It appears that when context is taken into account, and when searching interviews are part of the methodology, useful information about the prevalence and the meaning of violence can be garnered. Dobash and Dobash note that women’s accounts ‘reveal the nature of men’s violence, the sources of conflict leading to attacks, their own emotions and reactions’.[521] Men’s accounts in turn show them minimising their own violence and denying responsibility for their actions.[522] Hearn, for instance, quotes one man as saying: ‘I wasn’t violent… I picked her up twice and threw her against the wall. .. I’ve never struck a woman. . .’[523] Another said he had only been ‘really’ violent twice, although he admitted to slapping his victim frequently: ‘I don’t see slapping as being really violent.’[524]

Family violence research is not designed to reveal attitudes like these, and Dobash and Dobash[525] argue that family violence methodology gives rise to a skewed picture. The act-based approach relies on lists of items designed to mea­sure conflict and abuse. The main instrument used is the Conflict Tactics Scale.[526]

This research tool does not distinguish between serious and trivial consequences. Nor can the meaning of acts and their outcome be discerned.[527] The Conflict Tactics Scale ignores motivation such as self-defence.[528] In addition, act-based measures often do not reveal frequency or seriousness.[529] The studies also conflate physical and sexual violence with behaviour such as shouting. And although Dobash and Dobash agree that non-violent acts of abuse are significant, they argue that it can be misleading not to separate them from physical violence.[530] Indeed, Dobash suggests that the ‘conflation of physical attack with conflict, intimidation and threats . . . may be a primary source of the notion that men and women are equally likely to be “violent” to an intimate partner’.[531]