Determinants for Completion of Online Undergraduate Degree Studies in the Anglophone Caribbean

Internet availability has made online education more accessible and, for many, a more flexible way of pursuing higher education. Some authors agree that accessibility and flexibility have significantly contributed to the growth in online education offerings (Lee & Choi, 2011; Zimmerman, 2012). However, as the number of students pursuing online studies continues to increase, so too does the number of online learners who fail to complete, particularly when compared to their traditional counterparts (Stone et al., 2019). At the study institution, the number of online students moved from 3,648 in the academic year (AY) 2007–2008 to 6,348 in the AY 2017–2018, with an attrition rate of approximately 25% (Warrican et al., 2014).

The growth in online education, particularly at the higher education level, has fueled much research, with authors seeking to understand the online environment better, attributes of learners, and factors that contribute to or predict attrition, dropout, persistence, and continuance completion, in online education. Although some research has been conducted within the Caribbean into online education, it has mainly considered: learner perceptions and attitudes to online learning in a single country (Barclay et al., 2018; Kerr, 2015); engaging the digital natives (Veira et al., 2014); first-year learner perceptions of student support (Hunte, 2012); the impact of technological readiness, past learning experience, and transactional distance on success (Dottin, 2016); the utility of online learning to higher education (Boisselle, 2014); and predictors of student success in specific courses (Warrican et al., 2014).

The Warrican et al. (2014) study investigated the predictors of success for online learners pursuing undergraduate studies. That study focused on two courses with high failure rates: Introduction to Financial Accounting and Introduction to Macroeconomics. Archival admissions and academic records for online students from the AY 2008–2009 cohort to Semester 1, 2012– 2013 were reviewed. Warrican et al. concluded that sex, student location, cumulative grade point average at the end of the first year, and use of course resources were significant predictors of program completion.

Unlike Warrican et al. (2014), this study is neither course-specific nor purely quantitative. It considers success in program persistence, retention, completion, and dropout in program attrition. This case study targets first-time online students who enrolled in the academic year 2013– 2014 who persisted and those who dropped out and sought to understand the factors contributing to their persistence or dropout. It tracked progress from enrollment through to degree program completion, dropout, or to the AY 2018–2019, whichever happened first, with the view of ascertaining the factors which contributed to their persistence decisions.

The study is based at The University of the West Indies, a multimode regional institution, which, through its Open Campus, offers online degree programs from associate degree to doctoral level. The University of the West Indies, Open Campus was established in 2008 as a mechanism, among other things, to provide access to higher education, particularly to the noncampus countries within the Anglophone Caribbean region.

The University of the West Indies is a five-campus institution with physical locations in 16 Anglophone countries. Although some countries do have local universities, The University of the West Indies is perceived to be the primer university, and as such, many learners dreamed of attending but were hindered due to the costs and other factors associated with living and studying in another country. Therefore, online offerings provide a measure of equity and an opportunity for learners from countries without a physical campus to achieve their higher education dreams (Kent, 2016; Pollard, 2018; Smith et al., 2017). This view is consistent with those of Caribbean aspirants.

The demand for and provision of online learning continues to increase, particularly for higher education and nontraditional learners. Nontraditional learners are mostly in paid employment, over 24 years of age and have responsibilities for the care of children and other family members, e.g., parents (O’Shea et al., 2015; Ragusa & Crampton, 2018; Signor & Moore, 2014; Stone et al., 2019). As such, their needs would not be consistent with those of traditional-aged learners. This demand is presently exaggerated by the COVID-19 pandemic, which forced educators and educational institutions at all levels across the globe to find alternative ways to engage learners. Perhaps this phenomenon is signaling the future of education at all levels, and research into this modality will continue to be fertile ground.

Several authors have emphasized the importance of understanding the online learning environment from the learner perspective, including concerning student satisfaction (Kara & DeShields, 2004); student self-determination (Chen & Jang, 2010; Wehmeyer et al., 2003); student selfefficacy (Garrison et al., 2000; Shea & Bidjerano, 2010;); students as digital immigrants (Chu & Tasi, 2009; Prensky, 2001;); the effect of learning theories (Silva et al., 2009); student connectivism (Siemens, 2014); family, work, and personal responsibilities like caring for elderly parents and children (Greenland & Moore, 2014; Ilgaz & Gulbahar, 2015; Markle, 2015); technological and pedagogical readiness (Dottin, 2016); and role strain, the conflict that may arise between competing responsibilities (Markle, 2015). This research focused on providing institutions with information that could be used to support online learners from enrollment to completion.

