Learner characteristics attributable to completion and duration of completion in engineering and science degree programmes of OUSL

Abstract

Purpose

The purpose of this study is to identify the learner characteristics attributable to the likelihood and the duration of programme completion in the Bachelor of Science (BSc) and Bachelor of Technology Honours in Engineering (BTech) degree programmes of the Open University of Sri Lanka (OUSL).

Design/methodology/approach

Data were gathered from the re-registrants for the degree programmes in the academic year 2020/2021, using a questionnaire developed as a Google form. The sample consisted of 301 and 516 re-registrants from the BTech and BSc programmes respectively. Influential factors were identified using Kruskal Wallis test (for duration of completion), binary logistic regression (for likelihood of completion) and Chi-squared test (associations between presage and process factors).

Findings

Entry qualification, age and time management skills at entry had significant effects on duration of completion. Attendance at academic activities, organizing time for self-studies and the competency in English at enrolment had significant effects on the likelihood of completion. Prior open and distance learning (ODL) experience had no significant effect on any of the product factors considered.

Research limitations/implications

Inaccessibility of dropouts and using only the responses from the first administration of the questionnaire are limitations. Active learners are more likely to respond, in the first administration and may bias the results.

Practical implications

Findings are useful for designing future studies to identify at-risk students and thereby enhance the programme completion and reduce prolonged time for completion.

Social implications

Effective strategies to control the identified factors will uplift programme completion and reduce drop-out rates.

Originality/value

Decision making using inferential techniques makes the study distinct among studies undertaken on the same population. The study enriches the limited current research on factors affecting programme completion in ODL mode.

Citation

Navaratna, C.W., Bandarage, G., Rajapaksha Appuhamilage, D.N., Pasqual, H., Rajendra, J.C.N., Ranasinghe, M.D.D. and Ratnayake, U.W. (2022), "Learner characteristics attributable to completion and duration of completion in engineering and science degree programmes of OUSL", Asian Association of Open Universities Journal, Vol. 17 No. 1, pp. 65-83. https://doi.org/10.1108/AAOUJ-09-2021-0114

Introduction

In the recent years, open and distance learning (ODL) has drastically captured the interest of learners as an attractive and viable pathway to higher education. Marked growth in the economic value placed on higher education, along with increased family and work commitments, has boosted the demand for higher education in an accessible mode more than ever before. The current pandemic situation has driven even the conventional mode institutions to adopt ODL methodologies at least as a temporary measure to continue offering the regular progammes. ODL methodologies enable reaching large numbers of spatially dispersed learners and permit catering to learners seeking flexible hours for studies. Yet, many researchers worldwide have identified poor completion rates and prolonged time for completion as issues manifested in ODL systems (Ibrahim and Van der Heijden, 2019; Simpson, 2013; Taniguchi and Kaufman, 2005). Literature reveals many studies conducted internationally (Li and Wong, 2019; Radheshyam et al., 2017) as well as locally (Aluwihare and De Silva, 2016; Dedigamuwa and Senanayake, 2012) that identify factors affecting learning and learner support is reported as one such crucial factor.

Learner support covers a wide spectrum extending to intellectual, emotional, financial and social needs of learners. Exogenous factors such as changes in secondary education curriculum and learning environments, growing opportunities for acquiring different modes of learner support, changes in family and work commitments, etc., demand variations in the support that learners seek. Technological advancements, emergence of new knowledge related to pedagogical and andragogical aspects, developments in infra-structure facilities, etc., necessitate institutions to repeatedly rethink the extent and the mode of learner support that should be provided, as well as how to optimally integrate different modes of learner support with diverse learning environments. Consequently, research on changes that can uplift the learning processes turns out to be a dynamical, ongoing integral component, especially in ODL systems.

An aspect that needs focus is the inherent diversity among learners in ODL systems, with respect to demographics, psychological factors, learner intentions and competencies (Ariadurai and Manohanthan, 2008). Learning styles, learning approaches, learner requirements and accessibility for resources, apart from varying across learners, fluctuate over time as well. Providing customized learner support giving opportunity for the learners to choose from as opposed to providing a monotone mode of learner support, emerges as a vision, for which the feasibility, and strategies suitable for the local scenario need to be investigated. This report presents the findings from a preliminary study undertaken as a first step in this direction with the specific aim of identifying factors that have significant impact on the learning. The findings from this study enhance the awareness of learner characteristics and learner support requirements and will also be useful in selecting input variables for predictive analytic models for performance of students.

Rest of the article is organized in subsections, in the order, related studies, methodology and the results of the empirical study. Finally, we present the conclusions and discussion with suggestions for further work.

Related studies

Literature reports multifaceted factors affecting student persistence in ODL. This section reports the articles that we closely followed in choosing factors for our study.

