Which approach should be preferred to create the interaction term?

To answer this question, Henseler and Chin (2010) ran an extensive simulation study, comparing the approaches in terms of their statistical power, point estimation accuracy, and prediction accuracy.

When formative measures are involved, the two-stage approach must be used.

If the exogenous construct and the moderator are measured reflectively, the further choice of method depends on the aim of the analysis.

1. When the objective is to determine whether or not the moderator exerts a significant effect on the relationship, the two-stage approach is preferred. This approach yields high levels of statistical power compared with the orthogonalization approach and especially to the product indicator approach.

2. When the primary concern is minimizing estimation bias, the orthogonalizing approach should be preferred as it performs best in terms of point accuracy.

3. When the aim is to maximize prediction of the endogenous construct, researchers should apply the orthogonalizing approach as this approach yields high prediction accuracy.