An archival analysis of sexual assault victims’ age estimation accuracy when describing stranger offenders

ABSTRACT

Approximately 10% of all serious sexual assaults in England and Wales involve victims and offenders who are strangers. The victims often estimate the stranger offender’s age during police interviews. These age estimations, if accurate, can help identify offenders. This archival analysis examined the accuracy of 546 stranger sexual assault victims’ age estimations. It also examined whether their accuracy can be predicted by victim age – offender age differences, victim age – offender estimated age differences, victim race – offender race differences, victim intoxication, victims’ duration of exposure to offenders, the time delay between assaults and age estimations, whether offenders have weapons, and whether offenders use sighting precautions. Amongst the descriptive findings, we found victims’ mean age estimation error was 4.78 years, the degree of over- and underestimation was equivalent, that only 12.5% of age estimations were within 0.99 years of an offender’s true age, but 90% were within 9.99 years of an offender’s true age. Only victim age – offender age differences and, importantly, victim age – offender estimated age differences predicted age estimation accuracy: as age differences decreased, accuracy increased. These findings can help predict stranger sexual assault victims’ age estimation accuracy in future cases.

KEYWORDS:

• Sexual assault
• rape
• age estimation
• own-age bias
• memory

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Craig Thorley http://orcid.org/0000-0002-3207-435X

Notes

1 This value relates to studies where absolute age estimation error is calculated. Some studies have calculated average age estimation error only, so the calculations included negative values from underestimations and positive values from overestimations. These negative and positive values often cancelled each other out during the averaging, resulting in lower error rates than reported here (see Rhodes, 2009).

2 Fahsing et al. (2004) did not state how long participants were exposed to the offenders for in their studies, preventing direct comparisons between their work and that of Granhag et al. (2013). The same author (Granhag) wrote both articles and mentions that the exposure duration in the 2013 study ‘was considerably shorter than any of the cases examined by Fahsing et al. (2004)’ (p. 929).

3 Laboratory-based studies have shown participants’ verbal description of an offender can be impaired if the offender is armed (see Kocab & Sporer’s, 2016, meta-analysis).