Abstract

This study identified and examined the characteristics of some selected empirical studies on validity of UME in Nigeria. It investigated the effect of sample size on the selected studies with a view to establishing the validity of the reported coefficients of the studies. The study made use of an ex-post facto design. Statistical data from 30 empirical studies were purposely selected from internet and manual searching of journal. Psychometric meta-analytic method was used to code the different study characteristics while chi square for focused and diffused tests were used to compare and combine the study outcomes. The results showed that the Weighted Mean Fisher (WZr) and Mean Fisher (Zr) were the same (WZr)=(Zr) = 0.39 indicating that study characteristics of the selected empirical studies did not affect the validity of UME in terms of type of publications outlets, statistical measure, study coverage and year of publication. The effect size of published articles was significantly different from unpublished theses. (X2=5.421 p<0.05). However Weighted mean by sample size was greater than Mean Fisher WZr = 0.497> Zr=0.39. The implication of this is that the heterogeneity of the effect sizes of the empirical studies was influenced by the sample sizes. The study concluded that the validity of UME was relatively low (r = 0.375) and the variation in the results of the empirical studies on UME was as a result of differences in their sample sizes and not because of study methods or time of publication.