(PS2-B36) Suicide Cognitions Scale-revised: Psychometric Support in a Community Sample Using Bifactor Modeling

The Suicide Cognitions Scale -Revised (SCS-R; Bryan et al., 2021) assesses suicide specific cognitions (e.g. hopelessness, feelings of unlovability). Bryan and colleagues’ (2021) pilot study of the SCS-R demonstrated a superior bifactor solution for the instrument. Although promising, this investigation was conducted in a sample of adults (N=2,000) with limited recent history of suicidal thoughts and behaviors (STBs). The current study investigated the psychometric properties of the SCS-R in a larger sample to confirm/replicate the measure’s factor structure as well further test the criterion validity of the SCS-R. Adults (N= 10,265) were recruited through Qualtrics Panels and completed a self-report measure of past-month STBs as well as the SCS in an online study. Bifactor models in addition to traditional confirmatory models were conducted to determine the extent to which the scale can be interpreted as unitary as seen in Bryan and colleagues (2021). Results indicated that the bifactor model with 5 factors demonstrated the best overall fit but had overall low factor loadings (< .20). However, the Exploratory Structural Equation Modeling (ESEM) bifactor model with 3 specific factors demonstrated good fit and had higher item factor loadings and was thus chosen (c²(62) = 587.59, p < .05, TLI = .998, CFI = .999, RMSEA = .028, SRMR = .005). All 16 items loaded onto the general factor ( >.81), while only 7 items loaded onto specific factors ( >.20), and only one had a loading >.30. All items appeared to be predominantly influenced by the general factor (IECV >.85, PUC = .94, wH=.98). Multinomial logistic regressions indicated that the total score of the SCS-R is useful in distinguishing varying levels of STBs such as past month planning for suicide without attempt versus past month suicide attempt (c²= 1,270.62, p < .001) as well as past month SI but no planning and those with both past month planning and SI (χ²=2,516.22, p< .001). Replicating the work of Bryan and colleagues (2021), the SCS-R demonstrated a strong unidimensional latent factor and can be used to discriminate amongst individuals with varying severity of STBs. These results support the clinical utility of the SCS-R as a tool for determining risk for STBs in line with the Fluid Vulnerability Theory (FVT; Rudd, 2006) that can be used for assessment and treatment monitoring purposes. Future research with the SCS-R should prioritize longitudinal designs to determine predictive utility of the measure.