Modeling the response style in continuous bounded responses: Model development and validation.

Area of Science:

• Psychometrics
• Statistical Modeling
• Survey Methodology

Background:

• Item response theory models are established for Likert-scale data but less developed for continuous formats like visual analogue scales (VAS) and slider bars.
• Response styles, such as extreme response style (ERS) and midpoint response style (MRS), can bias results in continuous measurement formats.
• Existing models often fail to adequately address or isolate these response styles in continuous data.

Purpose of the Study:

• To propose a novel item response model framework for analyzing continuous bounded response formats.
• To flexibly incorporate content traits, extreme response style (ERS), and midpoint response style (MRS) into a unified model.
• To validate the model's ability to accurately estimate ERS and MRS and mitigate their impact on observed responses.

Main Methods:

• Development of a hierarchical item response model framework utilizing pseudo-responses.
• Empirical validation using continuous bounded response data to assess ERS and MRS estimation.
• Simulation studies employing Markov chain Monte Carlo (MCMC) methods to evaluate parameter recovery.

Main Results:

• The proposed model demonstrated a superior fit to continuous bounded response data compared to existing approaches.
• The model effectively estimated extreme response style (ERS) and midpoint response style (MRS).
• Markov chain Monte Carlo (MCMC) methods accurately recovered model parameters across various simulation conditions.

Conclusions:

• The novel item response model framework provides a robust and valid approach for analyzing continuous bounded response data.
• The model successfully isolates and quantifies response styles (ERS, MRS), mitigating their adverse effects on observed responses.
• This framework advances psychometric modeling for continuous measurement instruments, enhancing data accuracy and interpretability.