It was based on percentages, and mostly based on studies done by law enforcement.

It has been a long time since I looked at that data, and it would take some time to look it all up - but I'm thinking that Kenneth Lanning's research into claims of ritual child abuse (over 20 years ago) distinguished between "situational offenders" (non-pedophiles) who committed about 80% of abuse and "fixated offenders" (pedophiles) who committed the other 20%, and in distinguishing between the two groups used pretty much the same methods (including penile plethysmography) that in separate studies found 90% of men somewhat attracted to prepubescent girls and 20% to 30% of men with an attraction to prepubescent girls equal to or greater than their attraction to women. I'm thinking there were some other studies that found similar percentages of abuse by "situational offenders", including a study by Australian law enforcement about 10 years ago.

Of course, these criminologists never put the two numbers together, so they never came to what is for us the obvious conclusion: that at worst, pedophiles commit sexual abuse at the same rate as teleiophiles, and most likely are a bit less likely to abuse children.

Perhaps someone else with a better memory and bookmarks can give you some of the particulars - or you can search for it yourself. I have other priorities at the moment.