(PS1-D88) Measure Development of the Sexual and Negative Dating Inventory (SANDI): Psychometric Evaluation Using a College Student Sample

Background: Alcohol use and adverse sexual outcomes (e.g., sexual assault, risky sex, and regretted sexual experiences) are strongly associated on college campuses. Understanding mechanisms and behaviors that can reduce adverse sexual outcomes is an important area of public health research. Harm reduction is a framework aimed at developing a protective behavioral repertoire. Protective Behavioral Strategies (PBS) are a set of individually implemented harm reduction strategies that effectively reduce both the rate of harm and severity of consequences from external behaviors. The aim for the proposed study is to develop a measure that assesses dating and sexual protective behavioral strategies; the Sexual and Negative Dating Inventory (SANDI).

Method: At baseline, n = 1722 individuals from a large Southeastern University filled out demographic information, the SANDI, the DSPARS, DBS, RSS, and the safer-sex PBS scale to assess for convergent validity. Test-retest reliability was assessed at one month follow-up with n = 448 participants. Alcohol, negative consequences, and adverse sexual outcomes were assessed using the DDQ, BYAACQ, and SES-SFV, respectively. Individuals were on average 19.67 (2.85 SD) years old and primarily female (69.4%). Participants identified as White (69.8%) African American (9.9%), Asian (9.1%), Multiple Races (5.7%), an unlisted race (4.5%), Native Hawaiian/Pacific Islander (0.5%), or American Indian/Alaska Native (0.5%). A total of 60 items were proposed for the SANDI.

Results: A four-factor structure was initially hypothesized, however, EFA identified a five-factor structure. A CFA was conducted with the 5 factors. Initially, 26 items were identified with high cross-loadings across factors. Eleven items did not load well on their factors (< .05), and 3 items had high modification indices within factors that suggesting high correlated error, these were iteratively removed. This resulted with 20 items which showed good fit to the data: χ²(160) = 1039.79, p < .001, CFI = .964, RMSEA = .060 (90% CI = .056, .063), SRMR = .035. Correlation revealed test-retest reliability across factors which was adequate (F1: 0.49, F2: 0.80, F3: 0.75, F4: 0.62, and F5: 0.62). Negative binomial regression showed an inverse association between the SANDI and alcohol consumption (B = -0.51, p < .001), negative consequences (B = -0.59, p < .001), AUDIT score (B = -0.08, p = .015), and sexual risk (B = -0.07, p = .037). A logistic regression showed a positive association with sexual victimization history (OR = 1.62, p < .001).

Conclusions: The SANDI is a new measure of sexual and dating protective behaviors. This measure provides a factor structure of protective behaviors that can serve as targets for future intervention research.