The Advantages of LISREL for Accounting Researchers

LISREL (linear structural relations) is used to explore causal relationships among variables, and it is compared to path analysis as a method of exploring such relationships. The data from Rebele and Michaels (1990) and Lachman and Aranya (1986), which used ordinary path analysis, were re-examined using LISREL. There are differences between the final models obtained through LISREL analysis that included measurement error and the models obtained by the prior studies with ordinary path analysis and no measurement error. If a model that fits the data is identified using regression-based path analysis, reanalyzing the data using LISREL will result in less biased parameter estimates due to the explicit incorporation of measurement error in the model. The LISREL analysis has the advantages of including measurement error and assessing via the chi-square statistic the comparable fit of competing models. Future accounting researchers should be aware of the potential benefits of LISREL.