Hart (2012) identifies satisfaction with online learning, a sense of belonging to the learning community, motivation, support from peers and family, time management skills, and increased communication with the instructor as factors associated with student persistence in an online programme. Au et al. (2018) report that factors linked with success of learners vary depending on levels of achievements of the learners and recommend improving time management skills and student skills of weak students to enhance persistence in ODL.

Li and Wong (2019), based on a review study, report 194 influential factors for student persistence along with the changes and trends in the identified factors over time. These authors have broadly categorized these factors into student factors, institutional factors and environmental factors. Student factors are further classified as pre-enrolment factors to cover demographic factors, prior educational experience and skills and post-enrolment factors to cover the learning process inclusive of planning, managing and resource allocation for studies. In listing out the independent variables for the study, we closely followed this classification.

Radheshyam et al. (2017) based on a review of 85 papers, present 48 influential factors for academic performance at university level. They broadly classify these independent factors as academic factors and non-academic factors. Five dependent factors are proposed to assess the performance, namely, semester cumulative grade point average, a dummy variable representing success or failure of the student, relative retention rate, time to graduation and marks obtained at end-semester examination. Since our focus is on programme completion, we use two product factors, which are a dummy variable representing completion or non-completion of the programme and the time for completion of the degree.

Soto and Anand (2009) applying chi-square test of independence have shown that completion of pre-requisites, passing the laboratory component of cell biology, accomplishment of homework, and attendance were related to passing a cell biology core course, where they have defined passing the course as obtaining a C grade or higher. Applying logistic regression, the same authors have shown that perfect attendance followed by GPA are the most important factors associated with passing the course. Since our focus is on the completion of a programme and not about a single course, attendance was regarded as a factor that is more applicable to our context.

Turning towards local studies, Aluwihare and De Silva (2016) have studied personal factors and institutional factors associated with prolonged completion. Among the personal factors, performance at the secondary education Level (G.C.E. Advanced Level) and students' awareness on ODL, work-related challenges, travel time and cost of commuting to the main Centre located in Colombo and inability to spend the required time expected of the programme are reported as barriers that delay programme completion. Among the institutional factors, poor academic counselling and guidance, laboratory sessions being inappropriate or insufficient to understand the course material; inadequate library resources and laboratory facilities at the regional centres are reported as contributing factors for prolonged degree completion.

The prevailing conditions applicable for learners at the Open University of Sri Lanka (OUSL), especially the geographical, environmental and financial aspects may be different to those faced by the learners focused on studies done in other countries. This is evident from a comparison of the institutional factors, identified as influential for the local scenario, by Aluwihare and De Silva (2016) with those of other studies stated here. On the contrary, time management skill is a personal factor identified in all the studies stated here, as crucial for successful completion. In recent years, the OUSL has taken steps to enhance learner support, including provision of a pre-academic readiness training programme for learners to empower for independent learning. Major changes in programme delivery have also happened including enhanced support through supplementary online courses delivered via LearnOUSL platform. Aligning with these, the influential factors identified in the previous studies for the local scenario may have changed. It is therefore vital to re-investigate the factors affecting current learners in the focused degree programmes.

Methodology

Here onwards, the Bachelor of Science and Bachelor of Technology Honours in Engineering (BTech) degree programmes of the OUSL will be referred to as Bachelor of Science (BSc) and BTech degree programmes, respectively.

Study sample

The sampled population comprised 6,423 re-registrants for the BSc and BTech degree programmes in the academic year 2020/2021. The sample comprised 817 respondents to the survey in the first administration and included learners at various stages of the learning process and degree requirement completers awaiting graduation. The questionnaire was pilot tested and revised before administering to the students in the survey sample.

The data were gathered anonymously by making the link for the Google form accessible to BSc students via SMS, since Faculty of Natural Sciences had a database of mobile numbers used for communication. For BTech students, questionnaire was made accessible via MyOUSL student web portal, as it is a common platform used for communication, by the Faculty of Engineering Technology.

Survey instrument and factors

The factors identified from the related studies reviewed were shortlisted to a manageable number, and were customized to suit the local scenario, to cover the three broad areas: student factors, institutional factors and environmental factors (Li and Wong, 2019). A questionnaire was developed as a Google form focusing on information under the broad categories: Learner profile (demographic data, entry qualifications, English language and ICT skills; time management skills at entry, etc.), Learning environment (place of study, support from mentors, financial support for education, etc.), Learning behaviour (time management during studies, study practices, attitude and commitment towards studies, etc.), Learner support (learning resources, opportunities for interactions with instructors and peers, flexibility available in the programme in scheduling academic activities, etc.) and Learner constraints (extensive commitments required for family, work and other activities, financial difficulties, other unexpected constraints such as unfavorable life events, etc.).

Conceptual model

Literature reports (Li and Wong, 2019) several models for student retention, perseverance and attrition. We organized our study closely following the Bigg's 3P model (Biggs, 1987) that describes the relationships between Presage factors (prior to enrollment), Process factors (during learning) and Product factors (learning outcomes). Figure 1 illustrates the conceptual model used, where we have closely followed the notation used in Bigg's model. The letters (a), (b) and (c) denote the effects of presage factors on product factors, presage factors on the process factors and process factors on the product factors, respectively.