Additionally, other researchers have considered the institutional factors that may contribute to learner persistence, including transactional distance and technological challenges (Dottin, 2016; Yoo Joo & Huang, 2013), quality of course design, material, and delivery (Devlin & McKay, 2016); facilitation of student engagement (Alman et al., 2012; Eliasquevici et al., 2017; Lambrinidis, 2014; Pinchbeck & Heaney, 2017; Shah & Cheng, 2018; Vincenzes & Drew, 2017); facilitation of learning (Fredrickson, 2015; Garratt-Reed et al., 2016; Wuellner, 2013); and course design (Hammond & Shoemarker, 2014; Pittenger & Doering, 2010).

Interestingly, more recent studies have focused on the “importance of understanding and recognizing the diversity of the online student cohort, contending that only through recognizing and valuing this cohort can an equitable experience be achieved” (Stone et al., 2019, p. 27). Therefore, knowledge of these factors should influence institutional provisions and support for persons who choose the online course of study. This knowledge is particularly vital, as online education is a popular option for nontraditional learners due mainly to its perceived flexibility, accessibility, and convenience (Boling et al., 2012; Kuyini, 2011; O’Shea et al., 2015; Stone et al., 2016).

As much as flexibility is considered an essential attribute of online learning, this is not often the case due mainly to university regulations (Stone et al., 2019), usually legacy regulations that were not adapted to suit the online environment. Todhunter (2013) found that there was often a mismatch between students’ expectations of flexibility; for example, more relaxed times lines for submission of assignments, and the reality of the institutional provision, noting that “rarely does the level of flexibility extend beyond how students interact with staff, learning resources, and fellow students” (p. 240). So, although online learning increases access to education, notably higher education for adults and persons living in rural and other areas without a physical university, as well as those persons who, due to work and family responsibilities, may be unable to attend face-to-face classes, it may often lack the flexibility to provide the equity which these learners expect and require for persistence.

This study investigated persistence and dropout of undergraduate online learners from the academic cohort 2013–2014. Even though much research has been conducted into online attrition, also known as drop out or nonpersistence, there is still no universally accepted definition for this concept, which has meant that attrition rates cannot be accurately compared across institutions due to differences in reporting (Atchleyet et al., 2013; Cauble, 2015; Hart, 2012). The definition for attrition ranges from ‘a learner being required to be active in the class during the first 3 weeks’ (Balter et al., 2013); to “learners who registered but dropped out: prior to the start of class, prior to start of instruction, during orientation, and or after orientation” (Frydenberg, 2007).

Given the varying definitions associated with online attrition, in this study, it is defined as learners who enrolled, paid the requisite tuition fees, completed at least one semester, completed at least three courses on the program, were in good financial and academic standing, but failed to register for a course for at least two consecutive academic years, that is, voluntary withdrawal from the program. This definition eliminates those learners who were required to withdraw for academic purposes, who failed to enroll in any courses since initial registration, who enrolled but never attended classes or attempted assignments, or attended only for a short time (less than half a semester).

As a final precaution, the survey instrument asked respondents to indicate whether they had actually left the program or taken undocumented leave and planned to return to their online studies at some time in the near future. Retention is defined as learners who are still pursuing their online program. Persistence is defined as those learners who successfully completed their online studies.

The study used a survey instrument consisting of 6 demographic questions, 5 open-ended questions, and 17 multiplechoice questions. Of the 679 past learners who met the requirements, personal email addresses were collected.¹ were available for 325. Of the target figure, 101 past learners responded to the survey representing a confidence level of 5%t, a margin of error of 8.11%, and a response rate of 31%.

The sample comprised predominately female respondents (90%), which was fairly consistent with the target cohort, which was 84% female. The majority of the respondents were in the 3544 (37.6%) age group. The sample consisted of 41 (40.6%) respondents who dropped out and 60 (59.4%) who completed their online studies. Approximately 84% of the respondents were in full-time employment while they were studying, and 50% were single. Of the 41 respondents who dropped out, 63% indicated no plans to return to their online studies. The sample consisted of 24% two-parent family structures with under 18 age children and 22% single parents with under 18 age children. Table 1 provides a sociodemographic profile of the respondents who participated in the study.

The profile of respondents was compared to the 2013–2014 cohort to determine whether the sample was representative of the typical online learner at the institution. Overall, the respondents in this study matched well with the overall cohort in terms of age and gender. The archival data contained no information on marital and employment status or family circumstances.

