I believe the general principle, just not the actual numbers.

Another reason for skepticism: googling "ward clapham recidivism" I found a bunch of pages suggesting the recidivism rate dropped not to 8% but to 5%, an equally unsourced and implausible-sounding number.

Of course recidivism rates depend on what you're measuring, anyway. 70% of burglars are re-arrested for burglary within three years of release, while only 3% of rapists and 1% of murderers get re-arrested for the same crimes.

> 1% of murderers get re-arrested for the same crimes.

The latter may well be due to the combination of (1) murder is (mostly) a young man's crime and (2) murder sentences are typically long enough that folks aren't young when they're paroled/released.