The Norway-vs-US recidivism comparison (20% vs 76.6%) is one of the most widely shared 'proof points' for rehabilitation-oriented prison policy. This piece checks whether the headline numbers are actually measuring the same thing, in the same populations, over the same time window -- and finds they are not.

Compared official Norwegian reconviction statistics (2-year and 5-year windows, 2018-2020 cohorts) against US Bureau of Justice Statistics multi-state release-cohort studies (2005 and 2012 cohorts, 5-year and 9-year follow-ups).

Cross-referenced incarceration rate and population-composition data from World Prison Brief and Prison Policy Initiative's global incarceration dataset. Reviewed definitional differences (rearrest vs. reconviction vs. reincarceration) that make cross-national recidivism comparisons methodologically fragile.

6%. Four times better rehabilitation, case closed, ship every warden to Halden. I want to like this claim. My colleague Isabel Vargas at COLMEX sent me the volleyball photo three separate times this month, each with more exclamation points than the last. But close reading of the actual government reports gives you a much less flattering national morality play.

Start with what's actually being measured. Norway's own reconviction figures -- not "recidivism" in the loose meme sense, reconviction, meaning a new sentence handed down -- run about 16 to 18 percent at two years post-release for recent cohorts, climbing toward 25 percent at five years for the 2018 release cohort.

The US Bureau of Justice Statistics tracks a different outcome entirely: rearrest. Their 2012 34-state release cohort found 71 percent of released prisoners were rearrested within five years, and 46 percent returned to prison for a new sentence or a supervision violation in that same window.

Compare Norway's 25 percent reconviction-with-new-sentence at five years to the US's 46 percent return-to-prison figure at five years and the gap shrinks from "4x" to under 2x.

Compare it instead to the 71 percent rearrest figure, which includes arrests that never produce a conviction, and you're comparing a strict outcome to a loose one and calling the difference a national character trait.

Then there's who gets counted at all. Norway incarcerates about 54 people per 100,000 residents. The US incarcerates roughly 629 per 100,000 -- more than eleven times the rate.

When you lock up a much larger share of your population, you mechanically pull in more first-time, low-level, and marginally-deterrable offenders alongside serious repeat offenders, diluting your cohort's baseline risk profile in ways that push measured recidivism in unpredictable directions.

Norway's ordinary sentencing structure also caps most prison terms well below US norms, with preventive detention reserved as a narrow, judicially reviewed exception, meaning its released population skews toward people who served short-to-moderate terms in a system that already filtered out the highest-risk profile upstream through different charging and diversion practices.

None of the viral comparisons I've seen control for any of this. They compare an apple output rate to an orange output rate and attribute the entire difference to volleyball.

I am not arguing Halden's model is fake, or that architecture and staff-to-inmate ratios don't matter -- there is real evidence that shorter, more normalized incarceration reduces institutionalization harm on its own terms.

6 percent headline, is doing far more rhetorical work than the underlying data supports, because the two figures were never measuring the same event, over the same time window, in populations selected the same way.

Prison Policy Initiative's own global incarceration dataset makes this point implicitly every year by treating incarceration rate, not recidivism percentage lifted without its denominator, as the more policy-relevant number to interrogate.

If you want to know whether Norway's model would cut US recidivism, the honest answer is nobody has run that experiment, because you cannot retrofit a 54-per-100,000 incarceration philosophy onto a system built around 629-per-100,000 and expect the same population mix to walk through the door. What we actually have is two countries measuring two different outcomes, over two different time windows, in two differently-selected populations -- and a very good story.