The data was collected using a modified survey instrument created by Su and Waugh (2018) to study perceptions of two groups of students, persisters and nonpersisters, who participated in WebIT online master of science in instruction technology program at the University of Tennessee at Knoxville. This instrument was chosen as the researchers studied a similar phenomenon, all be it at the program level. The instrument was piloted to 10 online students to determine whether: (1) the respondents understood the questions; (2) items, as framed, caused reluctance/ apprehension when responding; (3) questions were simply stated; (4) the length of time respondents took to complete; (5) there were any ambiguities; and (6) questions were in keeping with the research focus. After the pilot, the research instrument was amended as necessary. The data was collected over 3 weeks via the SurveyMonkey tool.

The cohort archival data showed a graduation rate of 4.7%, 9.2%, and 13.3% after 4, 5, and 6 years respectively. When these findings were compared to the National Centre for Education Statistics (NCES; 2018), it was found that the cohort’s graduation rates after 4 and 5 years were below those of NCES, which were 6.5% and 10.6%, respectively. However, the cohort’s graduation after 6 years was 1.3% higher than the NCES.

Further analysis of the graduation, dropout, and retention rates by age range showed that over 50% of all cohorts are likely to persist after 6 years of initial enrollment. It further showed that the highest dropout rate was experienced by the under 24s and the lowest by the over 55s. Interestingly, graduation rates for all age ranges were similar, ranging from 15.8% for the 45–54 to 11.9% for the 35–44 age range. The graduation by year and age range and dropouts by age range is shown in Table 2. 

Many studies have found online education dropout rates to be usually between 20 and 50% (Carr, 2000; Greenland & Moore,2014; Patterson et al., 2012; Rovai & Downey, 2010). Therefore, with a dropout rate of 36%, the cohort is just below the middle spectrum of international norms. As noted earlier, Warrican et al. (2014) reported a dropout rate at the same institution of 25% for the 2007–2008 cohort. This 11% rise in attrition, over 5 AYs, should be carefully evaluated to address what might be a negative trend.

The data further showed that of the 474 students from the AY 2013–2014 cohort, who failed to persist, 428 dropped out by the end of Year 1 and 41 by the end of Year 2. Of the sample, 83% left by the end of Year 1, 15% by the end of Year 2, and 2% by the end of Year 3. These findings are consistent with literature that suggests that learners are more likely to drop out during Year 1 of their studies (e.g., Pascarella & Terenzini, 2005; Tinto, 2012). Table 3 shows the cohort dropout data at the end of AY 2013–2014 to 2017–2018.

The study found that respondents who considered the actual workload of the program to have met (48%) or exceeded (51%) their initial expectations were more likely to persist than those who felt that it was less than expected. Interestingly, only 35% of dropouts felt that it met their expectations, and 20% felt that it was below. Further, respondents who indicated that they spent more time than expected on their online studies were twice as likely to succeed. Of note, the respondents who felt that they spent less time than expected were seven times more likely to fail. As it related to actual time spent on studies, respondents who spent less than 5 hours per week on their studies were three times more likely to fail than those respondents who spent more than 25 hours, but only twice as like to fail as those that spent between 5–10 and 11–15 hours weekly.

When a Spearman rank order (ρ) procedure was run to investigate the relationship between perception of actual workload and program completion, the results showed a small positive correlation (ρ = 0.290). This relationship was statistically significant ( p = >0.002). The literature notes that learners do not usually consider the workload and the resultant level of personal involvement required to pursue online learning (Bawa, 2016). Learners are usually more focused on the perceived flexibility, convenience, and reduced costs of this modality.

However, due to this lack of awareness, many online learners are unpleasantly surprised and leave the online environment due to this unexpected reality. When the open-ended questions were analyzed, there was consistency about the online workload being too extensive. The sentiments are best summed up in the words of Respondents A, M, and K. Respondent A noted

Similarly, Respondent M felt that there was a need to have more

Respondent M was stating that there was a need for greater learner autonomy and flexibility to learn at their own pace. Similarly, Respondent K, a dropout, also highlighted concerns associated with workload, group work, and flexibility and noted that

The literature suggests that flexibility is a key reason why persons join and persist in online education. As early as 2000, Bates found that the expectation of tertiary education students had evolved from being mainly concerned with the completion of a program and the on-campus experience to being “interested in small modules and short programs ... and in learning that can be done at home and fitted around work, family, and social obligations” (p. 5). Similarly, as noted previously, and as highlighted by Respondent K, there is a disconnect between the online learner’s perception of flexibility and the reality of the online environment.

