Overstating the Behavioral Effects of the Seattle Social Development Project

I would like to raise 2 issues regarding a recent report of the effects of the Seattle Social Development Project (SSDP) on sexual behavior and an accompanying editorial in the ARCHIVES.^1- 2

First, the article reports data from 144 subjects (the "full-intervention" group) who participated in the SSDP program in grades 1 through 4 as part of a randomized experiment and continued in the intervention condition in fifth grade when the study changed to a quasi experiment. The article states that the SSDP has exhibited "consistently high sample retention rates," with 93% of subjects followed up at age 21 years. For the full-intervention group, the follow-up rate was calculated using the 156 fifth-graders who had parental consent to participate in the quasi experiment as a "baseline." However, involvement of these subjects in the study commenced 4 years earlier. Previous reports from the SSDP do not give the sample size of the full-intervention group at the first-grade baseline, but show that at the second grade follow-up, it comprised 285 subjects.³ Thus, the correct follow-up rate for the full-intervention group at age 21 years was no more than 50%. This low retention rate seriously undermines the methodological integrity of the evaluation.