If you mean, you'd expect the 0.5% of men using the facility to commit more than 4% of the assaults (which is my intuition too) I think that comes from two sources:
- I haven't looked at where the 90% of assaults committed by men comes from, but that sounds low to me: perhaps it depends on the kind of assault we're talking about
- sadly, I think it's reasonable to think that the 0.5% of female-toilet users who are men are likely not to be typical of men, but instead to be far more likely than the average man to commit assault. Good men stay out so that bad men stand out.
Also, in any individual workplace female toilet, it's quantised. It's not that each one will have 0.5% of its users being male, because few workplace toilets are used by as many as 200 people (which is what it would take for 0.5% of users to be as many as 1 user). Far more likely it's either 0% of users are male (no transwomen around), or it's a much higher percentage, if there is a man who uses that particular toilet. Overall it's a silly argument anyway - we're not really talking about the actual statistical likelihood of being assaulted, we're talking about the fear of being assaulted which comes from a whole-society set of data plus personal experience, and we're talking about privacy and dignity.