We examined two learning outcome variables (product factors), namely the likelihood of program completion and the time taken for completion. The presage factors and the process factors we selected for this study will be described in the results section.

Objectives of the study

The specific objectives of this study are:

1. To identify

   • The presage factors that have direct impact on the product factors

   • The presage factors that have significant associations with the process factors and

   • The process factors that influence the product factors considered in this study

2. Make recommendations for improving the completion rates of the focused degree programmes

3. To examine the generalizability of the Bigg's 3P model for the local context

Data analysis tools used

Descriptive statistical tools were used to quantitatively describe and graphically illustrate basic features of the data and to explore associations between variables.

In our study, the response variable “time duration for programme completion” assumes integer values. Hence, Kruskal–Wallis (KW) test that applies to ordinal scale data (Kruskal, 1952; Ostertagová et al., 2014) was applied to identify the factors that are significantly associated with the time duration for program completion. The other response variable that represents the likelihood of programme completion is a binary scale variable that was defined to take values 1 and 0 depending on whether the degree requirements are “completed” or “not completed”. Hence, binary logistic regression (Agresti, 2013; Kleinbaum and Klein, 2010; Tyler et. al., 2021) was applied to identify the factors that are significantly associated with the likelihood of programme completion. Presage factors and process factors considered in this study are either measured at the nominal scale or at the ordinal scale. Hence, Chi-squared test of association (Agresti, 2013; Radheshyam et al., 2017)) was used to identify presage factors that have significant associations with the process factors, as well as to identify process factors that have significant impact on the product factors. The data were analysed using the SPSS Version 27 statistical software.

Results

Background information

The programmes at the OUSL comprise levels, as years in a conventional university system. The university offers a foundation programme, currently named Advanced Certificate in Science, as a pathway for students who lack required qualifications to directly enter the degree programme. Levels 1 and 2 are reserved for the foundation level. The degree programmes start at Level 3 and continue up to Level 5 or Level 6 depending on whether the programme is a three-year programme or a four-year programme. At each level, students are required to fulfill a workload amounting to a total of 30 credits, subject to stipulated conditions. One credit is considered as a workload amounting to 50 notional hours, which we describe in detail later.

In the BSc degree programme, out of the 3,878 registrants contacted, 516 respondents (around 13%) comprised 62 (12.0%) at Level3, 182 (35.3%) at Level4, 199 (38.6%) at Level5 and 73 (14.1%) degree requirement completers. In the BTech degree programme, out of the 2545 registrants contacted, 301 respondents (around 12%) comprised 32 (10.6%) at Level3, 62 (20.6%) at Level4, 143 (47.5%) at Level5 and 64 (21.3%) degree requirement completers.

Analysis of presage factors

Demographic characteristics

Learners in the two degree programmes had entered with similar ODL experience (see the Appendix). A remarkable percentage (BSc (86%), BTech (81%)) has entered with no experience in ODL mode. As well, the age composition in the two programmes is similar. However, the two programmes substantially differ with respect to gender composition, stream of study at the secondary level and entry qualifications. The BTech degree is mostly followed by males whereas the BSc is the contrary. Around 34% of entrants to the BTech degree have entered with a diploma as opposed to none in the BSc degree.

Some key features of the data (see Appendix) that is worth highlighting are:

1. Both programmes have attracted (immediate school leavers (below 20 age group).

2. Around 88% of students in both programmes belong to the 20 to 29 age group. According to our experience in the past, there is a shift in the intake towards attracting younger group. Further work is needed to find out whether the learner support requirements differ across age groups.

3. Over 2% of students had opted to switch from the biological science stream to follow an engineering degree programme. Around 3% had followed the Technology stream. BTech degree programme directly admits students from the physical science stream at the secondary level. Those who have switched from other streams have followed foundation courses to qualify for admission to the BTech degree. The depth in mathematics covered at the secondary level in the Technology stream is different to that in the physical science stream. Further study is needed to find out whether the learner support requirements differ across groups with different entry qualifications.

Pre-enrollment skills

As highlighted in Table 1, the distribution of learners according to self-judgements of knowledge in English, ICT, and other skills (presentation and report writing, teamwork, interpersonal and time management) at the entry to the two degree programmes are quite similar. Only less than 2% had assessed themselves as weak in English and ICT skills. Also, not more than 5% had assessed themselves as weak in report writing and presentation, teamwork, interpersonal and time management skills.

Analysis of process factors

Under process factors, we examine the usage of learning resources during studies and learning behaviours.

Usage of learning resources

Learning resource usage was very similar across the programmes and is summarized collectively for the entire sample of respondents. Course material (94%), MyOUSL (83.8%), past papers (71.7%) and online platforms (69.1%) stand out as the widely used learning resources (see Figure 2).

