A mini-dissertation submitted to the Faculty of Education in partial fulfillment of the requirements for the master's degree (Educational Psychology). This study examines the validity of the General Scholastic Aptitude Test as a tool for predicting academic success for first-year information technology (IT) students.
LIST OF TABLES
MOTIVATION FOR THE STUDY TO BE UNDERTAKEN
Over the same period, the view has developed that the predictive value, for academic success at tertiary institutions, of the Grade 12 school certificate has declined (Beecher & Fischer, 1999; Miller & Bradbury, 1999; PE Technikon, 2003; Watson, A, Foxcroft , Seymour, Kock & Streicher, 2001). Higher education is expensive for the student, their family and for the government of the day.
THE PROBLEM STATEMENT
Therefore, sound selection is a critical factor both for the psychological and emotional health of the student, for the financial well-being of the student's family, and for the continued financial support of the institution from the state. Thus, this study aims to determine the validity of the GSAT as a predictive tool for academic success and to consider other factors that may influence the prediction of academic success of a prospective freshman.
RESEARCH AIMS
HYPOTHESES
DEFINITION OF TERMS
THE RESEARCH METHOD
- Research design
- Sampling
- Literature Review
- Instruments
- Scoring and Analysis of Data
THE RESEARCH PLAN
This will include quantitative and qualitative interpretations of the results and conclusions drawn from these results. Findings will be reported in the same sequence as they appear in the study objectives.
INTRODUCTION
STUDIES ON PREDICTIVE VALIDITY AND PSYCHOLOGICAL TESTS (PREDICTION AND SELECTION)
- Prediction and Selection – concepts and reasons
- The predictive validity of psychological tests for academic success – general terms and with special reference to the GSAT
- Psychological tests, academic success and racial differences
- Psychological tests, academic success and gender differences
- Other predictors of academic success (SR & English ability) and their correlation with psychological tests
- Selection procedures around the World
- Selection procedures in South Africa
- Predictors of academic success for IT students – with a focus on psychological tests
In their opinion, with increased diversity in the South African system and in light of the changing education system. Ayaya (1996), Greyling (2000), Miller and Bradbury (1999) and Venter (1995) are of the opinion that students' ability with the English language may be the key when the school leaving results are to be used.
A PREDICTION AND SELECTION MODEL FOR PORT ELIZABTH TECHNIKON (PET)
In addition, many of the students who emerge from such environments have low academic self-confidence (Greyling 2000:20). In addition to the GSAT, students are assessed for English comprehension on subtest 4 of the Academic Aptitude Test (AAT).
THE GENERAL SCHOLASTIC APTITUDE TEST (GSAT)
- Background
- Use at PET
Thus, provided they are not used as an intelligence test, the GSAT has a valid application as a measure of scholastic ability and therefore may yield useful results for predicting scholastic success in the entire population. Regarding the use of the GSAT in PE Teknikon, it is used for the purpose of assessing scholastic abilities for all population groups.
DIRECTION OF THIS RESEARCH
INTRODUCTION
THE RESEARCH METHOD
- Research design
- Sampling
- Instruments
- Scoring and Analysis of Data
A full academic load consisted of four subjects, two of which were offered per semester, but where the student got the average of the two semesters. A successful student was defined as a student who had passed 3 or 4 of the first year subjects in which they had enrolled and thus would be able to enroll in all or most of their second year subjects.
EXPLANATION OF THE USE OF THE MEASURING INSTRUMENTS
- The General Scholastic Aptitude Test (GSAT)
- The Swedish Rating System (SRS)
- The Academic Aptitude Test (AAT) English Comprehension (subtest 4)
The AAT is a measure of the student's English comprehension ability and is only used in cases where both the SR and GSAT provide borderline assessments. In such cases, where the student appears to have a high level of English proficiency, they may be given the opportunity to enroll in the course (Naude, 1995).
THE TESTING PROCEDURE
THE STATISTICAL TECHNIQUES
After establishing these relationships, the strength of the correlations was examined by means of a Spearman rank correlation to determine if any of these relationships were statistically significant. Finally, a series of logistic regressions was performed in an attempt to determine whether any of the variables played a significant role in predicting academic success.
INTRODUCTION
DESCRIPTIVE STATISTICS
- The relationship between GSAT and the number of subjects passed .1 The dichotomized variable
- The original variable
- Racial and gender groups
- The relationship between Swedish Rating (SR) and the number of subjects passed
- The dichotomized variable
- The original variable
- Racial and gender groups
Upon further examination of the histograms (Figure 4.2), about 19% of the successful group of students fall below the GSAT cutoff. A comparison of the GSAT cut-off means and the number of subjects who passed, significant at p<0.05. The analysis of variance (Table 4.21) and the Levene test for homogeneity of variance (Table 4.22) confirm that the GSAT cutoff mean for successful female students is.
Analysis of variance (Table 4.24) confirms that the average SR of the successful group of students is statistically significantly different from the unsuccessful group at the 0.05 significance level. Comparison of mean Swedish grade and number of subjects completed, significant at p<0.05. Analysis of variance (Table 4.28) confirms that the average SR of the successful group of students is statistically significantly different from the unsuccessful group at the 0.05 significance level.
