Efficacy of a Social Intervention on Facial Encoding in Children with High-Functioning Autism Spectrum Disorders
Rodgers, Jonathan David
MetadataShow full item record
This study evaluated the impact of a summer social development program on the facial encoding of emotions in children with high-functioning autism spectrum disorders (HFASDs). A total of 60 children, ages 6-12 years with HFASDs from two summers of a randomized controlled trial were included. Participants were randomly assigned to the treatment or wait-list control group within each summer. Due to some overlap in participants across summers, the current study included only data from the first summer of participation for each case, which resulted in n = 30 treatment vs. n = 30 control. Each participant was photographed using a standardized script making posed facial expressions at the beginning and end of the treatment or wait-list period. Photographs were rated on three metrics assessing various aspects of facial encoding by eight female undergraduate and graduate student raters. Ratings were obtained on (a) the extent to which each facial expression photograph reflected one of the six prototypical emotions (i.e., happy, sad, angry, fearful, surprised, or disgusted), (b) which emotion was portrayed in the photograph (i.e., a forced-choice selection of encoding accuracy), and (c) the overall "oddness" of the facial expression. From these ratings, five measures were calculated to assess components of facial affect encoding. These included two measures of display accuracy, one measure of oddness, and two measures of expression ambiguity. After controlling for pretest differences using an ANCOVA model, the current study found a significant increase in the facial encoding of sadness at posttest for the treatment group. This increase was demonstrated on the extent ratings, forced-choice accuracy and both ambiguity measures, but was not significant for the oddness ratings. Effect size estimates for statistically significant comparisons of adjusted posttest scores were d = .56 to .65. Results for no other emotions reached statistical significance. Strengths, limitations, and directions for future research are discussed.