Learning behaviours of the participants

According to the guidelines elaborated in the Sri Lanka Qualification Framework (SLQF, 2015), the volume of work for a course is measured by the credit rating, where one credit is considered as equivalent to 50 notional hours; this comprises time spent for self-studies, participation for academic activities and examinations. Through course guidebooks, information sheets, awareness at orientation and academic counselling, as well as Student Academic Readiness and Training Programme (StART@OUSL), such information is communicated to students to enable a smooth switch from the spoon fed system of learning at secondary level to the independent learning expected at the university level.

Among the academic activities, attendance for day schools is not compulsory. However, the students are strongly advised to attend the day schools with prior preparation enabling adequate learner-instructor interaction. Students are expected to allocate adequate time for self-studies, using notional hours as a guidance. Under the learning behaviour, we investigated whether the learners have used notional hours as a guidance in allocating time for self-studies, regularly attended the day schools, organized the studies according to an appropriate time schedule or have just studied near to the examinations.

Organizing self-studies

Table 2 illustrates that learners in the two degree programmes have remarkable similarity in organizing time for self-studies. Less than 2% have fully spent time on studies. This is not unusual, given that the programmes are designed and offered in distance mode, opening access to learners with family, employment, and other commitments. However, the fact that considerable percentages of students (approaching 25% at later stages) studying only close to examinations cannot be ignored. These learners are more likely to be surface learners or strategic learners as opposed to deep learners (Biggs, 1987). Despite that the learners have assessed themselves as possessing fairly good time management skills at entry to the programme and have undergone an academic readiness training programme, less than 50% of the respondents have organized time according to a time schedule. Since time management is a crucial component especially in ODL, we remark this as an observation that deserves attention.

Attending academic activities

Unlike the other process factors studied, programme-wise differences exist with regards to participation for academic activities (Table 3), In the BSc degree programme, poor attendance is more prominent at higher levels with the peak at Level 5. On the contrary, in the BTech degree programme, poor attendance is more prominent at lower levels with the peak at Level 3. In both programmes, among students who had completed the degree requirements, close to 40% had indicated good attendance. Among degree completers, poor attendance is more prominent among learners in the BSc degree programme (37%) compared to BTech programme (around 19%).

Time commitment with reference to notional hours

Despite that the notional hours allocated for courses are made available at the start of the programme, as illustrated in Table 4, a considerable fraction of learners (over 50%) has not used notional hours as a guidance.

Analysis of product factors

The two product factors selected for the study are “the time duration for completion” and “the likelihood of completion”. In identifying factors associated with duration of completion, we confine to the subsample of completers.

Duration of programme completion

The BSc degree is a 90 credit programme with a minimum duration of 3 academic years. The BTech degree is a 150 credit programme, with a minimum duration of 4 academic years. However, students with lateral entry to the BTech programme (e.g. Diploma and Higher Diploma holders) are generally exempted from Level 3 and some may even receive further exemptions for selective courses at Level 4. Such entrants can complete the BTech degree requirements in three academic years. For ease of reference, here onwards, we simply drop the term “academic” and refer as ‘years.

Table 5 presents the distribution of learners according to the time taken for programme completion.

Considering the higher credit rating for the BTech degree, learners taking longer to complete compared to BSc is not unusual. However, prolonged time for completion evident in both programmes needs attention.

Influence of presage factors on product factors (a)

Next, we present the findings from analysing the presage factors that are likely to have influenced the time taken for programme completion.

Association between age and duration of programme completion

Figure 3 illustrates a plot of cumulative percentage of completers against the number of years taken for completion across different age groups.

As illustrated in Figure 3, the degree programmes remarkably differ with respect to how the time taken for completion vary across different age groups.

In the case of BSc degree, around 2% of 20–24 age group had completed in three years. However, the rest in this age group had taken longer than the below 20 age group. The longest time for completion (up to 9 years) is seen in the 30 to 34 age group and closest to this is the 25 to 29 age group. Majority of 35–39 age group has completed in less time than the 30 to 34 age group. We hypothesize that the 25 to 29 and 30 to 34 age groups have more family and work commitments and thereby have less engagement in studies. A definite answer needs further study.

The picture seen in the BTech degree is different to that of the BSc degree. In the BTech degree, learners in the below 20 age group have taken longer times (up to 12 years) for completion, whereas a large majority in the 35 to 39 age group have completed by 5 years. We note that the BTech degree is a more skills oriented programme. Most of the courses have practical components and in addition, the programme has an industrial training component. The programme allows exemption in the industrial training component, provided they have work experience in a related area. Consequently, such learners can save around six months from the total duration of the programme. The middle age group is likely to have more related work experience compared to the younger age groups. Therefore, students who are employed in a related area and have more work experience (e.g. Learners with work experience in the electrical field and following a degree in electrical engineering) are at an advantage to complete in less time compared to the younger age group.

