The Double Interview Task: Assessing the Social Communication of Children with Asperger Syndrome

Reliability

Six interviews were randomly selected for interobserver reliability. This represented 30% of the total interviews. Two graduate students in speech language pathology served as reliability judges, one for transcription and one for coding. Reliability was calculated by dividing the total agreements by the total number of agreements plus disagreements.

Reliability measures were obtained first for transcription and then for coding. The transcription reliability for participant # 1 was 98%, participant # 8 was 100%, participant # 9 was 96%, participant # 10 was 100%, participant # 14 was 97%, and participant # 16 was 100%. The across participant reliability was 98%.

Disagreements in transcription were resolved prior to coding by discussion and review of the videotapes. Intercoder reliability was then used to measure the agreement when the observers assigned codes to the same transcript. The intercoderagreement for participant # 1 was 86%, participant # 8 was 88%, participant # 9 was 88%, participant # 10 was 88% participant # 14 was 90% and participant # 16 was 88%. The across participant reliability was 88%.