The criminal justice system\u2019s most fundamental promise is that innocent people will not be punished. Yet decades of exoneration data show that innocent people are convicted, imprisoned, and sometimes sentenced to death. That raises a direct empirical question: when innocent defendants appeal their convictions, are appellate courts any better at identifying and correcting these errors than trial courts were at preventing them?<\/p>\n\n\n\n
Convicting an innocent person is a classic Type I error: the system finds guilt when guilt is not present. In criminal law, that error is not a minor statistical inconvenience. It is a profound system failure. An innocent person loses liberty, reputation, years of life, and sometimes far more. The legal system is supposed to recognize the gravity of that mistake by building in layers of review. Trial courts are supposed to prevent error in the first instance, and appellate and habeas courts are supposed to correct serious mistakes when they occur. If those reviewing courts cannot distinguish wrongful convictions from ordinary convictions in the most serious cases, then one of the system\u2019s central safeguards is not functioning as promised.<\/p>\n\n\n\n
Brandon Garrett\u2019s article Judging Innocence<\/em> provides unusually useful data for answering that question. Garrett studied the first 200 people exonerated by postconviction DNA testing in the United States and compared a subset of those cases to a matched set of similar non-DNA cases. His results permit us to ask a fundamental question:<\/p>\n\n\n\n Do defendants later proven innocent obtain appellate or postconviction reversals at higher rates than similarly situated defendants who were not later proven innocent?<\/p>\n<\/blockquote>\n\n\n\n If appellate courts can identify miscarriages of justice, one would expect DNA exonerees to have received relief more often than similar defendants. If they do not, that suggests ordinary appellate review is poorly equipped to distinguish wrongful convictions from other serious criminal convictions.<\/p>\n\n\n\n Garrett\u2019s full innocence group included the first 200 people exonerated by postconviction DNA testing in the United States, from Gary Dotson\u2019s exoneration in 1989 through Jerry Miller\u2019s exoneration in April 2007. The study remains unusually valuable because it does not just describe exonerees; it pairs a subset of those innocence cases with a structured matched comparison group.<\/p>\n\n\n\n But not all 200 cases can be used for this specific comparison. Garrett focused much of his appellate analysis on the 133 exonerees who had written judicial decisions, because without written decisions he could not reliably determine what claims were raised, how courts ruled, or why relief was denied.<\/p>\n\n\n\n For the matched comparison, Garrett narrowed the data further to 121 noncapital innocence cases with written decisions. He excluded capital cases from the comparison because death-penalty cases have unusually high reversal rates and raise separate procedural issues.<\/p>\n\n\n\n He then created a matched comparison group of 121 noncapital cases. These were not \u201cknown guilty\u201d cases. They were cases with no later DNA evidence proving either innocence or guilt. Garrett describes them as standing in for the broader universe of serious criminal defendants who never obtain DNA testing.<\/p>\n\n\n\n The matching procedure was designed to make the comparison cases resemble the DNA-exoneration cases. For each of the 121 noncapital innocence cases, Garrett searched Westlaw for a case from the same state, involving the same or comparable conviction offense, with a reported decision in the same year. He accepted the first case meeting those criteria and only later checked whether the conviction had been reversed.<\/p>\n\n\n\n That design matters. The comparison does not simply ask whether exonerees did better than a random group of criminal appellants. It asks whether exonerees did better than defendants convicted of similar serious crimes, in the same jurisdictions, during the same time periods, whose cases also produced written appellate or postconviction decisions.<\/p>\n\n\n\n The outcome is whether the defendant received a qualifying reversal before DNA exoneration. Garrett defined a reversal in a strong sense: a court ordered a new trial or vacated the conviction, and that relief survived appellate review. The analysis therefore does not count every favorable comment, remand, sentence modification, or later DNA-based vacatur. It asks whether ordinary appellate or postconviction review corrected the conviction before DNA testing did.<\/p>\n\n\n\n The results are almost identical across the two groups.<\/p>\n\n\n\n The difference is negative 0.8 percentage points: the DNA-exoneration cases had one fewer reversal than the matched comparison cases. The difference is not statistically significant. Garrett\u2019s own conclusion is that exonerees \u201cfared no better\u201d during review than matched rape and murder defendants.<\/p>\n\n\n\n This is the core result. In the comparable noncapital cases, defendants later proven innocent by DNA evidence did not receive reversals at a higher rate than similarly situated defendants. The appellate and postconviction process did not distinguish the innocent from the comparison cases in any meaningful way. <\/p>\n\n\n\n This finding does not prove that appellate courts never detect innocence. Some innocent defendants did receive relief. But the result shows something more important and more disturbing: the courts were not doing better than the background rate in the very cases where we now know the convictions were wrong.<\/p>\n\n\n\n That means ordinary review was not reliably performing its core error-correction function. Innocent defendants were no more likely to benefit from it than other serious criminal appellants.<\/p>\n\n\n\n The result is especially troubling because these innocence cases are not speculative. These are not merely defendants with plausible innocence arguments or lingering doubt. They are people later proven innocent through postconviction DNA testing. If appellate courts were effectively identifying wrongful convictions, their reversal rate should have been meaningfully higher. It was not.<\/p>\n\n\n\n Garrett\u2019s broader findings help explain the pattern. The DNA exonerees were often convicted based on evidence now known to be unreliable: eyewitness identifications, faulty forensic evidence, informant testimony, and false confessions. Yet appellate courts often did not directly revisit factual guilt. They reviewed legal claims, procedural defaults, harmless error, and the sufficiency of the record as it appeared at the time.