Admission Groups and Academic Performance: A Study of Marginal Entrants in the Selection to Higher Education

Linn Karlsson; Magnus Wikström

doi:10.1515/bejeap-2020-0434

Published by De Gruyter January 4, 2022

Admission Groups and Academic Performance: A Study of Marginal Entrants in the Selection to Higher Education

Linn Karlsson and Magnus Wikström

From the journal The B.E. Journal of Economic Analysis & Policy

https://doi.org/10.1515/bejeap-2020-0434

Showing a limited preview of this publication:

Abstract

The purpose of this paper is to study whether Swedish admission policies successful in selecting the best-performing students. The Swedish universities select students based on two different instruments, which each form a separate admission group. A regression model is recommended to estimate the achievement differences for the marginally accepted students between the admission groups and is applied to a sample of 9024 Swedish university entrants in four different fields of education. Marginally accepted students in the group selected by school grades on average perform better than students accepted by an admission test, suggesting that a small reallocation of study positions towards the grade admission group may increase overall academic achievement. However, the achievement difference appears to vary concerning university programme selectivity. We found that increasing selection by grades in less competitive programmes would improve overall achievement, while we do not find any effect for highly competitive programmes.

Keywords: higher education; selection; admission groups; marginal admit; achievement

JEL Classification: I21; I23; I28

Corresponding author: Linn Karlsson, Department of Economics, Umeå School of Business, Economics and Statistics, Umeå university, Samhällsvetarhuset, Bibliotekgränd 6, A24407, 901 87, Umeå, Sweden, E-mail: linn.karlsson@umu.se

Funding source: Handelsbanken

Award Identifier / Grant number: P2016-0299

Funding source: Vetenskapsrådet

Award Identifier / Grant number: 2014-01990

Acknowledgements

The authors would like to thank David Granlund, Gauthier Lanot, Alejandro Vega, two anonymous referees, seminar participants at the LEER conference in Leuven, 2019, and the SweSAT conference in Umeå, 2019, for valuable comments. Research grants from the Swedish Research Council (ref. 2014-01990) and Handelsbanken (ref. P2016-0299) are gratefully acknowledged.

Competing interests: The Authors declare that there is no conflict of interest.

Appendix

The following figures illustrates the proportions of students in relation to the cut-off for each field of education. The x-axis shows the functions f and h for each student (see Eq. (1)). To make the upper secondary school GPA and SweSAT comparable, the variables s and g are defined as the relative distances between the individuals’ scores and the minimum scores in the programmes. See page 15 for a detailed description of the equation.

Figure 2:

The proportion of students and distance to cut-off: pooled sample.

Figure 3:

The proportion of students and distance to the cut-off: engineering.

Figure 4:

The proportion of students and distance to the cut-off: business and economics.

Figure 5:

The proportion of students and distance to the cut-off: social work.

Figure 6:

The proportion of students and distance to the cut-off: law.

Figure 7:

Distribution of admission scores in pooled sample unconditioned on admission group.

Table 6:

List of variables available in the datasets.

