Whenever some academician tells the media that this program or that program has “reduced recidivism,” or that “this group of offenders aren’t likely to commit more crimes” there are three questions you should always ask:

  • how long were the offenders tracked after they got out of prison?
  • how were offenders selected for (or excluded from) study?
  • who paid the academician?

I have an especially hard time trusting studies that are designed to test one specific program or sentencing initiative.  Such studies are usually designed by people who have a vested interest in proving the program a success — either the program directors themselves or some professor or consulting firm hired to evaluate their outcomes.

It’s sort of like telling a bunch of ambitious eleventh graders to grade their own performance on the SAT’s . . . based on effort.

Unfortunately, there is no graveyard where skewed studies go to die: they live on in debates about recidivism, sentencing, and crime.  This is how myths like “sex offenders almost never re-offend” seep out into the conventional wisdom.

How do you cook the books on recidivism? You follow tiny pools of offenders.  You pick offenders who have already shown initiative by enrolling in a program or being admitted into one — self-selecting, ideal participants.  You use partial information: convictions instead of arrests; post-plea sentencing instead of pre-pleaded charges.  Mostly, you follow offenders for very short periods of time after they are released, like, down the street to the first stoplight.

When you don’t do these things, this is what the headline looks like:

Recidivism rate worse than statistics indicate, Memphis-area study finds

20 years of research discovers 81 percent of former inmates end up back behind bars

Yikes.

Jeff Smith had been free of drugs for four years. Two of those years were during a stay at the Shelby County Correction Center and two were while working at the Salvation Army after his release from jail.

It was at the Salvation Army that Smith, 54, says he felt “a sense of purpose for the first time in years.” He was doing what he says he loves best — working as a carpenter and furniture refinisher. And he counseled other former inmates to try to keep them from repeating their mistakes.

Smith wishes he had followed his own advice. “I was tempted by the devil, and I failed,” he says. Carpentry, counseling and church services at the Salvation Army weren’t enough to break the “revolving-door” cycle that means, like Smith, up to 94 percent of former inmates will be rearrested and up to 81 percent will wind up behind bars again.

94% re-arrest rate.  This is from a 20-year study that recorded every re-arrest and re-conviction, avoiding the “partial information” scheme.  The study itself was conducted by people who have a program of their own to promote: they claim that their moral reconation therapy (MRT) resulted in a 25% decrease in recidivism:

About 94 percent of inmates receiving only standard counseling had been rearrested and 82 percent of them wound up back behind bars.  Of those receiving MRT therapy, 81 percent had been rearrested and 61 percent again wound up behind bars. It was reduction of about 25 percent from the group that did not receive MRT therapy.

Well, OK.  It’s not that I think that there’s no such thing as rehabilitation.  Consequences and 12-steps and therapy do work.  But I’d need to know a lot more about their selection process to buy the 25% claim.

Besides, when anti-incarceration activists claim that we save X amount of money by not incarcerating someone, that’s just untrue.  Most offenders receive significant social service dollars — housing, medical, food stamps — when they are out of prison as well, not to mention the price of policing them and the costs that arise every time they commit an additional crime, which 94% of them apparently will do.  Offenders who return to abusing substances when they get out of prison are particularly costly as their health deteriorates and their habits drag down the families and neighborhoods around them.  Innocent bystanders and misinformed taxpayers pay the tab either way.

Without acknowledging these costs, statements like this are, frankly, meaningless:

[T]he cost of housing an inmate like Smith is more than $24,000 a year, so cutting total costs by 25 percent would mean a huge savings.

Yet public policy debates rise and fall on questionable claims like these. The media needs to do a much better job of skeptically approaching all research claims.  After all, if there is reliable research showing that everything policymakers have been believing is not only wrong, but staggeringly wrong, the debate needs to be re-calibrated:

Tennessee Department of Correction studies show recidivism rates of about 51 percent over a three-year period, and national studies show recidivism averages of roughly 65 percent over three years. But [Dr. Greg] Little and [Dr. Kenneth] Robinson say the numbers keep going up over time, and the numbers are higher because most studies don’t count re-incarcerations that took place in other states or in courts other than the original case. For instance, an inmate released on state probation or parole is seldom counted as a recidivist if later jailed for a federal crime.

There is a very large difference between 51% recidivism and 94% recidivism.  You don’t need to throw out the rehabilitation baby with the research bathwater just because the research bathwater is hopelessly dirty, but you should wash the baby in clean water.