Another possible cause that needs investigation is the opportunity for lateral entry to the BTech degree. Younger age groups are likely to have joined immediately after schooling, whereas middle age groups are likely to have more diploma holders. This persuaded us to examine whether the completion time varies across learners with different enrolment qualifications. As illustrated in Figures 4 and 5, even among learners with G.C.E. (A/L) as prior enrolment qualification, the association of completion time with age has a similar pattern. Therefore, related work experience is a more likely factor than lateral entry, which causes middle age groups to complete the BTech degree in a shorter time. Further work is needed to test the validity of this hypothesis.

Influence of presage factors on process factors (b)

In examining the presage factors associated with the process factors presented in Table 6, we focused on the entire sample. This included learners who are still at the early stages of the programme. In Table 6, numbers indicated in the brackets give more details about the process factors.

In applying Chi-squared test, to avoid small cell frequencies, we had to combine age groups with the neighboring age groups. Thus, we created the age groups of below 25 years and above 34 years, reducing the number of age groups to 4. The analysis presented in Table 6 considers these four age groups. Likewise, prior ODL experience was categorized into two categories: “having no prior ODL experience” and “with some prior ODL experience”.

The results presented in Table 6 reveal no significant effect of prior ODL experience on any of the process factors considered. However, age and gender have shown significant effects on the process factors. The findings support the observation that for the local scenario, process factors have stronger influence on the product factors, compared to the presage factors considered in the Bigg's 3P model.

Organizing time for studies

How learners have organized time for studies was assessed based on their response to the question on when they have mostly studied, with four options to choose from, which are, “only study near to examinations”, “study as and when time is available”, “organize according to a pre-schedule” and “focus fully on studies”. Only younger age groups had learners who were fully engaged in studies. In the BSc degree, studying only close to examinations were more prominent in the older age groups (40% in the above 34 group and 31.3% in the 30 to 34 age group). In the BTech degree programme, significant gender-wise differences (p-value = 0.05) exist in organizing time for studies. No males as opposed to 4.8% of females had been fully engaged in studies. Also, studying only before examinations was more prominent among males (19.4%) compared to females (14.3%).

Attending academic activities

With respect to attendance for academic activities, significant age-wise differences were evident from both degree programmes. In the BSc degree, poor attendance is most prominent in the 25 to 29 age group (65.2%) followed by 30–34 (50.0%) and above 34 age group (46.5%). On the contrary, in the BTech degree programme, poor attendance was more prominent in the 20 to 24 age group (47.1%) followed by 30–34 (40.0%), 25 to 29 (36.5%), above 34 (33.3%) and below 20 (26.3%).

Using notional hours as a guidance

In the BSc degree programme, none of the presage factors considered were significantly associated with the attention on notional hours. In the BTech degree, only gender has shown significant associations with the attention on notional hours with evidence of male learners indicating less attentiveness on notional hours compared to females. Data show that in the BTech degree, 22.9% of females and 27.6% of males had spent less time compared to notional hours.

We highlight that the associations of age with the learning behaviours are in line with the prolonged time for completion across different age groups (Figure 3). Thus, we postulate that learner engagement in studies is a determinant for the prolonged time for completion and examining this will be left as further work.

Presage factors associated with product factors

Factors that have significant influence on the time duration for programme completion were identified by applying the KW test. The results are presented in Table 7, where we have highlighted the significant factors in bold case. We have used two significant levels. The 5% level is a commonly used significance level for identifying influential factors in studies and providing strong evidence for the effect. The 10% significance level permits identification of influential factors with more power (less Type II error) by compensating for the higher Type I error probability and is suitable when sample sizes are not substantial to detect critical effects. Since this study is a preliminary study to identify suitable input variables for a predictive analytics model that we intend to build as further work, using a 10% significance level is considered justifiable so that important variables have less risk of dropping from the model with no further study.

Identification of influential factors for the time taken for programme completion

The results presented in Table 7 indicate that the age at initial registration, entry qualification and time management skills at the entry point (BSc only) show significant effects on the mean time taken for programme completion. Gender, Knowledge on English, Knowledge on ICT, Presentation and Report writing skills, Teamwork, and Interpersonal skills at entry to the programme and prior experience in ODL through previous learning in a course in the ODL mode have not shown significant effects on the duration of programme completion.

Estimated time for program completion

In both degree programmes, the time durations have skewed distributions. Therefore, we report the median and mode of completion times and not the mean.

In the BSc degree programme, the median and mode of time taken by learners with GCE A/L and A/L with Foundation courses as entry qualification were 6 years and 5 years respectively. The entrants through the OUSL Foundation programme had taken slightly longer time with median and mode 6.5 years and 6 years respectively.

In the BTech degree programme, median and mode of time taken for completion of entrants with GCE A/L qualification were 7 and 6 years respectively. Median and mode of time for completion of entrants with a combination of A/L with Foundation courses, were both equal to 7. Entrants with Diploma/Higher Diploma had taken shorter time for completion with median and mode 5 years and 4 years respectively.