Variable name	Description	Source
Student ID	Student’s ID	The national education records, 2012/2013
Age	Student’s age in the year of 2012	The national education records, 2012/2013
Gender (woman)	A dummy variable taking the value one if	The national education records,
	the student is female	2012/2013
Immigrant background	A dummy variable taking the value one if	The national education
	at least one of the student’s parents is born	records, 2012/2013
	outside Sweden
Parental highest education	Dummy variables taking the value of one	The national
	if parent’s highest education level 9 years	education records,
	of schooling; two if parent’s highest	2012/2013
	educational level is at upper secondary
	level or three if parent’s highest
	educational level is at post-secondary level
GPA	Student’s upper secondary school GPA	The national
		education records, 2012/2013
SweSAT	Student’s SweSAT score	The national education records,
		2012/2013
Admissions group	Dummy variables indicating the student’s	The national
	admissions group. Students admitted by	education records,
	other instruments than GPA or SweSAT	2012/2013
	are excluded
University ID	ID for each university	The national education records,
		2012/2013
Programme ID	ID for each programme	The national education records,
		2012/2013
Completed credits	Student’s completed credits in the	The national
	academic year of 2012/2013. Students	education records,
	with zero completed credits are excluded	2012/2013
	from the analysis
Enrolled credits	Student’s enrolled credits in the academic	The national education records,
	year of 2012/2013	2012/2013
First-choice applicants	The number of first-choice applicants per	National
	programme	admissions records,
		UHR, 2012
Number of admitted	The number of admitted students per	National
applicants	programme. Since all programmes in this	admissions records,
	study are oversubscribed, the variable	UHR, 2012
	serves as a proxy of available spots
Transformed variables
Achievement	Achievement is defined as the fraction of
	completed credits over enrolled credits
	s ̃	s ̃ is defined as the relative distance from
	the student’s SweSAT to the minimum
	score in the programme. i.e. the
	admissions score from the last admitted
	student subtracted from the student’s
	admission score divided by the range
	between the highest and lowest admissions
	score
g ̃	g ̃ is defined as the relative distance from
	the student’s GPA to the minimum score in
	the programme
FIELD	Dummy variables for each field
	(Engineering, Business and Economics,
	Law and Social Work)
UNIVFIELD	Dummy variables for each university and
	field
Competition/Applicantss	The fraction of first-choice applicants over
	the number of admitted applicants
Competition: Low	A dummy variable taking the value of one
	if competition is less than two first-choice
	applicants per study position
Competition: Medium	A dummy variable taking the value of one
	if competition is more than two but more
	than four first-choice applicants per study
	position
Competition: High	A dummy variable taking the value of one
	if competition is more than four first-
	choice applicants per study position

The data were collected by Statistics Sweden and consist of several datasets. The variables parental highest education and immigrant background have been modified to be consistent in the pooled sample.

Table 7:

Marginal admission score by university.

Higher Education Institution	Business and Economics		Engineering
	GPA	SweSAT	GPA	SweSAT
Umeå University	15.2	10	10.4	1
Luleå University of technology	14.6	9	9.9	6
Uppsala University	18.7	15	11.2	6
University of Gävle	13.6	7
Dalarna University	13.4	2
Mälardalens University	14.6	7	11.2	1
KTH			12.8	11
Örebro University	16.8	11
Stockholm University
Linköping University	15.3	10	11.7	5
Jönköping University
Chalmers University of Technology			15.5	12
University of Gothenburg	17.2	13
Karlstad University	15.4	9	13	4
University of Skövde	12.8	5
University of Borås	15.8	11
Lund University	18.3	13	15.9	12
Halmstad University	15.78	11
Stockholm School of Economics	19.5	19
University West	14.32	6
Blekinge Institute of Technology			11.2	4
Mittuniversitetet	11.9	1	11.7	7
Malmö University
Ersta Sköndal University College
University of Gotland (now Uppsala)	12.3	2
Linnaeus University	15.4	9
Higher Education Institution	Law		Social Work
	GPA	SweSAT	GPA	SweSAT
Umeå University	18	13	13.8	5
Luleå University of Technology
Uppsala University	19	15	17.5	11
University of Gävle			12.2	3
Dalarna University			15.2	8
Mälardalens University			15.3	7
KTH
Örebro University	18.5	13	15.9	8
Stockholm University	18.2	15	15.5	9
Linköping University			16.1	9
Jönköping University			16.2	10
Chalmers University of Technology
University of Gothenburg	18.5	15	17.1	2
Karlstad University			16	8
University of Skövde
University of Borås
Lund University	19.0	15	15.7	8
Halmstad University
Stockholm School of Economics
University West
Blekinge Instititute of Technology
Mittuniversitetet			14.7	6
Malmö University			15.5	8
Ersta Sköndal University College			16.9	14
University of Gotland (now Uppsala)
Linnaeus University			15.8	7

The marginal admission score is defined as the smallest GPA or SweSAT student score of all offered programmes within each university and fields of education.

Table 8:

Summary statistics by admission group.