The lack of program flexibility was the primary reason identified by respondents for leaving. In addition to Respondent K, several other learners shared their views on flexibility. These include Respondent M, a dropout, who said that there needed to be “(1) Workable deadlines. (2) Remove mandatory expectations that students must interact on the course page at least once a week. (3) If not (2) above, remove course mark allotted to mandatory posting/interaction on course page at least once per week.” In other words, the learners needed more flexibility to complete and submit assignments at their convenience without having to meet institutional mandated deadlines. Similarly, Respondent C dropped out because “The program needs to be more flexible and not so rigid, allowing for students to move at their own pace.” These comments were agreed to by Respondent D, who, in keeping with Bawa (2016), acknowledged

Notably, the program’s flexibility was identified by persisters as the second more critical factor contributing to their decision to complete the online program. However, the views of the persisters were dissimilar to the dropouts. For example, Respondent F noted

Likewise, Respondent G, a persister, said, “Flexibility, high standards, and the fact that I don’t have to sit in a class with people.” Along the same lines, Respondent H summarized, as follows

Perhaps recognizing the limitations of the concept of flexibility from an institutional perspective, Respondent I noted that the online program provided the “ability to review class session if absent or stepped away from the computer, flexibility towards completing online tests (any time before deadline).” Several components or views on the concept of flexibility were highlighted by the dropouts, including Respondent D, who expected greater flexibility and cited having a job and family, as justification for this need, which is in keeping with Bawa (2016), who noted that “common assumptions [misconception] related to online learning are that because face-to-face presence is not required, an online platform will be less demanding on time, will require less effort to manage workload, and will not disrupt the learners’ lifestyle” (p. 3).

However, what was even more interesting about learners’ perceptions of flexibility, perhaps could be summed up in the words of Respondent I, a persister, who acknowledged that flexibility did not relate to course deadlines but to the ability to access the course materials and recordings at your convenience. Maybe, this is the key difference between persisters and dropouts; persisters have mentally shifted their perception of flexibility to not include flexible submission dates and times, whereas the findings suggest that dropouts had not made that mental adjustment and still view flexibility as total and not just related to the convenience of access to learning materials and recordings. Additionally, this finding highlights a role for institutions to assist online learners in understanding, appreciating, and making the mental shift in their concept of flexibility through tailored communication and training or to adapt their institutional regulations to accommodate better the forms of flexibility that the online learners require.

The study revealed several perceptions about the online learning environment that are mostly misguided. Unlike most teaching and learning environments in the Caribbean, the online environment is mainly constructivist, largely self-driven, and dependent on the learners’ ability to manage academic responsibilities. Usually, it consists of fewer props than may be available in the face-to-face learning environment. As such, learners who have not previously experienced this kind of selfdirected academic discipline may be more likely to experience demotivation and resultant drop out as they may be uncomfortable with self-learning and constructing knowledge. Respondent A, a dropout, provided an excellent example of this when they noted that they experienced “lack of discipline in self learning” and Respondent B, who noted that “the course was not properly geared toward me grasping the information ... preferred more teacher-student interaction.”

Although a key component of transactional distance, the preference for teacherstudent interaction may also be prevalent due to the past learning experience of some learners, which were traditionally didactic in nature as opposed to constructivist (World Bank, 2005). Conceivably the didactic orientation to learning and the need for interaction and direction meant that those responses from self-learners who participated in discussion forums and Blackboard Collaborate sessions hosted by facilitators were twice as likely to persist as those who did neither. Further, the respondents who persisted were twice as likely to have preferred and attended more synchronous sessions with facilitators than the dropouts. Once again, this highlights the importance of tutor-led learning for this cohort of online learners.

The need for tutor-led learning may have resulted in a disconnect between the expectations of some learners, informed by their past learning experiences, and the realities of the online learning environment. For example, Respondent C, a dropout, said, “interaction on Backboard for a 3-hour class once per week would have also assisted in the student understanding the course material and facilitated facilitatorstudent interaction.” Respondent B, a dropout, raised an interesting concern about the institution’s inability to accommodate learners with learning challenges adequately. To address the concern, Respondent B first identified the tutor and next the availability of learning resources as the solution. No consideration seems to have been given to the role of the learner in the teaching and learning paradigm. Respondent B noted that

Notably, the respondent incorrectly equated learning challenges with variation in learning styles or learning preference as it is now more widely known.

Although learning style was not defined in the survey, respondents were asked to indicate whether the online environment suited their learning style. Most significantly, 87% of the persisters indicated that their learning style was suited to the online environment, compared to 46% of dropouts. The study also found that most respondents who persisted (61%) preferred to work independently to a fixed facilitator schedule compared to 45% of respondents who dropped out. Similarly, 50% of the dropouts preferred to work at their own pace compared to 30% of the persisters. Of note, 70% of the persisters preferred a structured, logical sequence to their program compared to 50% of the dropouts. When a Spearman rank order (ρ) procedure was run to investigate the relationship between suitability of learning style to program completion, the results showed a medium positive correlation (ρ = 0.488). This relationship was statistically significant ( p = >0.0001).

