Open to interpretation? Inconsistent reporting of lifetime non-suicidal self-injury across two common assessments
Non-suicidal self-injury (NSSI) is typically assessed using either single-item questionnaires or checklists of common behaviours, but preliminary research suggests that checklists produce higher lifetime prevalence rates. In two pre-registered studies (combined n = 1364), we tested whether memory cueing afforded by behavioural checklists accounts for this discrepancy. Participants reported their lifetime NSSI history using both a single-item and a checklist, with presentation order randomised across participants. Nearly a third of participants reported inconsistent NSSI histories on the two assessments, with participants 1.57 times more likely to report an NSSI history on a checklist than on a single-item. Counter to the memory account, this discrepancy was evident even when participants completed the checklist first, suggesting that the increased prevalence estimates captured by checklists are unlikely to simply reflect memory facilitation. Across the two samples, 12.5% of participants would have been incorrectly screened out in two-step assessments; these participants were more likely to have engaged in NSSI historically, less likely to self-injure by cutting, and (in Study 2 only) were more likely to be men. These studies suggest that the inconsistencies across two of the most common NSSI assessments arise because people dissimilar to the lay conceptualisation of self-injury are less likely to endorse a single-item, even when they have affirmed engaging in self-injury behaviours on a checklist. We argue that single-item and checklist assessments capture different aspects of NSSI, such that future research should distinguish between behaviourally-identified NSSI assessed with behavioural checklists and self-identified NSSI assessed with single-item assessments.