Person misfit on the Inventory of Depressive Symptomatology: Low quality self‐report or true atypical symptom profile?

Abstract

Person misfit on a self-report measure refers to a response pattern that is unlikely given a theoretical measurement model. Person misfit may reflect low quality self-report data, for example due to random responding or misunderstanding of items. However, recent research in the context of psychopathology suggests that person misfit may reflect atypical symptom profiles that have implications for diagnosis or treatment. We followed-up on Wanders et al. (Journal of Affective Disorders, 180, 36-43, 2015) who investigated person misfit on the Inventory of Depressive Symptomatology (IDS) in the Netherlands Study of Depression and Anxiety (n = 2,981). Our goal was to investigate the extent to which misfit on the IDS reflects low-quality self-report patterns and the extent to which it reflects true atypical symptom profiles. Regression analysis showed that person misfit related more strongly to self-report quality indicators than to variables quantifying theoretically-derived atypical symptom profiles. A data-driven atypical symptom profile explained most variance in person misfit, suggesting that person misfit on the IDS mainly reflects a sample- and questionnaire-specific atypical symptom profile. We concluded that person-fit statistics are useful for detecting IDS scores that may not be valid. Further research is necessary to support the interpretation of person misfit as reflecting a meaningful atypical symptom combination.

Satisficing in Mental Health Care Patients: The Effect of Cognitive Symptoms on Self-Report Data Quality

“…Respondents may answer extremely not only because they simplify the response scale (i.e., use a satisficing strategy) but also because they are truly extreme in their behavior (e.g., He, Bartram, Inceoglu, & van de Vijver, 2014). Combining information from different validity indices may thus decrease the possibility that an unexpected response pattern is actually valid and meaningful (e.g., Conijn, Spinhoven, Meijer, & Lamers, 2017; Wanders, Wardenaar, Penninx, Meijer, & de Jonge, 2015).…”

Person-fit feedback on inconsistent symptom reports in clinical depression care

“…One way to better identify different causes of inconsistent symptom reports would be to utilize and develop more objective measures of the factors found to be associated with the quality of the responses. For example, studies associated self-report quality indicators like an extreme or agreement response style (van Herk et al 2004; Conijn et al 2016), interviewer evaluation of language problems (Conijn et al 2016), and external indicators of cognitive difficulties (Conijn et al 2017) with poorer person fit. Another promising approach would be to analyze the patterns of time spent by a patient on answering each item in computer-based assessments (Marianti et al 2014).…”

“…to prevent unwanted hospitalization). In addition, within psychopathology measurement clinical causes like comorbidity (Wanders et al 2015 a ; Wardenaar et al 2015) or concentration problems (Conijn et al 2016, 2017) may lead to inconsistent symptom reports where patients report severe symptoms without milder symptoms (Woods et al 2008; Conrad et al 2010; Wardenaar et al 2015; Conijn et al 2016). Detecting such inconsistent response patterns may have diagnostic value in clinical practice (Pallant & Tennant, 2007; Reise & Waller, 2009; Thomas, 2011).…”

“…Although many statistics are available to assess person fit, there are only few studies that examine the reason of this unexpected test behavior and how this behavior can invalidate test interpretation by a clinician. Recent person-fit studies have suggested that inconsistent symptom reports could reflect clinical factors, like an atypical suicide risk (Conrad et al 2010), or atypical presentation of symptoms (Wanders et al 2015 a ; Wardenaar et al 2015), but could also reflect causes associated with the quality of self-reported data (Conijn et al 2016), caused by, for example, cognitive problems (Conijn et al 2016, 2017). These studies analyzed large-scale archival data, which makes it difficult to retrieve and study the factors that underlie the responses for specific patients and from a more qualitative point of view (see Meijer et al 2008, for an exception in the education field).…”