CHAPTER 6: CONCLUSIONS AND POLICY IMPLICATIONS
6.1. Conclusions
6.1.2. Regression analyses: OLS and Logistic
In terms of statistical inference, the study followed closely the steps of existing studies that investigated the determinants of student achievement. As discussed in the methodology section, the educational production function was to be estimated in terms of students’ characteristics alone: the characteristics of education institutions (Iij) and the characteristics of academic/non-academic staff members (Aij) were treated as constant.
Theoretically, a regression analysis estimates or predicts the average value of the dependent variable on the basis of the fixed values of the explanatory variables. The results of regression analysis consolidated in Table 4-32 established the relationship between students’ performance at the intake level and their matric scores at the school leaving level. These results confirm the correlations sweep findings that students generally do better in first-year accounting and economics modules if they have been exposed to these subjects at high school. Characteristics of students that include inter alia total matric points, proficiency in Maths and English I, English as home first language, and other designated individual matric subjects scores
178 (such as matric Accounting and Economics scores) were good predictors of student performance in first- year accounting and economics modules.
Other personal characteristics and student demographics also play a role. For example, the age of the student has negative causal effects on student performance in first-year accounting and economics modules. That is, the predicted final examination marks for younger students would be higher than for older students, holding all other variables constant. Since the national average age at which students complete their matric in South Africa is between 17 and 20, it may be assumed that an older student or returning student of above 25 years of age at the point of university entrance would more likely not perform well or achieve slower progress than their peers of between 17 and 20 who just entered the higher education. Underperformance might have been caused by the fact that the students repeated some years in primary and secondary education (delayed educational career). The student may also have taken a significant break after completing secondary education, or may not have studied for a period of time (educational career break or a gap year). It might be because the student followed an alternative or longer educational route (for example enrolled via an access/foundational or extended programme) before attending university. Therefore, studying at a later age can be assumed to have a negative causal effect on Student academic performance.
Empirical results also give support to a marked improved in pass rates when students are selected and admitted strictly on the basis of total matric points of 36. Total matric points (or APS) and selected designated matric subject scores including mathematical knowledge and prior knowledge of accounting, economics, and English are good predictors of student performance in undergraduate accountancy and economics modules in the College at the upper end. Students admitted through Dean’s discretion or the total matric points (or APS) requirements for the BAdmin or BBus Admin, or in access programmes which require less than the total matric points of 36 when placed in the BCom (Accounting) and BCom (General) degree curricular routes are less likely to be successful. These students exhibit a lower level of academic ability as measured by their performance in the matriculation examination and are expected to have poor success rates in their university study. These students compare unfavourably with their peers. This demands the development of a new curriculum for these students or their placement into appropriate existing curricular routes different from the BCom (Accounting) and BCom (General) degree, where they can be more likely to succeed. This is a problem that the advocates of multidisciplinary curricula have to address as it is a challenge to offer the same standard of accounting and economics modules to everyone.
These findings are in line with South African and international studies.
179 From this econometric analysis, it emerges that weak predictors of student performance are total matric points, matric Maths score and matric English score, students who have English as home first language, and non-designated matric subjects scores that include matric Accounting score and matric Economics score.
However, the estimation of the parameters of the econometric model indicated that the variables included in the regression analyses play only a limited role in predicting the variance of student performance in the CLMS. This is evident from some of the pseudo R2 and R2 which are as low as 2 percent pointing out low explanatory power. The empirical results from the OLS and Logistic regression analyses reinforced the findings of the pair-wise correlations sweep. Determinants of students’ academic performance are ambiguous as they are not straightforward measures of student quality, making the prediction of student performance a far more complex process.
As cautioned in the methodology discussed in Chapter 3, a certain proportion of students’ academic performance is determined by the one or two predictors variables in this thesis’ model but as for the rest of the causation – this resides in the mystery of the other factors that either were not incorporated in the model and thus not counted (such as some of those mentioned in the findings of the focus groups) or even conceptualized in the “African belief systems” (Spirit of Ubuntu, cleansing or ancestors’ prayers; divine and fasting prayers; superstition, guardian angels’ wings or lucky charms). These African belief systems affect the attitude of the student and this in turn affects his or her academic performance. For example, it was widely hypothesized that higher failure rates in ACCT200 over the years result from the practice of giving prior learning credit for ACCT101 and ACCT102 to students transferring from other (non-accredited) HEIs.
These findings add a new dimension to the existing puzzle in the CLMS.
Other variables such as the context, the characteristics of UKZN and its institutional climate, environment, policies and services to students (Cohen et al., 2009), socio-economic background, student demographics, intellectual leadership, proper learning infrastructure, motivational and psychological attitudes, the characteristics of academic and non-academic staff members (administrators and support staff), amongst others are equally important in determining Student performance. The context can encompass several academic, financial, social and other non-university related explanations that can broadly be classified in four categories that include inter alia: (1) late registration (often after an appeal process that runs till late in the semester), and failure in the first test which knocks the students’ confidence and minimize their chances of their best marks being considered for the duly performed (DP) certificate, (2) absenteeism or sporadic attendance of lectures, while trying to get their heads around managing the time table, (3) lack of prescribed textbooks, and (4) lack of financial resources. There is a long list of other possible factors such as
180 university choices and behavior; class attendance, laziness or over-sleeping, partying, study location and commuting distance, employment status, interaction with academic and non-academic staff members and with peers among others, that could also influence student success.
Evidence to support these findings is drawn from a variety of sources. Guney (2009) examined designated endogenous and exogenous factors to search for determinants of students’ academic performance and identified lecturers, assessment, teaching material, students with better numeracy backgrounds, attendance, work experience, future career, degree course, age, and the ability to perform better in accounting amongst the endogenous factors. Exogenous factors such as learning disability, part-time work, and personal problems cause students to lose concentration and therefore underperform in accounting. McPherson (1993) recommended a system of indicators, rather than a single indicator, taking into account that value-added indicators may fail to factor in the fact that educational institutions may have a differential effect on the performance of different types of student.
These results suggest that admission eligibility consider additional mechanisms in the selection of candidates and their placement into appropriate curricular routes where they are more likely to succeed.
HEIs in South Africa are becoming innovative in screening their would-be students and placing them into appropriate curricular routes. Stakeholders at UKZN have to explore and identify the characteristics that are enhancing student achievement in the College coupled with other contextual variables in a quest to improve the pass rates and throughput rates, since this study has found them to be equally or more important determinants of student success.