Information Systems researchers are often interested in comparing outcomes across different groups of interest, as in the case of experimental and quasi-experimental studies. These designs have traditionally been modeled, as we show through a review of our literature, by using analysis of variance techniques on observed scores, typically the sum or average of all items measuring the dependent variable of interest. These designs, however, can be analyzed with structural equation modeling and latent variables (SEM-LV) techniques, which can better accommodate measurement error and more complex models than would otherwise be possible using the traditional techniques. This research introduces the foundations of the SEM-LV approach for these research designs and highlights these advantages, and provides several examples that underscore the flexibility of the latent variable techniques discussed here. It also compares the two main alternatives for implementing this approach and discusses the advantages and disadvantages of each.
"Improving Between-group Comparisons in IS Research Through the Use of SEM and Latent Variables: An Introduction and Some Examples,"
Communications of the Association for Information Systems: Vol. 34
, Article 30.
Available at: https://aisel.aisnet.org/cais/vol34/iss1/30