Statistical Audit of Data Analysis of Educational Researches

Abstract

The present study is a part of the research project approved and supported by ICSSR, New Delhi, under its Senior Fellowship Programmme. 126 Ph.D. and 11 M.Phil. theses submitted to five universities of the five North-West Indian States were examined with a view to identify dominant statistical practices adopted by the researchers in education. The study revealed that 56.9 per cent of the research studies used t-test, while 33.6 per cent employed Analysis of Variance technique to test the hypotheses. Thus, over 90 per cent of the researches used either t-test or Analysis of Variance (ANOVA) in particular. However, it was found that in 51.3 per cent applications of the total t-tests and 21.7 per cent of the total ANOVA uses were inappropriate. Next most preferred (29.9 per cent) practice comprised of testing relationships using coefficient of correlations technique. Furthermore, the assumptions underlying these techniques were seldom tested. Consequently, only one researcher was found to have used non-parametric test of hypothesis. There was just one study wherein the researcher has reported effect size for each of the factors without interpreting it. The findings of the study could serve as a useful guide for chalking out need-based intervention programmes for researchers in education.