic goal orientations, and task value), expectancy components (i.e., control beliefs and self-efficacy for learning and performance), and affective components (i.e. test anxiety).

Learning strategies are used by students to self-regulate their personal processes, their environment, and their behavior. The learning strategies scales consists of 31 items regarding students’ use of different cognitive and metacognitive strategies (i.e., rehearsal, elaboration, organization, critical thinking, and metacognitive self-regulation) and 19 items concerning student management of different resources such as time and study environment, effort regulation, peer learning, and help seeking. The questions use a Likert type scale ranging from 1 (not at all true of me) to 7 (very true of me). The MSLQ was used in this study to determine levels of motivation and learning strategy use. The MSLQ questionnaire’s subscales have high internal consistency. The Cronbach’s alpha range from .52 to .93. Help seeking scale had the lowest Cronbach’s alpha (α = .52) and self-efficacy had the highest Cronbach’s alpha (α = .93) (Pintrich et al., 1991).

A total of 347 students in the enrolled category were contacted and a total of 162 in the dropped category were contacted. Of the 509 students contacted, 22 from the dropped category and 91 from the not dropped category participated in the study by completing the survey. Overall, there was a low rate of participation from both dropped and enrolled groups. However, the enrolled group had higher participation rate than the dropped group. The data were examined to check if there were any missing scores. The results indicate that there were no missing scores in the data. Means and standard deviations for each the MSLQ subscales across retention groups are listed in Table 1.

The Pearson r correlation was conducted to assess the motivation variables that correlate with retention in the distance education program. The analysis indicated that there was a significant positive correlation between selfefficacy for learning and performance and retention ( r = .241, p = .010), such that as the self-efficacy scores increases, the retention in the distance education program also increases. There were no other significant correlations found between the motivation subscales and student retention (see Table 2).

The correlation analysis indicated that there was a significant positive correlation between effort regulation and retention (r = .198, p = .036) and between peer learning and retention ( r = .185, p = .049). The results suggest that as effort regulation and peer learning increases, students’ retention in the distance program also increases. There were no other significant correlations found between the learning strategy subscales and student retention (see Table 3).

A Mann-Whitney U test, a nonparametric test was conducted to assess if there is a difference between the dropped and enrolled groups. This test was chosen as an alternative to the t test because of the two groups had unequal sample size and were not normally distributed. The Mann-Whitney U test is similar to the t test but it is more appropriate to be conducted when the data do not meet the parametric assumptions of the t test (McKnight & Najab, 2010). The results indicated that the self-efficacy score of the enrolled group (M = 5.906, SD = .748) is significantly different from that of the dropped group (M = 5.426, SD = .859), U = 634.00, p = .008).

The goal of this study was to be able to predict which students will drop out. The Mann-Whitney U test revealed that the two groups differed on the self-efficacy scores. However, a regression analysis is necessary to create a model that predicts student retention. The dependent variable (i.e., retention) has only two outcomes (i.e., enrolled or dropped). In

such situation, conducting a binary logistic regression analysis is more appropriate option than conducting simple regression analysis. A binary logistic regression using the forward likelihood ratio (LR) method was conducted. The chi-square test was significant (χ²(1) = 6.061, p = .014) which indicates that the new model is significantly better at predicting student retention compared to the baseline model. The Hosmer and Lemeshow test had a significance value greater than p = .05, which suggests that the model is a good-fit with the data (χ2(8) = 12.589, p = .127). The results indicated that self-efficacy was the only significant predictor of student retention in the program (β = .703, p = .015). The Exp(B) value associated with self-efficacy is 2.021. This means that with ever one unit increase in self-efficacy score, the students are 2.021 times more likely to stay enrolled in the program.

The stated reasons for dropping or not dropping give interesting insight into the motivations of the students. Of the 97 surveyed participants who provided information about their motivations for not dropping the program,67% (n = 65) stated that extrinsic goal orientation was the reason they did not drop the program. This shows that the main concern is not the learning process itself (i.e., taking the class and doing the work), but the means to an end, be it career, improvement, and grades, among others. Of the 15 surveyed participants who provided information about their motivations for dropping the program, 33% (n = 5) gave low task value as their reason to drop out. Low task value refers to the student’s perceptions of the program in terms of interest, importance, and utility. According to the Binary logistic regression, self-efficacy was found to be a good predictor of student retention, whereas, according to student report, task value and extrinsic goal orientation were found to be the reason for student retention and attrition in the program. Self-efficacy, task value, and extrinsic goal orientation fall under the motivation scale category of the MSLQ.