Learning behaviours associated with the duration of programme completion (c)

In studying duration of completion, we confine to programme completers. Under the learning behaviours, we examined whether the learners have organized time for self-studies, attended academic activities and have used notional hours as a guidance in allocating time for studies.

As evident from the results of the KW test presented in Table 8, none of the learning behaviours examined are significantly associated with the duration of programme completion.

Factors associated with the likelihood of programme completion ((a), (c))

One of the weaknesses in this study is the inaccessibility of data pertaining to programme non-completers who are currently inactive (dropouts). While emphasizing the importance of reaching the dropouts, as a feasible solution to identify factors that are associated with the likelihood of programme completion, we compared the learner profiles and the learning behaviours of the two groups who have completed within a specified period (this is described next) and those who have failed to complete within the specified period.

Recall that around 74% of the students had completed the B. Sc. Degree within a period of 6 years. This persuaded us to take 6 years as a reasonable time duration for completion of the BSc degree. To identify influential factors associated with programme completion, we created a binary variable with the two levels “programme completed” and “not completed” within the 6 year period. The sample of BSc degree programme comprised 107 respondents who had spent at least 6 years and out of them 67 students (62.6%) had completed the degree requirements.

On average, BTech students had taken longer times for completion of degree requirements. Around 67% of students had completed the degree within a period of 7 years. For BTech degree, 7 years was taken as a reasonable period for completion. Factors associated with BTech degree completion were identified by focusing on the subsample of students who had spent at least 7 years from initial registration. In the sample of BTech students, there were 66 students who had spent at least 7 years and out of them, 32 students (48.5%) had completed the degree requirements.

Results of applying binary logistic regression on the created dummy variables are presented in Table 9, where we have reported the p-value for comparing the null model and the model with the variable concerned, using Wald test statistic.

From the results presented in Table 9, we conclude that among the learner profile variables (presage factors), knowledge of English has a significant effect on the likelihood of programme completion. Among the learner characteristic variables (process factors), attendance at academic activities had significantly affected the likelihood of programme completion of the BSc degree, within 6 years. In the BTech degree, attendance was not significant even at 10% significance level. The p-value of 0.11, that is close to 10% level, made us speculate whether the sample sizes resulting from cut off at 7 years, was not adequate to detect the effect. Therefore, the analysis was repeated using 6 years as cut-off thus giving bigger sample sizes for the analysis, As expected, attendance then turned out to be a significant factor. Same argument applies to the effect of organizing time for self-studies, where p-values are again small in both programmes. Thus, we identify attendance and organizing time as factors that possibly influence the likelihood of programme completion. This hypothesis needs to be examined through further work with bigger samples.

Conclusion and discussion

This study focused on identifying learner characteristics attributable to prolonged time for completion and likelihood of programme completion in BSc and BTech degree programmes offered by the OUSL. In identifying factors, we closely followed the Bigg's 3P model (Biggs, 1987). Among the presage factors, age at initial registration and time management skills at entry to the programme appeared as important predictors of the time taken for completion of programmes. Further analysis indicated that age is associated with learning behaviours (process factors). Even though the age at initial registration appears as a factor directly influencing time for completion, further work is needed to uncover possible confounding process factors with age that may have led to this finding.

As a generalization of the Bigg's 3P model to the local scenario, we emphasize that the process factors are stronger determinants of the product factors compared to the presage factors. Based on the findings, we recommend empowering students with time management skills and uplifting learner engagement in studies to reduce prolonged time for completion and to improve completion rates.

Course material, past examination papers and support through online tools are the widely used modes of support accessed by the learners in the BSc and BTech Degree programmes of the OUSL. However, we note that there may be many other forms of learner support in the present day (e.g., support via ZOOM interactive sessions, virtual laboratory classes) that may be attractive to the learners that our study has missed out.

Aluwihare and De Silva (2016) have identified prior experience in ODL as an important determinant of prolonged programme completion of BTech graduates. However, this study revealed no significant effect of prior ODL experience on the time taken for completion. Findings presented in Aluwihare and De Silva (2016) were based on descriptive analysis of a sample of 284 respondents drawn from 956 graduates from the Faculty of Engineering Technology of the OUSL, who have graduated during the period from 2006 to 2016. After this period, the university extensively improved learner support in many directions. One such change is the introduction of the Student Academic Readiness and Training Programme (StART@OUSL) for all new entrants, to prepare for the ODL mode. The institutional support was further enhanced through providing supplementary online courses. The inconsistency in the findings of this study with those of Aluwihare and De Silva (2016) could be due to such improvements in learner support and need to be further studied to identify precise reasons.