<\/p>\n\n\n\n That may help explain the failure. A narrow concern for procedural due process is not enough if the reviewing court is focused on technical rules that are only loosely connected to the truth of the conviction. Courts may be performing legal review without performing meaningful conviction review. If so, they are checking whether procedures were followed closely enough, not whether the system has imprisoned the wrong person.<\/p>\n\n\n\n Garrett also found that courts often denied relief after concluding that the evidence of guilt outweighed the asserted error, sometimes describing the evidence as \u201coverwhelming\u201d in cases later proven to be wrongful convictions. That is a serious warning about record-based review. A trial record can appear strong precisely because unreliable evidence was admitted and credited. If appellate courts take the record as given and then ask only whether a particular error seems harmless in light of that record, they may simply reinforce the original mistake.<\/p>\n\n\n\n The implications should not be softened to please the courts. These data show that appellate courts objectively fail to do what they are supposed to do in the most serious cases. They are meant to correct serious trial-level errors, including the gravest error of all: convicting the innocent. But in Garrett\u2019s matched comparison, innocent defendants did not outperform the background reversal rate. On this measure, the system\u2019s supposed corrective machinery did not work.<\/p>\n\n\n\n Garrett\u2019s additional figures make the failure even harder to dismiss as a matter of a few close cases. Among the 133 exonerees with written decisions, only 18 received reversals at all. Seventy percent, or 93 of the 133, received no relief of any kind during their appeals or postconviction proceedings. Garrett also found that 86% of the exonerees with written decisions ultimately had their claims denied. Even when courts reached the merits, they often denied relief. Those are not numbers that suggest a functioning error-correction system.<\/p>\n\n\n\n Nor did the courts reliably correct these convictions once DNA evidence began to point toward innocence. Garrett reports that courts denied at least twelve exonerees relief despite at least preliminary DNA test results excluding them; each was later exonerated only after an executive or higher court granted relief. Forty-one of the 200 exonerees, or 21%, received pardons from state executives, often because they lacked an available judicial forum for relief. Only two received DNA testing and a vacatur through federal habeas corpus. The others generally obtained vacaturs in state courts, typically through newly discovered evidence procedures. Even after biological evidence excluded them, the ordinary judicial process was often unwilling or unable to provide a remedy.<\/p>\n\n\n\n This sharpens the institutional point. A narrow concern for procedural due process is not enough if courts are focused on procedural and technical questions that are only weakly connected to actual guilt or innocence. Appellate courts may be resolving legal disputes while failing to perform meaningful conviction review. If the system cannot separate the innocent from similarly situated convicted defendants, then its error-correction function is not just imperfect. It is misaligned with its most basic purpose.<\/p>\n\n\n\n That also helps explain the rise of conviction integrity units and other specialized innocence-review mechanisms. Their growth reflects a recognition that wrongful convictions require focused factual review. But there is something deeply unsettling about the need for these institutions. Appellate courts and habeas courts were supposed to be our original conviction review units. If separate executive-branch or prosecutor-based mechanisms are increasingly doing work that ordinary judicial review failed to do, that is not merely innovation. It is evidence of institutional deficiency.<\/p>\n\n\n\n Garrett\u2019s data allow a straightforward test. Among comparable noncapital cases, DNA exonerees were reversed in 9.1% of cases, while matched comparison defendants were reversed in 9.9% of cases. The difference is essentially zero.<\/p>\n\n\n\n The result is not merely that innocent defendants sometimes lose. The result is that innocent defendants did not win more often than similarly situated defendants. In these data, ordinary appellate and postconviction review did not reliably sort the innocent from the rest.<\/p>\n\n\n\n That is not just a disappointing empirical finding. It is evidence that one of the criminal justice system\u2019s central safeguards is failing where it matters most.<\/p>\n","protected":false},"excerpt":{"rendered":" The criminal justice system\u2019s most fundamental promise is that innocent people will not be punished. Yet decades of exoneration data show that innocent people are convicted, imprisoned, and sometimes sentenced […]<\/p>\n","protected":false},"author":2,"featured_media":288,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[9,3],"tags":[],"class_list":["post-274","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-criminal-justice","category-law-and-courts"],"_links":{"self":[{"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/posts\/274","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/comments?post=274"}],"version-history":[{"count":3,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/posts\/274\/revisions"}],"predecessor-version":[{"id":287,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/posts\/274\/revisions\/287"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/media\/288"}],"wp:attachment":[{"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/media?parent=274"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/categories?post=274"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.poliscidata.com\/blog\/wp-json\/wp\/v2\/tags?post=274"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n
Garrett\u2019s Dataset and Case Selection<\/h2>\n\n\n\n
The Relevant Outcome Is Reversal Before DNA Exoneration<\/h2>\n\n\n\n
Innocent Defendants Did Not Win More Often<\/h2>\n\n\n\n
Court Outcome<\/th> DNA Innocence Cases<\/th> Matched Comparison Cases<\/th><\/tr> Reversed<\/td> 9.1%<\/td> 9.9%<\/td><\/tr> Not reversed<\/td> 90.9%<\/td> 90.1%<\/td><\/tr> Total<\/td> 121<\/td> 121<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n Why This Result Matters<\/h2>\n\n\n\n
The Problem Appears To Be Structural, Not Accidental<\/h2>\n\n\n\n
Appellate Courts Are Failing at Their Basic Job in the Most Serious Cases<\/h2>\n\n\n\n
The Bottom Line<\/h2>\n\n\n\n