	Upper secondary GPA			SweSAT
	N	Mean	SD	N	Mean	SD
Pooled
Age	5379	20.07	1.57	3645	22.22	4.19
Gender (woman)	5379	0.53	0.50	3645	0.30	0.46
Foreign background	5379	0.14	0.35	3645	0.09	0.29
Parental highest education
Less than upper secondary schooling	5372	0.02	0.12	3627	0.02	0.15
Upper secondary schooling	5372	0.35	0.48	3627	0.34	0.47
More than upper secondary schooling	5372	0.64	0.48	3627	0.64	0.48
Engineering
Age	2859	19.75	1.14	1775	20.97	2.69
Gender (woman)	2859	0.35	0.48	1775	0.16	0.37
Foreign background	2859	0.15	0.36	1775	0.10	0.29
Parental highest education
Less than upper secondary schooling	2855	0.02	0.13	1767	0.01	0.10
Upper secondary schooling	2855	0.29	0.45	1767	0.30	0.46
More than upper secondary schooling	2855	0.69	0.46	1767	0.69	0.46
Business and Economics
Age	1364	20.28	1.78	965	22.37	3.87
Gender (woman)	1364	0.63	0.48	965	0.31	0.46
Foreign background	1364	0.12	0.32	965	0.09	0.29
Parental highest education
Less than upper secondary schooling	1362	0.01	0.12	963	0.03	0.18
Upper secondary schooling	1362	0.42	0.49	963	0.38	0.49
More than upper secondary schooling	1362	0.56	0.50	963	0.59	0.49
Social work
Age	658	20.92	2.26	533	25.65	6.41
Gender (woman)	658	0.94	0.24	533	0.73	0.45
Foreign background	658	0.16	0.37	533	0.11	0.31
Parental highest education
Less than upper secondary schooling	658	0.02	0.12	527	0.07	0.25
Upper secondary schooling	658	0.49	0.50	527	0.46	0.50
More than upper secondary schooling	658	0.49	0.50	527	0.47	0.50
Law
Age	498	20.21	1.43	372	22.84	3.94
Gender (woman)	498	0.71	0.45	372	0.31	0.46
Foreign background	498	0.10	0.30	372	0.06	0.24
Parental highest education
Less than upper secondary schooling	497	0.01	0.10	370	0.01	0.07
Upper secondary schooling	497	0.26	0.44	370	0.24	0.42
More than upper secondary schooling	497	0.73	0.45	370	0.76	0.43

Summary statistics of socioeconomic background by admission group. Parental highest education is a categorical variable which indicates the education level of the parent with the highest education.

Table 9:

Gender-specific regressions.

Full sample	Women	Men
Pooled	−0.0147	−0.0276
	(0.0176)	(0.0175)
N	3909	5106
R-squared	0.175	0.127
Engineering	−0.0134	−0.0425^*
	(0.0311)	(0.0217)
N	1280	3351
R-squared	0.144	0.104
Bus. And Econ.	−0.0729^**	0.00414
	(0.0338)	(0.0365)
N	1155	1173
R-squared	0.125	0.111
Social work	0.00302	−0.189^***
	(0.0302)	(0.0636)
N	1003	183
R-squared	0.087	0.177
Law	0.0289	−0.00952
	(0.0486)	(0.0481)
N	471	399
R-squared	0.140	0.129

Parameters of the term DIFFG for women and men are displayed in the table. The estimates correspond to model 1 in Table 4. Standard errors are clustered at programme level and reported within parentheses. Significance levels are denoted by asterisks, where ^*** is significant at the 1 per cent level, ^** at the 5 per cent level and ^* at the 10 per cent level. The order of the polynomials f(^.) and h(^.) is K = 2. Field fixed effects and university fixed effects are included in the estimation.

Table 10:

Marginal achievement differences; background variables included.

	Model 1	Model 2	Model 3
	(Constant)	(Quadratic)	(Categorical)
Intercept	−0.030^**	−0.076^***
	(0.015)	(0.027)
Gender (woman)	0.044^***	0.045^***	0.045^***
	(0.007)	(0.007)	(0.007)
Age	0.001	0.001	0.001
	(0.001)	(0.001)	(0.001)
Foreign background	−0.054^***	−0.054^***	−0.054^***
	(0.008)	(0.008)	(0.008)
Parental highest education
Upper secondary schooling	−0.024	−0.024	−0.024
	(0.015)	(0.015)	(0.015)
More than upper secondary schooling	−0.010	−0.010	−0.010
	(0.015)	(0.015)	(0.015)
Competition		0.030*
		(0.017)
Competition squared		−0.003
		(0.002)
Competition: Low			−0.046^***
			(0.017)
Competition: Medium			−0.012
			(0.017)
Competition: High			−0.002
			(0.016)

Observations	8999	8999	8999
R-squared	0.177	0.178	0.179

Parameters of the term DIFFG are displayed in the table. Standard errors are clustered at programme level and reported within parentheses. Significance levels are denoted by asterisks, where ^*** is significant at the 1 per cent level, ^** at the 5 per cent level and ^* at the 10 per cent level. The order of the polynomials f(^.) and h(^.) is K = 2. Field fixed effects and university fixed effects are included in the estimation. The number of observations is 8999 (full sample) instead of 9024 because of missing values.

