Generalizability analysis of the communication and symbolic behavior scales: A pilot study

摘要

In this pilot study, generalizability theory was used as an exploratory analysis of the various sources of error in the scoring of the Communication and Symbolic Behavior Scales. The different score conversions, raters, and growth of participants were the facets used in a fully crossed, random effects design. The results indicated that the score conversions effectively controlled for the variability in typically developing children in the same age range. The results showed that the raters accounted for small amounts of the estimated variance. In addition, the results indicated that the CSBS was sensitive to the different rates of development in the participants in the study. The results suggested that the CSBS successfully controls for the variability in the development of communicative behaviors without sacrificing sensitivity to different rates of growth and uses cluster scores to effectively differentiate strengths and weaknesses in individual children. The findings and their implications for futures studies are discussed.