A Mixed Methods Exploratory Examination of Victim Injury and Death: Effect of Weapon Type and Victim Resistance During Sexual Assaults by Strangers

Abstract

Extending prior research regarding victim injury during sexual assault, the current study employed mixed-methods sequential explanatory research design to identify predictors of victim injury and victim death in 361 attempted and completed sexual assaults committed by 72 repeat sex offenders who assaulted strangers. Results from multinomial path analysis indicated that being female and offender coercion increased the likelihood of victim resistance, which in turn elevated the likelihood of victim injury. Divergent from the predictors of victim injury, the risk of victim death increased with victim age, offender alcohol use, and offender weapon possession. Exploratory analyses suggest that certain weapons resulted in higher probability of victim injury or victim death within the context of victim resistance. Exploratory analysis of qualitative data indicated that crime event order was not uniform—in some cases victim resistance preceded victim injury and in some assaults victim injury occurred prior to or was unrelated to victim resistance. The study findings highlight the need for further investigation of victim survival strategies when offenders possess different types of weapons.

Keywords:

• sexual assault
• victim death
• victim injury
• victim resistance
• weapon effects

Notes

1. Scores from the full Grasmick scale and scores of the subset of impulsivity items were analyzed with similar results. Results reported in the text are based on the impulsivity subscale. Results of analysis using the full scale are available upon request.

2. As model fit indices are not provided with multinomial path analysis, path analysis was first conducted with victim injury coded and specified in the model as an ordinal variable with three possible attributes—no physical injury, physical injury, and death—in order to assess general model fit. The path analysis evidenced adequate fit to the data with a nonsignificant Chi-square, χ²(4) = 4.10, p = .39. All other model fit indices demonstrated good model fit (NC = 1.03; CFI = .99; TLI = .99; RMSEA = .01; WRMR = .46) (Kline, 2005).

3. A multinomial path analysis model with all 12 predictors was analyzed to see if the results changed when all measures were included. Results showed that the inclusion of the additional predictors did not alter any of the effects (in terms of sign and significance) of the other variables. Results available upon request.

4. There was no significant difference in victim injury/death across assaults with no resistance and assaults with passive resistance (χ² (2) = .48, NS).

5. There was no significant difference in victim injury/death across assaults with physical only resistance and assaults with verbal and physical resistance (χ² (2) = 2.28, NS).

6. Pseudonyms were used in place of the names of the study participants.

7. When including only assaults by offenders with varied assault outcomes in terms of victim injury or death, there was no significant change in the results comparing differences in mean ratings. Results available upon request.

8. When including only assaults by offenders with varied assault outcomes in terms of victim injury or death, there was no significant change in the results comparing differences in mean ratings. Results available upon request.