In the study the author linked to, it says that the recidivism rate for sex offenders is 5.3% -- lower than most of the other rates that the author mentions. He

In the study the author linked to, it says that the recidivism rate for sex offenders is 5.3% -- lower than most of the other rates that the author mentions. He simply uses the "only 5 percent of new sex offenses..." statistic to make a point, and view the data from another angle. There's nothing misleading about it, and the author didn't misinterpret the data. Coincidentally, the numbers happen to be similar. When you say "It's obviously true that "those who commit sex crimes are more likely to reoffend" than the average citizen." to point out the author's logical "mistake," I believe you might be misunderstanding the data as well. You have some number of sex offenses, SO, and some population P, that contains some number of sex offenders, P_O. .053 x P_O = sex offenders who reoffend SO x .05 = Sex offenses caused by reoffenders SO x .95 = Some subset of the general population, P, that commits a sex offense. Only if .95SO = .053(P-P_O) does the rate of reoffense by sex offenders equal the rate of sex offense by the general non-offender population (with some simplifications like "people commit one offense at a time, then are persecuted, the only people who have committed two offenses are reoffenders"), which requires a LOT more information to decide, and is probably not true