Right Reason in the Integration of Group and Single-Subject Research in Behavioral Disorders

Abstract

In this discussion of quantitative research synthesis (meta-analysis), past criticisms of meta-analysis are reviewed along with the ways in which they have been addressed. The problems associated with synthesizing single-subject research are discussed, particularly the difficulties in obtaining a metric equivalent to the effect size calculated for group-design research. After analyzing recommended approaches, the authors endorse the percentage-of-nonoverlapping-data metric. Their earlier single-subject synthesis showing limited efficacy for social skills instruction is then reviewed. The findings are discussed in relation to the nature of single-subject research, the proper role of research integration, and the influence of subjectivity in interpreting research findings. Finally, possible reasons for disappointing findings regarding the effectiveness of social skills instruction are explored. The authors conclude that, until expressed concerns are remedied, it is neither incorrect nor illogical to question the efficacy of social skills instruction.