I should begin by noting that, to the extent that John Hill fully captures it in his Providence Journal article, I support programs like OpenDoors’s “9 Yards.” That said, even with programs on which there is broad agreement, somebody has to make sure the public is reviewing its success honestly. To that end, this data ought to raise questions for the careful reader:
The OpenDoors report tracked how 50 parolees who went through the program in its first three years fared in their first year out of prison, compared to 51 parolees who didn’t.
In that first year after release, 27 percent of 9 Yards participants reoffended and were sentenced, versus 38 percent for those in the control group. Six percent of 9 Yards inmates were convicted of new felonies, compared to 21 percent of the control group.
The question not answered is whether the ones signing up for the program were just more likely not to re-offend to begin with. If that’s the case, then the study really doesn’t tell us much of anything, and the news article risks being mere boosterism.
Although the numbers of participants appears to be different than what Hill reports, it looks like this is the report, and it notes that, while there was an initial screening process for eligible participants, selection into 9 Yards was random. So, there is no “selection bias” based simply on who entered the program and who didn’t.
In other words, the good news is indeed good.