Aptitude Testing for University Entrance

Aptitude Testing for University Entrance: A Literature Review

Angus S. McDonald Paul E. Newton Chris Whetton

Pauline Benefield

Contents

List of Tables and List of Figures

Acknowledgements

Executive Summary

5

1.

Overview

8

2.

The use of aptitude testing in university entrance

10

2.1 Introduction

10

2.2 United States of America

10

2.2.1 The Scholastic Aptitude (or Assessment) Test

10

2.2.2 The American College Test

12

2.2.3 Use of test scores by institutions

13

2.2.4 Use of test scores by students

15

2.3 Aptitude testing in Israel

16

2.3.1 The Psychometric Entrance Test (PET)

17

2.4 Aptitude testing in Sweden

17

2.4.1 The SweSAT

18

2.5 Aptitude testing in Singapore

18

3.

Predicting college performance

20

3.1 Introduction

20

3.2 The predictive validity of the SAT and other indices of

20

performance

3.2.1 Issues to consider in assessing predictive

20

validity

3.2.2 Evidence on predictive validity

21

3.2.3 Do aptitude tests tell us any more than we

24

already know?

3.3 Is aptitude testing fair?

26

3.3.1 The concept of test bias

27

3.3.2 Evidence of score differences

28

3.3.3 SAT scores, academic attainment and scores

31

on test measuring similar constructs

3.3.4 Item-level analysis of SAT data

33

3.3.5 Does the evidence suggest the SAT is biased?

35

3.3.5.1 Sex differences

35

3.3.5.2 Ethnic differences

37

3.4 Do test scores predict fairly for all groups?

39

3.5 Discussion

41

4.

Consequences of aptitude testing

45

4.1 Introduction

45

4.2 Test preparation and coaching

45

4.3 Aptitude testing and educational opportunities

49

4.4 What would happen if the SAT were abolished?

51

4.5 Discussion

52

5.

Research in Britain

54

5.1 Introduction

54

5.2 Predicting success in British universities

54

5.2.1 Previous research conducted by the NFER

54

5.2.2 Research on A-levels and other tests

58

5.2.3 Summary and issues for consideration

61

5.3 Current developments in British universities

63

5.3.1 Dr Jane Mellanby, Department of

63

Experimental Psychology, University of Oxford

5.3.1.1 Background

63

5.3.1.2 Trial test and initial findings

63

5.3.1.3 Planned research

65

5.3.1.4 Potential use of the test and other issues

65

5.3.2 Professor Michael Worton, Vice-Provost,

66

University College London

5.3.2.1 Background

66

5.3.2.2 Trial tests and planned research

66

5.3.2.3 Application of the tests

67

5.3.2.4 Outstanding issues

67

5.3.3 Professor Dylan Wiliam, Head of the School of

68

Education, King's College London

5.3.3.1 Background

68

5.3.3.2 Testing for Medical School and

68

enhancing access to medicine

5.3.3.3 Planned research and developments in

69

access to medicine

6.

Conclusions

71

6.1 Reasons for interest in the SAT

71

6.2 Aptitude testing in the British education system

72

6.3 The consequences of aptitude testing

73

References

76

Appendix 1

Methodology for the Review

83

List of Tables

3.1

Correlations of first-year GPA with SAT and HSGPA

21

3.2

Percentage of variance in first-year GPA accounted for by

25

HSGPA and SAT

3.3

2000 average SAT scores by sex and ethnic status

28

3.4

2000 average ACT scores by sex and ethnic status

30

List of Figures

3.1

Possible relationships between the SAT, HSGPA

24

and college GPA

Note: The SAT I: Reasoning Test has two elements, a verbal section and a 'math' section. Since this American English term is that used in the test itself, it has been referred to in this way throughout this report.

The authors would like to thank Liz Gibson for her careful work in producing the final manuscript.

p. 5

Executive Summary

1. This report presents the results of a literature review on aptitude testing for university entrance. The work was commissioned by The Sutton Trust and undertaken by the National Foundation for Educational Research.

2. Evidence for the review was obtained by searching a number of electronic databases and the websites of the main organisations involved in admissions testing in the United States. The central areas covered in this review are an overview of admissions procedures in countries which use aptitude tests, the extent to which aptitude tests are fair predictors of university performance, and the effects aptitude testing may have on the wider education system. Research on the prediction of university performance in Britain is also reviewed, and the findings reported from interviews conducted at three universities currently developing additional ways of assessing applicants.

3. Probably the best known admissions test is the Scholastic Aptitude Test (SAT), which was introduced by the College Board in 1926 with the goal of standardising the admissions process to universities and colleges in the United States. In its current form, the SAT assesses verbal and math ability, mainly through multiple-choice items. Admissions tests in countries such as Israel and Sweden follow a similar format, and evidence suggests that they assess the same underlying constructs as the SAT.

4. Many colleges in the United States use SAT scores in conjunction with high school record as part of their selection process. However, there is considerable variation in practices, with some colleges not using admissions test scores for some or even any of their students.

5. High school record and admissions test scores are able to predict college grades to a modest degree. Of the two, high school record is generally the better predictor, and once this has been taken into account the additional ability of the SAT to predict performance is limited.

6. There has been considerable controversy over the fairness of tests such as the SAT, particularly in relation to sex and ethnic score differences. This review considers a range of evidence to reach a judgement on test fairness. Evidence suggests that both the verbal and math sections of the SAT are somewhat biased in favour of males, although it cannot be ruled out that this is due to male SAT takers being a more highly selected group. Evidence for students from ethnic minorities is somewhat less conclusive. Although groups such as African Americans score about one standard deviation lower on the SAT than Whites, this pattern is repeated on virtually all tests which measure aspects of intelligence. Even when overt attempts have been made to equalise the scores of Blacks and Whites, Blacks are still seen to score lower. The reasons behind this are not fully understood.

7. The question of test fairness is probably most adequately addressed through an examination of how accurate admissions test scores are at predicting college attainment for different groups. It is clear that test scores do not always equate to attained college grades, but as the SAT accounts for a relatively small proportion of the variance in college grades, this is not surprising. There is consistent evidence that the SAT under-predicts female attainment, even after differential selection of courses for grading difficulty is

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download