Factors Affecting Research Productivity of Faculty Members in Government Universities: Lisrel and Neural Network Analyses

Abstract

     The purposes of this research were 1) to study researcher's characteristics, researchership, research competence and institutional support for research work as factors affecting research productivity, 2) to test for invariance of research productivity models across groups with size difference in Pedagogy Department, and 3) to compare the results of factors affecting research productivity using LISREL and Neural Network analyses. The sample consisted of 300 faculty members from 16 government universities. The research instruments were rating scales measuring research productivity, researchership, research competence and institutional supports for research work. The reliabilities of the instrument ranged from .76-.96. Data were analyzed through descriptive statistics, LISREL, and Neural Network Analyses. The major findings were: 1) The average of each faculty member's research productivity was 0.40 research pieces per year; 2) Researchership and research competence were high in average, and institutional support for research work was moderate; 3) Research productivity model fitted well to empirical data (Chisquare= 80.007, p=0.132 df=67, GFI=0.963, AGFI=0.942, RMR=0.161). The test of model invariance across 2 groups of departments with different size indicated that the two models were invariant in form, but varied in loading and other parameters. The causal relationship using LISREL and Neural Network analyses suggested consistently that researcher characteristic, research competence, institutional support for research work and researchership had direct effects on research productivity; 4) The comparison of analyses with LISREL and Neural Network indicated similar results.