(PS8-91) The Tonic Immobility Scale and Consideration of Peritraumatic Use of Alcohol or Other Substances: An Analysis of Item Differences and Measurement Invariance of a Two-factor Model

Tonic immobility (TI) has been defined as “an involuntary component of the fear response that is characterized by freezing or immobility in situations involving extreme fear coupled with physical restraint” (Fuse et al., 2007) and is frequently endorsed as a peritraumatic response during rape. A meta-analysis concludes TI is closely associated with the severity of Posttraumatic Stress Disorder symptoms and the relationship is particularly notable when experienced by women and in the context of interpersonal trauma (Coimbra et al., 2023). Individuals endorsing histories of rape often report consensual or non-consensual consumption of alcohol or other substances. Yet, it is less understood how substances may play a factor in peritraumatic responses either by an amplification or dampening of fear responses or incorporation of alternate physiological processes. The Tonic Immobility Scale (TIS; Forsyth et al., 2000) was developed to evaluate fear-based freeze responses. Fuse and colleagues posited a two-factor solution to this scale, though continued replication work is needed as well as consideration of issues of potential measurement invariance between those who do and do not believe that substance use influenced their item responses. To further evaluate, 177 women endorsing sexual assault histories as part of a larger longitudinal investigation completed the TIS and answered a dichotomous question querying their belief that alcohol or substances were associated with their TIS responses. Between-group t-tests found those who believed substances were associated with their TIS responses (36%) reported significantly more paralysis (1.98, p = .049), inability to move (2.08, p = .04), detachment from self (2.2, p=.029), and detachment from environment (2.79, p = .006) than their peers who did not endorse this belief. We further sought to replicate the two-factor model identified by Fuse and evaluate potential measurement invariance based on peritraumatic substance use perceptions. Our configural CFA model demonstrated poor fit (CMin=322, p < .001; RMSEA=.149; CFI=.711; TLI=.606). Similarly, metric invariance with constrained parameters for factor loading by group membership was indicative of unacceptable model fit (CMin=331, p < .001; RMSEA=.141; CFI=.711; TLI=.658). Item loadings were particularly poor for four of the ten items for those believing responses were impacted by substances and for one item for those who did not endorse this belief. Due to poor model performance on configural and metric invariance testing, scalar and residual invariance were not investigated. Results are limited by cross-sectional survivor self-perceptions of the potential influence of substances, which may vary from substance-influenced physiological responding or require measurement closer to event occurrence. However, our research identifies variability in item response severity on TIS items between the two groups. We further suggest the two-factor model may not be psychometrically robust for measuring TI factor constructs generally, and particularly for those perceiving relationships between TI and contemporaneous substance use. Additional research is needed to identify the most beneficial facture structure of the TIS.