I highly recommend reading this response to the recent BJS report on Sex Offender Recidivism, it does a great job of explaining how language can be used to subvert the findings of official reports that are politically unpopular:
I also would add one other objection. To say that a sex offender is three times more likely to commit a new sex-offense intentionally leaves out the most critical information in that statistic: How likely someone is to commit a new sex offense.
If someone, for instance, is 1% likely to commit a new sex offense and you are three times more likely, your actual risk is still extremely low (just high relative to the person whose risk is 1%).
It is irresponsible to present this statistic out of content IMHO.
This is all in reference to this new report by the Bureau of Justice Statistics: