A modified hidden Markov model for detecting insufficient effort responses in questionnaires

Insufficient effort response (IER) significantly compromises the quality of questionnaire data, affecting the validity of resulting inferences. Traditional methods for detecting IER often fail to adequately capture various types of IER or consider participants' internal state transitions. This study expanded the hidden Markov model for analyzing participants' response strategies by reconstructing response and response time (RT) models that target the identification of IER in the context of questionnaires. The method takes into account the characteristics of IER in terms of response and RT, with the aim of dynamically detecting various types of IER. The simulation study demonstrated that a modified hidden Markov model (M-HMM) effectively recovers parameters, with its detection sensitivity primarily influenced by the prevalence of IER, differences in RT distributions between insufficient and effortful responses, and variations in IER severity and type among participants. Utilizing the M-HMM to analyze empirical data allowed for a deeper understanding of IER occurrences and improved item quality assessment, offering valuable insights for practitioners.