Predictive Validity of the Unified Tertiary Matriculation Examination on University Grades

A Validity Generalization Meta-Analysis

The Unified Tertiary Matriculation Examination (UTME) is a standardized test used for university admissions decisions in Nigeria. The examination is administered by the Joint Admissions and Matriculation Board (JAMB). The validity of the UTME in predicting academic performance has been investigated extensively in the literature, therefore this meta-analysis aims to synthesize the current validity evidence. This open-source web-page consists of the results, methodology, and data for a real-time updating meta-analytic database.
Authors
Affiliations

Matthew B. Jané

University of Connecticut

James Uanhoro

University of North Texas

James Steele

Solent University

Nduka Boika

University of North Texas

Lucija Batinovic

Linköping University

Blair T. Johnson

University of Connecticut

Current Results

Study Selection

Imputing Missing Reliability

The average inter-correlation between each year-end GPA (\(\bar{r}_{y_iy_j}\); see Methods) and the number of years that the students attend college, \(\kappa\), can be used to estimate the reliability of final cumulative GPA (CGPA) during their time in college. For example, a CGPA calculated from all four years of college is more reliable than a CGPA computed from just their first year in college. Since inter-grade correlations are rarely reported in primary studies, we can take the \(N\)-weighted average of available inter-grade correlations to fill in missing values. We then use the number of years of university the study computes the CGPA from and use those values to estimate the reliability of CGPA in the given study.

Estimated Reliability

# of Academic Years \((\kappa)\) Estimated Reliability \((r_{YY'})\)
1 0.738
2 0.849
3 0.894
4 0.918
5 0.934

Full Meta-Analysis

The full meta-analysis of all studies is reported below with an uncorrected (i.e., bare bones) and corrected version.

Bare Bones Model (No Correction)

\(k\) \(N\) \(\bar{r}\) \(CI_{L95}\) \(CI_{U95}\) \(SD_r\) \(CR_{L90}\) \(CR_{U90}\)
31 10675 0.227 0.169 0.285 0.157 -0.008 0.461

Artifact Correction Model

\(k\) \(N\) \(\hat{\bar{\rho}}\) \(CI_{L95}\) \(CI_{U95}\) \(SD_\rho\) \(CR_{L90}\) \(CR_{U90}\)
31 10675 0.268 0.203 0.334 0.156 0.003 0.533

Moderators

Disciplines

College disciplines/majors are broken down into broad categories.

Computer Based \(k\) \(N\) \(\hat{\bar{\rho}}\) \(CI_{L95}\) \(CI_{U95}\) \(SD_\rho\) \(CR_{L90}\) \(CR_{U90}\)
All Levels 31 10675 0.268 0.203 0.334 0.157 0.002 0.535
Business and Economics 5 2997 0.224 0.059 0.389 0.120 -0.033 0.481
Education 3 1314 0.441 -0.158 1.000 0.237 -0.251 1.000
Language, Humanities, and Art 3 604 0.339 -0.273 0.951 0.223 -0.313 0.991
Math, Science, and Engineering 15 4319 0.226 0.145 0.307 0.111 0.031 0.421
Mixed 1 438 0.035 -0.077 0.146 NA NA NA
Social Science 4 1003 0.353 0.115 0.591 0.125 0.059 0.647

Type of Institution

Whether the institution is a university, polytechnic, or college of education.

Institution Type \(k\) \(N\) \(\hat{\bar{\rho}}\) \(CI_{L95}\) \(CI_{U95}\) \(SD_\rho\) \(CR_{L90}\) \(CR_{U90}\)
All Levels 31 10675 0.268 0.203 0.334 0.156 0.004 0.533
college of education 3 1314 0.436 -0.152 1.000 0.233 -0.243 1.000
university 28 9361 0.248 0.186 0.310 0.134 0.020 0.475