Response style bias in tourism research: using item response tree models to reduce measurement error

Despite widespread use of rating scales in tourism research, participant response styles are a rarely considered form of measurement error. This research provides an account of response style bias in tourism research and introduces item response tree (IRTree) models as a promising method to identify and address these biases. Specifically, this research demonstrates how IRTrees can separate two prominent types of response style biases (extreme response style and midpoint response style) from the traits being measured. An R code is provided for use in future research, to assist tourism researchers interested in collecting high quality survey data.