According to Liptak, one conclusion to be extracted from the Garrett study is that "[t]he courts performed miserably in ferreting out the innocent."
To the extent that Garrett's study discusses the rate of wrongful conviction of the innocent -- thus far I have only skimmed Garrett's paper --, Michael Risinger's paper (scroll down in the page to reach the link) should be consulted.
Liptak asserts, in passing, that "rape [is] the classic crime in which DNA can categorically prove innocence." This is a familiar mistake. But sophisticated observers such as Liptak should not make it. DNA cannot "categorically" prove innocence -- anymore than it can "categorically" prove guilt. For example, it is well known (or it should be well known) that (i) the presence of one male's semen in a woman's vagina does not rule out the possibility that another man raped that woman (whether the man who was the source of the semen found in the woman and tested did so or not) and (ii) a laboratory may have tested the wrong semen sample. Is it necessary to recite such possibilities? The fact that in particular circumstances the results of a DNA test may be highly persuasive evidence of innocence does not change the fundamental point: DNA evidence is NEVER conclusive of ANYTHING. Tiresome. And a dangerous misunderstanding.
Liptak notes that Garrett concluded that "[t]he leading cause of the wrongful convictions was erroneous identification by eyewitnesses, which occurred 79 percent of the time." What are we to make of this statistic? Does it matter how frequently eyewitness identifications resulted in the conviction of the guilty? (I assume Garrett discusses this point. [Let me emphasize that I have not yet carefully scrutinized Garrett's study. So I really am NOT saying or even intimating that Garrett has not discussed the questions I raise here.])
Note: unless one demands absolute factual certainty (which is unattainable under any circumstances), on the basis of the above-quoted statistic alone -- or on the basis of that statistic and the additional statistic that in 1/4 of the wrongful convictions eyewitness identification was the sole evidence --, on the basis of such statistics alone one would not conclude (I would think) (i) that eyewitness identification evidence should be barred or (ii) that eyewitness evidence alone should never suffice for a conviction. (However, further arguments and considerations might justify the latter conclusion -- the conclusion that eyewitness evidence [e.g., of just one eyewitness] should not be suffice for a conviction. [Again, I want to make crystal clear that I am NOT asserting that Garrett does not discuss such issues. I have not yet studied his study.])I have lightly skimmed SOME parts of the Garrett study. The part of the study I enjoyed most (due, surely, in no small part to my personal theoretical prejudices) is the part in which Garrett stresses the importance of pretrial fact investigation and the substantial resources that are required (including, I would say, the diligence, discipline, and imagination of defense counsel) to conduct effective pretrial fact investigation.