How should intellectual impairment be assessed in death penalty litigation?

Joseph Clifton Smith, regarded as a “slow learner” since first grade, spent his school years in special education classrooms designed for those with intellectual or learning disabilities. In 1997, when in his late 20s, he beat a man named Durk Van Dam to death while stealing his tools and $140. The U.S. Supreme Court ruled in Atkins v. Virginia in 2002 that executing individuals with intellectual disabilities violates the Eighth Amendment’s prohibition on cruel and unusual punishment. To avoid the death penalty under Alabama law, Smith would need to show an IQ at or below 70, measurable deficits in adaptive behaviors, and manifestation of an intellectual disability before age 18. Smith’s IQ had been tested five times during his school years and the trial stage, with scores ranging between 72 and 78, but no single composite score was below 70.

IQ scores are reported, however, with a standard error of measurement (SEM) that accounts for variability in test performance due to environmental distractions, individual factors like mood or fatigue, or slight measurement imprecision. As such, an IQ score is understood to occur within a confidence interval of plus or minus several points—typically 3 to 5 points on most standardized tests of intelligence. A 3-point confidence interval around Smith’s lowest score of 72 meant that his score might fall as low as 69. The district court considered the range of scores as well as his adaptive behavior deficits and concluded that executing Smith would be a violation of the Eighth Amendment. The 11th Circuit, also using a holistic approach that considered multiple IQ scores and the SEM around those scores, affirmed the district court’s finding that Smith is intellectually disabled. The Supreme Court is now being asked to consider how intellectual impairment should be assessed in death penalty litigation, particularly whether there is a bright-line rule or whether SEMs should be applied and scores considered more holistically.

Psychologists are central to assessing the criteria set forth in Atkins v. Virginia and will want to consider how to deal with variability in patient IQ scores. Like Smith, some individuals will have multiple and varying IQ scores on record, and use of confidence intervals will broaden that range of scores even further. Clinicians will need to consider how to interpret and report discrepancies among test scores. Anette Bünger, PhD, and her colleagues at the University of Basel—who found that score discrepancies in IQ tests tend to be the rule rather than the exception—recommend against the use of rigid cutoff values, particularly in high-stakes decisions. Bünger and colleagues recommend the use of the 95% confidence interval, as it allows for the best trade-off between comparability and accuracy when comparing IQ scores (Journal of School Psychology, Vol. 88, 2021opens in new window). The American Association on Intellectual and Developmental Disabilities 2021 manual recommends consideration of a range of scores, noting that the scoring trends and clinical information are more telling or informative than any single score (Schalock, R. L., et al., American Journal on Intellectual and Developmental Disabilities, Vol. 126, No. 6, 2021opens in new window). The Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) Text Revision does not recognize a hard cutoff score for a diagnosis of intellectual disability and instead suggests a score of “around 70,” consideration of confidence intervals in test score reporting, and analysis of adaptive functioning.

Ultimately, Hamm v. Smith underscores the critical need for courts to align legal standards with clinical practice by interpreting IQ scores within their scientific limits and giving full weight to adaptive functioning, ensuring that constitutional protections for individuals with intellectual disability are meaningfully upheld.