Despite the short training given at the start, on studying in distance mode, the findings indicate that even at later stages, a considerable percentage of learners' study only near examinations. Attendance for academic activities and using notional hours as a guidance are not significantly associated with the time taken for programme completion. However, poor attendance for academic activities, inadequate organization of time for self-study were identified as influential factors for the likelihood of programme completion. These findings agree with the findings reported by Ali et. al. (2013) that study hours is a significant predictor of academic performance. Our findings also agree with the findings reported by Hassan and Jalila (2017) that reports poor study habits and absenteeism affect academic performance. Thus, the university must look for ways to make learners actively engage in studies throughout. Training only at the entry point seems inadequate and boost up actions are required to enhance learner engagement. Furthermore, learning styles that may demand variations in the learner support needs to be investigated and is left as further work.

Viewing from the perspectives of experiential learning, Kolb (1984) reports that performance, learning, and development form a continuum of adaptive postures to the environment. This study only focused on the time duration and the likelihood of completion of programmes. The data were collected anonymously, and performance levels of the students were not studied. The study necessitates further research on how the identified variables affect the performance levels of the students and more importantly how those affect learning and development as described in Kolb (1984).

The remarkable similarity in the two programmes with respect to learner characteristics attributable to learning lead to believe that the findings reflect real effects and not artifacts of the data. A major limitation of our study is not having access to dropouts. The study yielded several hypotheses that need to be tested through a carefully designed confirmatory study, inclusive of program completers as well as dropouts. We leave this as further work. Finally, we note that the findings of the study are useful in designing future studies and developing predictive models as well as prescriptive models with dash boards for timely identifying customized learner support.

Figures

Figure 1

Conceptual model

Figure 2

Usage of learning resources

Figure 3

Cumulative percentage of completers vs number of years

Figure 4

BSc completion time vs entry qualifications

Figure 5

BTech completion time vs entry qualification

Table 1

Self-assessment on English, ICT and other skills at enrollment

Knowledge/Skills at entry on	Programme	Excellent (%)	Good (%)	Average (%)	Weak (%)
English	BSc	19.0	62.0	18.4	0.6
	BTech	11.3	62.5	25.2	1.9
ICT	BSc	11.8	65.7	21.3	1.2
	BTech	17.9	58.1	22.9	1.0
Presentation/Report writing skills	BSc	13.6	57.6	26.7	2.1
	BTech	10.6	54.8	30.9	3.7
Teamwork skills	BSc	24.2	57.9	17.3	0.6
	BTech	28.6	56.1	14.3	1.0
Interpersonal skills	BSc	21.5	60.3	17.1	1.2
	BTech	21.9	59.5	17.9	0.7
Time management skills	BSc	15.3	59.1	22.5	3.1
	BTech	17.3	54.5	23.2	5.0

Table 2

Organizing self-studies at different levels of study

Level	Programme	Only studied close to exams (%)	Studied as and when time is available (%)	Organized time according to a schedule (%)	Time was fully spent on studies (%)	Total (%)
Level 3	BSc	16.2	41.9	41.9	0.0	100.0
	BTech	9.4	53.1	37.5	0.0	100.0
Level 4	BSc	10.4	44.0	45.6	0.0	100.0
	BTech	17.7	48.4	32.3	1.6	100.0
Level 5	BSc	28.1	42.7	29.2	0.0	100.0
	BTech	20.3	37.8	39.9	2.1	100.0
Degree completers	BSc	23.3	35.6	41.1	0.0	100.0
	BTech	15.6	31.3	51.6	1.6	100.0

Table 3

Attendance for academic activities at different levels of study

		Attendance is
Level of the study	Programme	Regular (%)	Good (%)	Poor (%)	Total (%)
Level 3	BSc	33.9	32.3	33.8	100.0
	BTech	37.5	18.8	43.7	100.0
Level 4	BSc	21.4	42.3	36.3	100.0
	BTech	27.4	38.7	33.9	100.0
Level 5	BSc	18.1	33.2	48.7	100.0
	BTech	31.5	31.5	37.0	100.0
Degree completers	BSc	21.9	41.1	37.0	100.0
	BTech	37.5	43.8	18.7	100.0

Table 4

Time spent on studies compared to notional hours

		Stage in the programme
Time spent with reference to notional hours	Programme	Level 3	Level 4	Level 5	Completed
did not consider notional hours	BSc	12 (19.4%)	33 (18.1%)	44 (22.1%)	19 (26.0%)
	BTech	4 (12.4%)	18 (29.0%)	35 (24.4%)	21 (32.8%)
less time than notional hours	BSc	27 (43.5%)	61 (33.5%)	83 (41.7%)	25 (34.2%)
	BTech	14 (43.8%)	13 (21.0%)	39 (27.3%)	12 (18.8%)
Same time as notional hours	BSc	13 (21.0%)	60 (33.0%)	46 (23.1%)	15 (20.5%)
	BTech	8 (25.0%)	19 (30.6%)	36 (25.2%)	22 (34.4%)
more time than notional hours	BSc	10 (16.1%)	28 (15.4%)	26 (13.1%)	14 (19.3%)
	BTech	6 (18.8%)	12 (19.4%)	33 (23.1%)	9 (14.1%)
Total	BSc	62 (100.0%)	182 (100.0%)	199 (100.0%)	73 (100.0%)
	BTech	32 (100.0%)	62 (100.0%)	143 (100.0%)	64 (100.0%)

