Analysis of partial least squares on split path models: underpinning delone & maclean theory

Abstract

The analysis of partial least squares utilizing Structural Equation Modeling is well known as a second generation technique .Its true that , this technique has not been fully utilized in the field of research to handle complex data analysis in simulation and modeling by a number of researches globally. Such has been experienced in universities within African continent and specifically Postgraduate students particular in Kenya . PLS handled by SEM is considered not friendly by a number of scholars hence it has taken a low profile trend. This research aimed at analyzing partial least squares on Multiple group model comparison. Research underpinned a Delone & Mclean theory, attention is directed towards multiple group comparison of Six constructs with 36 split path diagrams. Data was adopted from the public sector from a PhD Thesis. Findings indicated that: Multiple group modeling comparison indicated that: Technical operation skills had positive significant difference on IFMIS applicability; Group Comparison yielded a lower ratio index meaning it fitted well. Information Quality model performed equally well in terms of Model Ratio Index. In terms of Goodness of fit comparison it indicated good results above the threshold. Level of IT Infrastructure delivered results of poor goodness of fit,which was not significant and had a higher ratio Index above 5, indicating a poor model fitting. All in all Partial least squares on Multiple Group Modeling Comparison was successfully utilized in measuring and analyzing the constructs of Delone & Mclean Theory. In conclusion all the five models were optimum and indicated an effective measure on IFMIS applicability, apart from Level of IT Infrastructure. Study contribution include:Techniques on Partial least squares by Structural Equation Modeling using by Splitting models Comparison, analysis by AMOS, Ratio indices. Study recommends further investigation of New Techniques for Analysis: Bootstrapping and Nested comparison.