This is, of course, and empirical question, but it is far from clear just how to study it empirically. In order to have an idea how to do so, we would need to know more clearly what's meant by "the number one enemy". Do you mean the principal cause, statistically speaking? Or something else? Are you asking the question universally, or do you have in mind specific populations and specific time frames? And then, are you asking about the incidence of irresponsible actions, or about the capacity of persons to act responsibly (e.g. for their actions to be under their "rational control"). In other words, more precision has to be given to the question in order even to consider how one might investigate it empirically.
As I philosopher I'd be tempted to look at the meaning of 'responsible'. What do you mean here? Which language register? Moral, legal, everyday notion? It will affect how you can look at this question in an empirical study.
Of course there are many concepts to define, or rather make stipulated definitions of, before starting the study. One must nevertheless remember that there are quite a difference of "the experienced need of definitions" in common use of these concepts. Earlier interview studies I´ve conducted started with for example just responsible and than splitted into e.g. legal, moral etc.
Writing as a clinician from Toronto, currently attending a trauma conference in the Netherlands (www.estd2016.org) I would be inclined to say that 1 and 2 are of more central importance and that 3 and 4 mostly arise out of 1 and 2. Of course research has been attempting to tease this out for a long time ... and there is certainly political pressure on researchers to obfuscate the contributory role of developmental trauma and poverty on victimization, perpetration and bad things in general ... but then again, often it is in how we face trauma that we truly become moral and good ... so maybe this question is just too big to answer in a multiple choice format.