By Alex Fradera
Psychology can help people one person at a time, but it also holds the promise of changing society at a mass scale, through campaigns to change attitudes and behaviour. One such endeavour is the development of programmes to reduce the rates of sexual assault of women on university campuses. But in a literature review in Aggression and Violent Behavior, researchers from the University of California make the case that such programmes may not just be ineffective, but counterproductive.
In 2013 the US passed its Violence Against Women Act; in response most US university campuses launched programmes that aimed to reduce sexual assault by raising awareness, changing attitudes and behaviours and encouraging bystanders to take a stand. But there has been little systematic evaluation of these programmes, and the evidence for any benefits is thus far thin. Of the 140 college-based behaviour change programmes studied by the Center for Disease Control, it recommended only three on the balance of evidence; all three focused on changing bystander action and two of the three merely showed “promise of effectiveness”. Other reviews have produced slightly more positive conclusions but often with weak effects or without evidence of actual behaviour change.
“Newish interventions have a thin evidence base, more work needed” is not a clamouring call for alarm, but based on their review of the field, Neil Malamuth and his team suspect there may be something worse at play: a boomerang effect.
It’s certainly become clear that well-meaning interventions in other contexts can have the opposite effect to that intended; we’ve previously covered instances of backfiring interventions that aimed to reduce prejudice, correct false beliefs on vaccine, or encourage employees to behave better in the workplace.
The reason boomerang effects are such a concern in the domain of sexual assault is that most interventions are being designed and evaluated around a broad target group – men – rather than looking at effects on the most high-risk individuals. Yet we know from previous evidence that the people most likely to respond contrarily are those most disposed to the problem behaviour in the first place. For instance, an intervention intended to reduce acceptance of violent video games in students was effective for many participants, but post-intervention, those who had been classified as living a delinquent lifestyle became more enthusiastic about playing the games. Another study demonstrated that an anti-violence campaign was effective in increasing opposition to violence in people low in trait aggression, but those who were highly aggressive became more enthusiastic about violence.
Boomerang effects seem to happen when participants experience reactance: anger stemming from the sense that a mode of life or thinking is threatened combined with “motivated reasoning” – like finding reasons to doubt the source of the message. In the midst of reactance, the person sees themselves as defiantly resisting preachy attempts to deny them their entitlement.
Is reactance likely to happen with the men most likely to commit sexual assault? Well, consider this profile of high-risk individuals: they tend to have hostility towards women, experience sexual arousal when physical force is involved, and have an impersonal sexual orientation (intimacy is not a prerequisite for what they seek in sex). As well as these sex-related factors, they tend to be anti-social, highly entitled and narcissistic. Evidence suggests these individuals are less likely to respond to cues of sexual disinterest, feeling that women are really into them or will be eventually, and they react more aggressively to interpersonal rejection. This seems a potent recipe for reactance.
There isn’t much evidence from existing anti-sexual assault programmes on specifically how these high-risk individuals respond, but smaller-scale experimental work reveals troubling signs. Messages promoting gender equality and greater consideration for women lead to lower aggressive tendencies for men already low in sexist attitudes, but more aggression among those who held more sexist attitudes from the start. A video intended to reduce acceptance of rape myths and increase victim empathy worked for men who had no history of sexually aggressive behaviour, but the high-risk men showed a boomerang effect. On the other hand, some experimental studies have shown some intervention content can have the desired effect on high-risk individuals, e.g. improving victim empathy. But if you hit the target on some occasions, and shoot yourself in the foot on others, you would want to take a good look at your technique before rolling it out nationwide.
The danger is that interventions could produce impressive-seeming results simply because the low-risk majority perceive sexual assault as even more deplorable than they already did – while the (roughly 30 per cent) of higher-risk men become even more intent on their agenda. Of course, such interventions may prevent certain sexual assaults – for example, through limiting predation opportunities through galvanising bystanders to act. But surely we should be working to maximise impact on those likely to commit the bulk of the assaults, rather than fire up their motivation still further.