Direct and Indirect Indicators of Street Victimization Among Homeless Youth

Background: While living on the streets, homeless youth experience high rates of victimization. Such encounters are associated with adverse mental health outcomes, such as posttraumatic stress and depressive symptoms. However, few instruments exist to assess victimization experiences on the streets.  The current study adds to the existing knowledge base by delineating direct (e.g., physical assault) and indirect (e.g., witnessing assault) indicators of street victimization, and assessing the effectiveness of these indicators in predicting mental health outcomes. The purpose of this study was to inform development of an index for assessing experiences of street victimization with utility for researchers and practitioners concerned with safety and mental health outcomes among homeless youth.

Methods: Quantitative interview data were collected via a 3-city, cross-sectional study of homeless youth (N=601) ages 18-24 seeking homeless services in Denver (n=201), Austin (n=200), and Los Angeles (n=200). A shortened version of the Traumatic Life Events Questionnaire (TLEQ) was used and tested with homeless youth by retaining 10 items asking youth whether, since they left home for the streets, they experienced specific direct and indirect victimization experiences (e.g., robbery, physical assault, sexual assault, witnessing assault). Whether youth met DSM-IV criteria for posttraumatic stress disorder (PTSD) and major depressive episode was also assessed, using the Mini International Neuropsychiatry Interview. Using Mplus Version 6, formative indicators analyses were employed, in which multiple causes models (MIMIC) determined if all 10 items of the revised TLEQ were an adequate index of street victimization. A two-indicator model, consisting of indirect (4 items) and direct (6 items) indices, was then tested to investigate improvement in model fit. Common in formative indicators analyses, the inclusion of reflective measures is necessary for model identification; thus, PTSD and depression were included as reflective outcomes. Final analyses were conducted to examine an interaction effect of exposure to both direct and indirect street victimization on outcomes of PTSD and depression.

Results: The initial estimation of the MIMIC model with all 10 items yielded a good fit (X²=8.11, df=9, p=.52; RMSEA=0.00; CFI=1.00). The model significantly predicted PTSD (OR=1.30, p<.05), but not depression. The two-indicator model was then compared to the initial model and no significant deterioration of fit was noted (X²=6.41, df=8, p=.60; RMSEA=0.00; CFI=1.00), also suggesting a good fit. Indirect and direct street victimization predicted PTSD (OR=1.32, p=< .05; OR=1.95, p<.001), respectively; however, only direct street victimization predicted depression (OR=1.47, p<.01). Multiple-street-victimization (i.e., experiencing both direct and indirect victimization) significantly predicted meeting criteria for PTSD (OR=2.29, p<.05).

Conclusions and Implications: Given the dearth of instrumentation available for researchers and practitioners to assess homeless youths’ victimization experiences, these findings demonstrate that the development of a street victimization index delineating direct and indirect street victimization may be useful in predicting homeless youths’ mental health outcomes. Utilizing this two-indicator index of street victimization will enable future empirical investigation of street victimization experiences and its relation to various outcomes among homeless youth. In addition, the index may have utility for service agencies seeking brief screening tools to facilitate trauma-informed mental health treatment.