Several authors, including Markle (2015), have documented the challenges faced by learners when balancing family, work, and study commitments. Often referred to as role conflict or interrole conflict is known to impact persistence decisions significantly. This study found that respondents who prioritized: (1) family; (2) studies; (3) work; and (4) other personal commitments were more likely to succeed compared to those who prioritized (1) family; (2) work; (3) studies; and (4) other personal commitments. When a Spearman rho procedure was run to ascertain the relationship between family prioritization and completion, the results show a weak positive correlation (ρ = 0.217). The results were statistically significant (p = 0.023). When a similar test was run to determine the relationship between prioritization of studies and completion of the program, a weak positive correlation was again found (ρ = 0.265). The results were statistically significant at ( p = 0.005). These findings showed that role and interrole conflict might be mitigated by ordering or prioritizing studies since, for both persisters and dropouts, the family was ranked as the number one priority. However, once the placement of studies was moved from second to third place or replaced with work, the likelihood of failing to complete the program doubled.

Twelve factors were identified that affected persistence decision these were: effects or suitability of a program to learning style, prioritization of study, prioritization of family, time spent studying, preference for facilitator control, perception of the structure of the program, number of courses taken per year; family circumstance, age, employment status, gender and availability of internet access. A binary logistic regression was performed to understand better which of these factors were more influential on the persistence decision.

The binary logistic regression found that perceived suitability of program structure to learning style, age, and family circumstance are significant predictors of program persistence decisions [χ² = 39.257, df = 12 and p = 0.0005 (<0.05)]. The other nine predictors were not significant. The model explained between 39.5% (Cox and Snell R2) and 54.6% (Nagelkerke R2) of the variance in the results. All of the 12 predictors “explain” 54.6% of the variability associated with completion and dropout in online learning. Suitability of structure for learning style, age and family structure are significant at the 5% level (Suitability of structure to learning style: Wald = 9.313, p = 0.002; Age Wald = 4.242, p = 0.039; Family Structure: Wald = 6.764, p = 0.009]. The odds ratio for suitability of structure to learning style is 13.952 (95% CI 2.869 – 67.853); age 0.407 (95% CI 0.173 – 0.957); and Family Structure 1.708 (95% CI 1.141 – 2.556). The model correctly predicted 88.2% of persisters and 77.8% of dropouts, giving an overall percentage correct prediction rate of 84.6%.

In summary, this study found that although several factors influence persistence decisions, the best indicators of online learner persistence in this cohort were: suitability of program structure to learning style, family circumstance, and age, in descending order of importance.

The findings are based on a single sample cohort of Anglophone Caribbean online learners. Although all responses were voluntary and anonymous, one cannot determine the truthfulness of the individual responses, even though they are assumed to be accurate. Despite this, several assumptions and tentative conclusions may be derived from this study.

This study was exploratory in nature, targeted learners from the AY 2013–2014 cohort, and was intended to identify factors that contribute to the decision by online learners to persist or drop out. The study found that learners identified 12 factors as instrumental in their persistence decision and were also statistically significant, including the placement of studying in their list of priorities; for example, when studying was placed third, dropout increased. However, when a binary logistic regression procedure was performed, it was revealed that the best indicators of persistence for this cohort, in order of priority, were: suitability of program structure to learning style, family structure, and age.

The study further found that the learners who understood flexibility in terms of access to learning materials and recordings at any time and not restricted it to anytime submissions of assignments were more likely to complete their studies. It also found that priority placed on studies was vital to success, that is, priority number two. However, once the position of studies moved to position three or below, nonpersistence usually resulted.

These findings have several implications for future research and institutional policies and procedures; for example, using these findings, institutions should focus more time in orientation to time management, prioritization of studies, and ensuring that students fully understand flexibility in the online environment. Additional research may include conduct a comparative analysis to determine whether these findings are consistent across cohorts and, as such, representative of Anglophone Caribbean online learners. The comparative analysis would provide not only regionally specific data but information that would inform institutional practices. Finally, the findings have shed more light on the factors contributing to the success, and by extension, attrition of Anglophone Caribbean online students. Higher education institutions in the region should use this knowledge, and extra-regionally, to establish policies, procedures, and systems to address the educational needs of this category of learners.