There is another possible explanation that is not ruled out by cultural differences, which follows the argument developed by Emile Durkheim in _Suicide_. In cross-cultural comparisons, particular values and norms may vary, but the fact that all units of observation are organized social groups does not. Moreover, it has been demonstrated that level of inegration to a stable community has an effect of crime rates in a community. It is thus plausible that crime rates will be higher in those social groups that are less integrated.
Since men are more likley to be individualistic and anti-social than women who, in turn, tend to be more dependent on social networks in their everyday life activities, it follows that criminality rates rates will be higher among males. That explanation may work for religiou participation, which is one of the key expressions and vehicles of community participation. It should be noted that religious participation is different from religiosity which includes intensity of beliefs. Most surveys can measure the former (albeit inaccurately) - but not the latter.
Of course, that explanation does not rule out physiological deficiencies that make males more anti-social than females.
Wojtek