Table 11:

Marginal achievement differences; low admission scores excluded.

	Model 1	Model 3
	(Constant)	(Categorical)
	Intercept	Low	Medium	High
A. Full sample
Pooled data	−0.045^**	−0.066^***	−0.024	−0.010
(N = 6390)	(0.018)	(0.021)	(0.020)	(0.019)
Engineering	−0.052^**	–	–	–
(N = 4634)	(0.023)
Business and Economics	−0.055*	–	–	–
(N = 2329)	(0.031)
Social Work	0.003	–	–	–
(N = 1191)	(0.041)
Law	0.029	–	–	–
(N = 870)	(-0.057)
B. Restricted sample
Pooled data	−0.049^**	−0.088^***	−0.015	0.033
(N = 1901)	(0.021)	(0.025)	(0.028)	(0.036)
Engineering	−0.074^**	–	–	–
(N = 1369)	(0.029)
Business and Economics	−0.051	–	–	–
(N = 652)	(0.041)
Social Work	0.055	–	–	–
(N = 331)	(0.075)
Law	−0.006	–	–	–
(N = 298)	(0.027)

Parameters of the term DIFFG are displayed in the table. Programmes with GPA < 15 and SweSAT < 10 are excluded. Standard errors are clustered at programme level and reported within parentheses. Significance levels are denoted by asterisks, where ^*** is significant at the 1 per cent level, ^** at the 5 per cent level and ^* at the 10 per cent level. The order of the polynomials f(^.) and h(^.) is K = 2. Field fixed effects are included in the estimation.

Table 12:

Tabulation of competition and field of education.

Field of Education	Competition
	Low	Medium	High	Total
Engineering	3898 (43.2%)	601 (6.7%)	135 (1.5%)	4634 (51.4%)
Business and economics	1444 (16%)	391 (4.3%)	494 (5.5%)	2329 (25.8%)
Social work	170 (1.9%)	616 (6.83%)	405 (4.5%)	1191 (13.2%)
Law	0 (0%)	177 (2%)	693 (7.7%)	870 (9.6%)
Total	5512 (61.1%)	1785 (19.8%)	1727 (19.1%)	9024 (100%)

The table reports the number of students in each category of competition. The percentages in the parenthesis show the share of the total sample.

References

Aguinis, H., S. A. Culpepper, and C. A. Pierce. 2016. “Differential Prediction Generalization in College Admissions Testing.” Journal of Educational Psychology 108 (7): 1045. https://doi.org/10.1037/edu0000104.Search in Google Scholar

Akhtari, M., N. Bau, and J.-W. Laliberte. 2019. “Affirmative Action and Pre-College Human Capital.” In NBER Working Paper.10.3386/w27779Search in Google Scholar

Alon, S., and O. Malamud. 2014. “The Impact of Israel’s Class-Based Affirmative Action Policy on Admission and Academic Outcomes.” Economics of Education Review 40: 123–39. https://doi.org/10.1016/j.econedurev.2014.02.004.Search in Google Scholar

Angelo, C., and A. B. Reis. 2021. “Gender Gaps in Different Grading Systems.” Education Economics 29 (1): 105–19. https://doi.org/10.1080/09645292.2020.1853681.Search in Google Scholar

Arcidiacono, P., and M. Lovenheim. 2016. “Affirmative Action and the Quality-Fit Trade-Off.” Journal of Economic Literature 54 (1): 3–51. https://doi.org/10.1257/jel.54.1.3.Search in Google Scholar

Arcidiacono, P., M. Lovenheim, and M. Zhu. 2015. “Affirmative Action in Undergraduate Education.” Annual Review of Economics 7 (1): 487–518. https://doi.org/10.1146/annurev-economics-080614-115445.Search in Google Scholar

