In the DP controversy, it's important not to take the sweeping claims made by "studies" at face value. When researchers have a pre-destined agenda, it's easy to produce tendentious results.

For example, in a DP case, it is very common for an appellate court to ask for additional findings from the trial court, get the findings, then uphold the verdict. So if ANY court at ANY stage of the case returns the case back down a level for more information or procedural verification, your study counts that as a "mistake" rather than admirable diligence and due process.

In addition, many of the DP reversals were for purely ideological grounds, having nothing to do with the facts of the case. The notorious justice Rose Bird single-handedly accounted for about half of California's DP reversals, IIRC, and she openly and candidly admitted she didn't care about the facts of the case. She was a liberal activist judge who was against the DP, so routinely reversed all DP cases. So that's half of California's "error rate" right there.

Also, the Liebman study inexcusably mixes reversals of convictions with reversals of sentences. E.g., a man clearly guilty might have his DP sentence reversed, but since his life in prison sentence still holds, prosecutors may decide not to undergo the cost and risk of retrial. But this should not be confused with a reversal of conviction, where someone is found not guilty of the charges themselves and freed. But the study adds them in to artifically inflate and sensationalize the numbers.

http://www.ncpa.org/pi/crime/pd061600c.html

http://www.lib.jjay.cuny.edu/docs/latzer.htm