A Novel Partial Credit Extension Using Varying Thresholds to Account for Response Tendencies

Item Response Theory models with varying thresholds are essential tools to account for unknown types of response tendencies in rating data. However, in order to separate constructs to be measured and response tendencies, specific constraints have to be imposed on varying thresholds and their interrelations. In this article, a multidimensional extension of a Partial Credit Model using a sum‐to‐zero constraint for varying thresholds is proposed. The new model allows us to flexibly account for response tendencies and to model covariations between varying thresholds that are commonly found in empirical data. The model's ability to estimate different types of response tendencies under various data structures is shown in a simulation study. An illustrative multicountry analysis demonstrates that differences between respondents in terms of their response tendencies exist and can be captured by the new model. Furthermore, it is well suited to account for extreme and mid response styles, but also to accommodate unknown, previously unmodeled, response tendencies. Therewith, the sum‐to‐zero model can be considered a suitable candidate to examine the types response tendencies in rating data and to account for biases in construct measures due to response tendencies.