Public perceptions of internet, familial and localised sexual grooming: Predicting perceived prevalence and safety

Abstract

This paper examines public perceptions of three sexual grooming types: computer-mediated sexual grooming (CMSG), familial sexual grooming (FSG) and localised sexual grooming (LSG). Using data from a national survey of 557 respondents from the United Kingdom, we tested models that predicted perceptions of the prevalence of CMSG, FSG and LSG and the perceived safety of internet, familial and localised grooming spaces. Media-related factors were the most significant in predicting higher levels of perceived prevalence of CMSG and disagreement in relation to safety of internet and public spaces. Knowledge of a grooming victim was most significant in predicting higher levels of perceived prevalence of FSG and LSG and higher levels disagreement in relation to the safety of the home. The findings suggest that the public express too little concern over familial sexual grooming and that initiatives should be introduced to make citizens more aware of the distinctions between types of sexual grooming behaviours, settings and offenders.

Keywords:

• computer-mediated sexual grooming
• familial sexual grooming
• localised sexual grooming
• media and sexual grooming

Notes

1. For a full account of this legislation as well as misinterpretations of what the law can and cannot do, see Craven, Brown, and Gilchrist, 2007; Gillespie, 2004; Ost, 2004.

2. This includes webpages accessed via mobile phone, but not text messages or telephone calls. See McCartan and McAlister (2011) for an overview of mobile phone use and sexual abuse.

3. For example, the CEOP thematic assessment into localised grooming identifies the importance of mobile phones in the grooming process (CEOP, 2011).

4. Online idiolect refers to the specific language patterns formed by internet users which differ from the offline spoken and written word (Williams, 2006).

5. See http://www.survey.bris.ac.uk/

6. More than 40 Cardiff University students distributed the link to the online survey via their facebook pages, yielding a significant return with a wide geographical profile (the average number of connections per student was 254). Students were also instructed to send the link via email to family and friends who, in turn, distributed the link to work colleagues, significantly boosting the age range of respondents. Finally, the authors distributed the link via the Mumsnet and Netmums forums to boost numbers of respondents with children.

7. Where population estimates are provided they should be interpreted with a degree of caution.

8. See: https://aoir.org/documents/ethics-guide/

9. The majority of those reporting personal experience as their main source of crime information also reported employment in criminal justice-related fields.

10. This type of regression analysis is the most appropriate given the data type in each of the dependent variables (ordinal data).

11. Results from the tests for parallel lines, model fit statistics and multicollinearity diagnostics are not shown. Pseudo R ² statistics are presented in Tables III and IV.

12. Data not presented in this paper but available upon request.

13. This conclusion was reached by examining the pseudo R ² values of each submodel. For a discussion of the use of pseudo R ² see Aldrich and Nelson (1984).

14. Data not presented in this paper but available upon request.