Question order effects in partial least squares path modelling: an empirical investigation

Effects of question order in surveys have been extensively researched. In the light of partial least squares path modeling, only few studies exist. In this study we analyze effects of question order with data from two waves of an annual satisfaction study. The positioning of area satisfaction questions and randomized attribute satisfaction questions was changed between the two waves. Results confirm previously described positive effects on explained variance if area satisfaction is asked after attribute satisfaction. This increase in explained variance in area performance does not lead to an improved model performance in explaining overall satisfaction.