This study revealed a statistically significant correlation between self-efficacy, a motivation subscale, and student retention. The felt reasons given for attrition or retention (i.e., extrinsic goal orientation and task value) were also from the motivation category. Additionally, effort regulation and peer learning, subscales of learning strategy, were also found to be significantly correlated with student retention. This study poses significance from an instructional design perspective since the motivational and self-regulation aspects can be built into the courses. For instance, good instructional design can build courses that are perceived as having a high task value, encourage and motivate the students in their progress, and build learning strategies into the course and study work so that students can improve their performance as they work through the material. In short, good instructional design can boost motivation and self-regulation (Anderman & Dawson, 2011).

The two retention groups significantly differed from each other on the self-efficacy for learning and performance score. The selfefficacy scale comprises two aspects—namely expectancy for success and self-efficacy. Expectancy for success alludes to performance expectations. Self-efficacy refers to selfappraisal of one’s ability to master a given task. It also refers to one’s ability to accomplish a given task and to one’s confidence in one’s skills to perform the task (Pintrich et al., 1991). Moore (1989) suggested that in a setting where students and teachers are separated, additional responsibility is placed on students to evaluate and motivate their learning process. As the students progress through the learning process, they engage in what Bandura (1976) and Zimmerman (1989) referred to as a reciprocal causation among three influence processes: personal, environmental, and behavioral. This process can lead to interpretations and conclusions by the students that have a direct influence on their self-efficacy for learning and performance. As Zimmerman and Martinez-Pons (1990) indicated, self-efficacy beliefs influence achievement behaviors. Also, achievement has a reciprocal influence on selfefficacy. This study depicts significance of self-efficacy beliefs on retention in the online environment. Motivation and attitude are two key components of the learning and teaching process. A negative reciprocal causation can directly impact self-efficacy, a key component in distance education success (Morrison et al., 2013).

Effort regulation, a resource management strategy, refers to the students’ ability to control their effort and attention when faced with distractions and uninteresting tasks (Pintrich et al., 1991). Bandura and Schunk (1981) indicated the importance of self-motivation in the distance education environment where the students do not have external factors to motivate them. The results of the present study indicate that students who are not able to self-regulate their efforts are at a higher risk of dropping out.

Peer learning, a resource management strategy, involves collaborating and communicating with peers to clarify and to gain insights on course materials (Pintrich et al., 1991). Moore (1989) listed learner-learner interaction as one of the three significant interactions in the learning process. In this study, the correlation analysis indicated that peer learning (i.e., student-student interaction) is significantly correlated with retention in online program. The result therefore supports the findings reported by Moore (1989).

This study, unlike the studies in the literature review, uses a very specific definition of academic success. It defines academic success as the improper dropping of courses or attrition, and retention. This study directly addresses the issue of high attrition rate in distance education. This is one of the biggest problems of distance education. The significant correlation between retention and selfefficacy for learning and performance, effort regulation, and peer learning informs the instructional design process to minimize the impact of the separation of the students from the teacher and their fellow students.

The overall sample size of this study is small ( n = 113) and the number of dropped students who participated in the study was even smaller (n = 22). It is possible that a larger sample size for both the dropped and not dropped could have yielded different results. Some limitations of surveying students who have already dropped from a particular program are the difficulty of contacting them, their unwillingness to participate in a survey regarding a program they are no longer part of, and receiving survey results that may not accurately reflect the reality of when they were enrolled in the program and deciding to withdraw.

Studies seeking to correlate motivated strategies with retention and attrition may have a better return of student feedback if they are surveyed at the beginning of their program and near the end of the first semester while they are still enrolled. The results could then be correlated with future retention or attrition. One of the difficulties of distance education attrition studies is getting quality feedback from those who are no longer enrolled in the academic program. Future research is needed to explore the impact of the instructional design of a course that intentionally seeks to show the value of what was being learned, that encourages and strengthens the students’ perception of their ability to do the course, and that recommends several learning strategies such as peer learning and time management as part of the course.