We are regularly asked how many innocent people are in prison in the United States. It’s impossible to know the exact rate of wrongful convictions in our country. To make a sound estimate of a rate of wrongful conviction, one must identify all cases of actual innocence (the numerator) among the total population of similar cases (i.e., the wrongly convicted and the correctly convicted in similar cases, the denominator).
The best available study to date by Sam Gross and colleagues in 2014 reported a wrongful conviction rate of about 4% for capital cases. Of course, this is just the best estimate based on the information available to us at this time. This conclusion would be strengthened by additional data-based studies using defined populations and efforts to maximize the sensitivity and the specificity of the method(s) used to determine wrongful convictions.
Read more about the study below:
Study Summary: This study is based on 117 documented exonerations among 7,482 cases in which a defendant was sentenced to death in the United States from 1973 to 2004. This study population represents a well-defined group of criminal convictions within which extensive efforts were made to identify wrongful convictions; a disproportionate number of the wrongful convictions identified in the United States have been discovered in the relatively small number of cases in which defendants are sentenced to death. Appropriate statistical modeling was used to account for changes in the denominator due to changes in status over time (e.g., resentenced to life in prison). Based on this analysis, the authors made a “conservative estimate” of a rate of wrongful conviction of approximately 4% among people sentenced to death. This estimate reflects a process that is uniquely capable of detecting criminal justice system errors because of the increased scrutiny and advocacy around capital cases, as opposed to the vast majority of cases that involve little or no post-conviction review or advocacy. The nuances of the study population and implications regarding the generalizability of the results are thoroughly discussed (e.g., capital defendants not sentenced to death may be more likely to be innocent because sentencing juries may opt for prison rather than death in cases where they have lingering doubt; people sentenced to death receive far more resources and advocacy than people receiving other sentences).
Take-Away Message: The estimate = 4%, based on 117 documented exonerations among 7,482 people sentenced to death in the United States from 1973 to 2004. The strength of this study lies in the fact that it is based on a clearly defined population and draws both the numerator (exonerations of people under threat of execution) and denominator from this population (people sentenced to death), and the authors use an appropriate analytic method to account for varying lengths of follow-up and changes in the population over time (Gross et al. 2014).
Reference:
Rate of false conviction of criminal defendants who are sentenced to death.
Samuel R. Gross, Barbara O’Brien, Chen Hu, and Edward H. Kennedy. Proceedings of the National Academy of Sciences of the United States of America. 2014; 111(20):7230–7235.
* Special thanks to the Innocence Project and the Innocence Network for helping gather this information