Detecting Response Styles in Visual Analogue Scales
Year: 2026
Type: preprint
Abstract: Response styles are prevalent in self-report data and may bias parameter estimates. While there are various approaches for detecting response styles in Likert-type data, models for visual analogue scales remain scarce. We propose a novel modeling framework that extends the beta item response model by Noel and Dauvier (2007) to jointly detect extreme response style, midpoint response style, and acquiescent response style. Estimated in a Bayesian framework using Stan, the model performed well in a comprehensive simulation study. Parameter recovery was primarily influenced by the number of (reverse-scored) items and the variance of response styles, whereas sample size had a negligible effect. The model’s practical utility is illustrated using empirical data (Kuhlmann et al., 2017), for which it demonstrated superior model fit compared to the original beta item response model. Overall, the proposed framework provides researchers with a principled approach to account for response styles in visual analogue scale data.
Source: PsyArXiv (OSF Preprints)
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7214572505