Bertrand, M., R. Hanna, and S. Mullainathan. 2010. “Affirmative Action in Education: Evidence from Engineering College Admissions in India.” Journal of Public Economics 94 (1–2): 16–29. https://doi.org/10.1016/j.jpubeco.2009.11.003.Search in Google Scholar

Bhattacharya, D., K. Shin, and M. Stevens. 2017. “Are University Admissions Academically Fair?” The Review of Economics and Statistics 99 (3): 449–64. https://doi.org/10.1162/rest_a_00618.Search in Google Scholar

Bodoh-Creed, A., and B. R. Hickman. 2017. “Pre-College Human Capital Investments and Affirmative Action: A Structural Policy Analysis of US College Admissions.” In Becker Friedman Institute for Research in Economics Working Paper, no. 2017–10.10.2139/ssrn.3008669Search in Google Scholar

Borghans, L., B. H. H. Golsteyn, J. JamesHeckman, and J. E. Humphries. 2016. “What Grades and Achievement Tests Measure.” Proceedings of the National Academy of Sciences 113 (47): 13354–9. https://doi.org/10.1073/pnas.1601135113.Search in Google Scholar

Brenøe, A. A., and U. Zölitz. 2020. “Exposure to More Female Peers Widens the Gender Gap in Stem Participation.” Journal of Labor Economics 38 (4): 1009–54.10.1086/706646Search in Google Scholar

Cerdeira, J. M., L. C. Nunes, A. Balcão Reis, and C. Seabra. 2018. “Predictors of Student Success in Higher Education: Secondary School Internal Scores versus National Exams.” Higher Education Quarterly 72 (4): 304–13. https://doi.org/10.1111/hequ.12158.Search in Google Scholar

Clancy, P., and G. Goastellec. 2007. “Exploring Access and Equity in Higher Education: Policy and Performance in a Comparative Perspective.” Higher Education Quarterly 61 (2): 136–54. https://doi.org/10.1111/j.1468-2273.2007.00343.x.Search in Google Scholar

Cliffordson, C. 2008. “Differential Prediction of Study Success across Academic Programs in the Swedish Context: The Validity of Grades and Tests as Selection Instruments for Higher Education.” Educational Assessment 13 (1): 56–75. https://doi.org/10.1080/10627190801968240.Search in Google Scholar

Conger, D. 2015. “High School Grades, Admissions Policies, and the Gender Gap in College Enrollment.” Economics of Education Review 46: 144–7. https://doi.org/10.1016/j.econedurev.2015.03.003.Search in Google Scholar

Cotton, C., B. R. Hickman, and J. Price. 2014. “Affirmative Action and Human Capital Investment: Evidence from a Randomized Field Experiment.” In NBER Working Paper, no. w20397.10.3386/w20397Search in Google Scholar

Duckworth, A. L., and M. E. P. Seligman. 2006. “Self-Discipline Gives Girls the Edge: Gender in Self-Discipline, Grades, and Achievement Test Scores.” Journal of Educational Psychology 98 (1): 198. https://doi.org/10.1037/0022-0663.98.1.198.Search in Google Scholar

Fischer, F. T., J. Schult, and B. Hell. 2013. “Sex-Specific Differential Prediction of College Admission Tests: A Meta-Analysis.” Journal of Educational Psychology 105 (2): 478. https://doi.org/10.1037/a0031956.Search in Google Scholar

Fischer, S. 2017. “The Downside of Good Peers: How Classroom Composition Differentially Affects Men’s and Women’s STEM Persistence.” Labour Economics 46: 211–26. https://doi.org/10.1016/j.labeco.2017.02.003.Search in Google Scholar

FryerJr, G. Roland, G. C. Loury, and T. Yuret. 2008. “An Economic Analysis of Color-Blind Affirmative Action.” Journal of Law, Economics, and Organization 24 (2): 319–55.10.1093/jleo/ewm053Search in Google Scholar

Gipps, C. V., and P. Murphy. 1994. A Fair Test? Assessment, Achievement and Equity. Philadelphia: Open University Press.Search in Google Scholar

Grau, N. 2018. “The Impact of College Admissions Policies on the Academic Effort of High School Students.” Economics of Education Review 65: 58–92. https://doi.org/10.1016/j.econedurev.2018.03.002.Search in Google Scholar

Helms, R. M. 2008. “University Admissions Worldwide.” In Education Working Paper Series 15.Search in Google Scholar