Table 5

Time taken for completion by programme

	Programme
Time duration	BSc	BTech
3 years	1 (1.4%)	3 (4.7%)
4 years	10 (13.7%)	6 (9.4%)
5 years	22 (30.1%)	10 (15.6%)
6 years	21 (28.8%)	13 (20.3%)
7 years	7 (9.6%)	11 (17.2%)
more than 7 years	12 (16.4%)	21 (32.8%)

Table 6

Associations of presage factors with process factors

		BSc degree			BTech degree
Process factor	Presage factor	Chi-squared test statistic	df	p-value	Chi-squared test statistic	df	p-value
Organizing self-studies (Table 2)	Age	21.53	9	0.01	8.93	9	0.44
	Prior ODL experience	1.53	3	0.68	1.75	3	0.63
	Gender	2.05	3	0.56	10.35	3	0.02
Attending academic activities (Table 3)	Age	49.94	6	0.00	12.707	6	0.05
	Prior ODL experience	3.12	2	0.79	1.76	2	0.42
	Gender	0.73	2	0.70	3.45	2	0.18
Using notional hours as a guidance (Table 4)	Age	6.91	9	0.65	10.07	9	0.35
	Prior ODL experience	6.58	9	0.68	3.76	3	0.29
	Gender	2.08	3	0.56	0.99	3	0.00

Table 7

Learner profiles associated with duration of programme completion

		BSc degree		BTech degree
Category	Factor	KW test statistic	p-value	KW test statistic	p-value
Learner Profile	Gender	1.19	0.27	0.52	0.47
	Age at initial registration	8.05	0.09*	10.62	0.03**
	Entry qualification	5.72	0.06*	10.11	0.04**
	Knowledge on English	0.28	0.87	4.70	0.20
	Knowledge on ICT	1.19	0.76	4.43	0.22
	Presentation/Report writing skills	0.27	0.88	2.99	0.39
	Teamwork skills at entry	2.50	0.48	1.36	0.51
	Interpersonal skills at entry	1.15	0.77	2.67	0.26
	Time management skills	6.59	0.09*	1.09	0.58
	Prior experience in ODL	2.61	0.11	0.06	0.80

Note(s): * Significant at 10% level; ** significant at 5% level

Table 8

Learning behaviours associated with duration of programme completion

		BSc degree		BTech. Degree
Category	Factor	KW test statistic	p-value	KW test statistic	p-value
Learning behaviour	Organizing time for self-studies	2.38	0.50	1.55	0.67
	Attending academic activities	2.29	0.32	3.60	0.17
	Using notional hours as a guidance	3.16	0.37	2.40	0.40

Table 9

Factors associated with likelihood of programme completion

		BSc degree		BTech degree
Category	Factor	Wald test statistic	p-value	Wald test statistic	p-value
Learner Profile	Gender	0.04	0.84	0.03	0.86
	Age at initial registration	2.03	0.15	1.03	0.31
	Entry qualification	0.30	0.58	0.14	0.71
	Knowledge on English	4.21	0.04**	2.91	0.09*
	Knowledge on ICT	2.46	0.12	0.25	0.62
	Presentation skills	0.24	0.62	0.87	0.35
	Teamwork skills at entry	1.90	0.17	1.68	0.20
	Interpersonal skills at entry	1.80	0.18	0.95	0.33
	Time management skills	0.17	0.68	1.59	0.21
	Prior experience in ODL	0.10	0.75	1.87	0.17
Learningbehaviour	Organizing self-studies	2.33	0.13	2.08	0.15
	Attending academic activities	4.24	0.04**	2.58	0.11
	Using notional hours as a guidance	0.38	0.54	1.19	0.28

Note(s): * Significant at 10% level; ** significant at 5% level

Table A1

Demographic characteristics and prior experience in ODL of the sampled learners in the BSc and BTech. Degree programmes of the Open University of Sri Lanka

Learner characteristics
Description	BSc degree (%)	BTech degree (%)
Gender
Male	22.3	66.5
Female	77.7	33.5
Age at start
Below 20 years	2.5	3.7
20–24 years	66.5	59.5
25–29 years	22.3	28.2
30–34 years	4.3	5.3
35–39 years	2.3	2.3
40 years or more	2.1	1.0
Entry qualification
G.C.E. (A/L)	88.4	53.2
Combination of G.C.E.(A/L) and Foundation	5.0	6.3
Entirely through Foundation programme	4.7	5.3
Diploma	0.0	33.9
other approved qualifications	1.9	1.3
Stream followed at secondary education
Biological Science	78.5	2.3
Physical Science	21.1	94.7
Technology Stream	0.4	3.0
Prior ODL experience
No prior experience	86.4	81.1
had followed a course in ODL mode	13.6	18.9