A comparison of predictive validity of undergraduate achievement, controlling intelligence-free factor using an equate and non equate secondary school grade point average criterion
Abstract:
The main purposes of this study were to analyze the correlation between equate and non equate secondary school grade point average by using 3 methods namely; equipercentile method, Angoffs linear method design IV-C2, IRT grade response model and the undergraduate achievement, controlled non related intelligence factors. After that compare the predictive validity of the undergraduate achievement, controlled non related intelligence factors by using equate and non equate secondary school grade point average criterion. The sample of this study was 1,369 students in the first year, 2005 academic year of the public universities, i.e. Chulalongkorn, Thamasat, Kasetsart, Mahidol, Srinakarinwirot. The data were secondary school grade point average, entrance examination score, the information from the non related intelligence factors and the first year of the university GPA. The results were as follow : 1. The secondary school grade point average which non equate and equate by using equipercentile method, Angoff linear method design IV-C2, IRT grade response model correlated with the undergraduate achievement, controlled non related intelligence factors significantly at .05 statistical level. 2. The comparison of the prediction showed that the equate secondary school grade point average had higher predictive validity than non equate secondary school grade point average significantly at .05 statistical level. The comparison of the predictive validity among secondary school grade point average by using 3 methods had been found that the equipercentile method and Angoff linear method design IV-C2 yielded higher predictive validity than IRT grade response model, and which were equal.