Person Fit

Person fit was used to identify response patterns that are deemed unlikely to be observed based on the Rasch model. People with excessive fit residuals, greater than 2.0 logits, were examined to determine the reason for poor fit.⁵³ They were compared with those who fit the model using Fisher’s exact test of significance (for gender) or the Mann–Whitney U-test (for age, pain intensity, pain duration, disability and FreNAQ-J score). Response strings of misfitting participants were analyzed to identify patterns in their responses. The presence of seriously misfitting cases has been shown to influence item parameter estimates.⁵⁴ If there was no clear pattern in the response data, re-analysis was performed after excluding misfit participants to examine which FreNAQ-J responses were affected by misfitting cases.