The Analysis of Variance (Table 4.31) and the Levene Test of Homogeneity (Table 4.32) confirm that the means of the passing and failing students do not differ significantly at the 0.05 level of significance. The analysis of variance (Table 4.34) and the Levene test of homogeneity (Table 4.35) confirm that the SR mean of the passing and failing students is not statistically significant at the significance level of 0.05. The Analysis of Variance (Table 4.37) and the Levene Test of Homogeneity (Table 4.38) confirm that the SR means of the successful and unsuccessful groups are statistically different at the significance level of 0.05.
STRENGTH OF ASSOCIATIONS
LOGISTIC REGRESSIONS
Only two variables entered the model, first the Swedish score, followed by T, which is the total GSAT score and not just the amount above the cutoff, see Table 4.47 on p. Finally, a series of logistic regressions were performed to show the effect of GSAT and Swedish scores as the only predictors of student achievement. From the above, it appears that the Swedish grade has a higher correlation with performance and thus in predicting academic success.
However, since the GSAT and the Swedish assessment have different ranges, it might not be fair to apply a 5-unit increase to both. 1 To determine the validity of the GSAT as a predictor of academic success for first-year IT students. For white students, there is a statistically significant correlation (at p<0.05) between GSAT and academic performance in the first year of the IT class - see Tables 4.8;.
For the black and colored students the correlations were positive but not statistically significant at p<0.05 and therefore the GSAT cannot be considered as. There is a statistically significant positive correlation (at p<0.05) between GSAT and Swedish rating, although Swedish rating appears to be the strongest. No statistically significant positive correlation (at p<0.05) was found between English proficiency and academic success in the first-year IT class - see table 4.47.
FINDINGS
- Findings with regard to the predictive validity of the GSAT in relation to the academic success of first year IT students at PE Technikon
- Findings with regard to whether the GSAT is a good predictor of academic success for the various racial groups in South Africa
- Findings with regard to whether the GSAT is a good predictor of academic success for the gender groups at PET
- Findings with regard to whether the GSAT correlates well with
- The Swedish Rating System
- English language ability
This again suggests that perhaps the test is identifying students with potential, but that there are other factors not addressed in the selection and teaching and learning process that play an important role in students' academic success. However, in terms of research intent, this confound means that the GSAT is not a good predictor of academic success for all racial groups in South Africa. For both male and female students in the total sample there is a statistically significant but weak positive correlation between GSAT and academic success, which is consistent with the literature (Barro, 2001; Coley, 2001; Huysamen & Roozendaal, 1999; Greyling, 2000 ).
This correlation is stronger for male than female students and suggests that the GSAT plays an important role in predicting academic success for both male and female students. Both the Swedish Rating (SR) and the GSAT were found to be statistically significant, albeit at a rather low level, in terms of predicting academic success. When stepwise logistic regression was performed, these were the only two factors that were statistically significant predictors of academic success.
In other words, there appears to be a correlation between GSAT and SR in terms of predicting academic success, but in the current education system, SR appears to be a better predictor. English language proficiency did not appear to play a statistically significant role in this study, and there appeared to be no significant relationship between level of English comprehension, GSAT, and the prediction of academic success. Ayaya (1996), Greyling (2000), Miller and Bradbury (1999) and Venter (1995) suggest that English proficiency plays an important role in predicting academic success, but that it may be secondary to other factors.
LIMITATIONS OF THIS STUDY
This lack of success may not be due to assessment limitations, but rather other factors that stand in the way of academic success.
RECOMMENDATIONS FOR THE SELECTION POLICY OF IT STUDENTS AND THE USE OF THE GSAT IN THE BROADER SELECTION POLICY OF THE PE
In other words for both the IT department and all other departments within the institution it would seem that it would be wise to use such assessments, but use them with discretion, this is especially true for black, colored and female students. They should not be used in isolation, but rather in conjunction with other available information in terms of selection and placement of students. Furthermore, the process of selection and placement should not end after registration, but should be supported by other academic and social programs to support the student as acceptance into a program raises an expectation of academic success (Zaaiman, van der Flier & Thijs , 2000.
AVENUES OF FURTHER RESEARCH
An additional area that requires examination in the field of selection is the role that subject selection at the high school level plays in increasing success in the IT field of study. It appears that certain subjects may better prepare the student for the disciplines required to study IT successfully, and identifying these subjects may improve the selection process. It is therefore necessary that the tertiary institutions begin to address the issues likely to arise in this regard and research and design appropriate assessments that can be used in the selection and placement of students coming from such educational backgrounds.
The prediction of academic achievement in the first year: a case study at the National University of Lesotho. Diversity: Predicting academic performance of first-year students at Potchefstroom College of Agriculture. Curricular choice and the differential prediction of male and female tertiary academic achievement.
Psycho-social factors and academic performance among African women in a predominantly white university in South Africa. Selection of South African students with advantages and disadvantages for university admission South African Journal of Higher Education. The Use of Psychometric Instruments in Predicting Success in First-Year Computer Science and Information Systems Courses.
Analysis of GSAT data
ANNEXURE 2