Helms, R. M. 2009. “University Admissions: Practices and Procedures Worldwide.” International Higher Education 54: 5–7.10.6017/ihe.2009.54.8410Search in Google Scholar

Karlsson, L., and M. Wikström. 2021. Gender Differences in Admission Scores and First-Year University Achievement. Umeå: Umeå University, Department of Economics.Search in Google Scholar

Keiser, H. N., P. R. Sackett, N. R. Kuncel, and T. Brothen. 2016. “Why Women Perform Better in College than Admission Scores Would Predict: Exploring the Roles of Conscientiousness and Course-Taking Patterns.” Journal of Applied Psychology 101 (4): 569. https://doi.org/10.1037/apl0000069.Search in Google Scholar

Lyrén, P.‐E. 2008. “Prediction of Academic Performance by Means of the Swedish Scholastic Assessment Test.” Scandinavian Journal of Educational Research 52 (6): 565–81. https://doi.org/10.1080/00313830802497158.Search in Google Scholar

Mouganie, P., and Y. Wang. 2020. “High-Performing Peers and Female Stem Choices in School.” Journal of Labor Economics 38 (3): 805–41. https://doi.org/10.1086/706052.Search in Google Scholar

Murphy, R., and W. Gill. 2020. “Minority Report: The Impact of Predicted Grades on University Admissions of Disadvantaged Groups.” Education Economics 28 (4): 333–50. https://doi.org/10.1080/09645292.2020.1761945.Search in Google Scholar

Sawyer, R. 2007. “Indicators of Usefulness of Test Scores.” Applied Measurement in Education 20 (3): 255–71. https://doi.org/10.1080/08957340701431245.Search in Google Scholar

Sawyer, R. 2013. “Beyond Correlations: Usefulness of High School GPA and Test Scores in Making College Admissions Decisions.” Applied Measurement in Education 26 (2): 89–112. https://doi.org/10.1080/08957347.2013.765433.Search in Google Scholar

Silva, P. L., L. C. Nunes, C. Seabra, A. Balcao Reis, and M. Alves. 2020. “Student Selection and Performance in Higher Education: Admission Exams vs. High School Scores.” Education Economics 28 (5): 1–18. https://doi.org/10.1080/09645292.2020.1782846.Search in Google Scholar

Thiele, T., A. Singleton, D. Pope, and D. Stanistreet. 2016. “Predicting Students’ Academic Performance Based on School and Socio-Demographic Characteristics.” Studies in Higher Education 41 (8): 1424–46. https://doi.org/10.1080/03075079.2014.974528.Search in Google Scholar

Wedman, J. 2017. Theory and Validity Evidence for a Large-Scale Test for Selection to Higher Education. Umeå: Umeå University.Search in Google Scholar

Wikström, C., and M. Wikström. 2017. “Group Differences in Student Performance in the Selection to Higher Education: Tests vs Grades.” Frontiers in Education 2: 45. https://doi.org/10.3389/feduc.2017.00045.Search in Google Scholar

Wikström, C., and M. Wikström. 2020. “Merit-based Admissions in Higher Education.” In Higher Education Admissions Practices: An International PerspectiveCathy Wendler and María Elena Oliveri, 34–50. Educational and Psychological Testing in a Global Context. Cambridge: Cambridge University Press.10.1017/9781108559607.003Search in Google Scholar

Zwick, R. 2002. “Fair Game?” In The Use of Standardized Admissions Tests in Higher Education. New York: RoutledgeFalmer.10.1037/e672192007-009Search in Google Scholar

Zwick, R. 2013. “Disentangling the Role of High School Grades, SAT® Scores, and SES in Predicting College Achievement.” ETS Research Report Series 2013 (1): 1–20. https://doi.org/10.1002/j.2333-8504.2013.tb02316.x.Search in Google Scholar

Zwick, R. 2017. “Who Gets In?” In Strategies for Fair and Effective College Admissions. Cambridge, MA: Harvard University Press.10.4159/9780674977648Search in Google Scholar

Received: 2020-12-17

Accepted: 2021-12-20

Published Online: 2022-01-04

Admission Groups and Academic Performance: A Study of Marginal Entrants in the Selection to Higher Education

Abstract

Acknowledgements

References

Journal and Issue

Articles in the same Issue