Abstract

Although a rapidly growing body of scientific literature is leading to more understanding of autistic traits in women and girls, their diagnosis continues to occur at later ages. Diagnosis received at later ages limits access to early intervention and related positive outcomes. The purpose of this study was to add data to the weight of evidence approach in differential diagnosis of autism in females and other underserved groups. Specifically, I set out to identify patterns between verbal comprehension tasks measuring vocabulary knowledge (crystallized knowledge) and analogous verbal comprehension (similarities) tasks drawing on additional cognitive abilities. Verbal analogy tasks measure knowledge of vocabulary in concrete terms with the added ability to recognize abstract aspects of concepts and possible relationships between given terms (crystalized knowledge and fluid reasoning). The study utilized commonly used cognitive assessment measures that include both vocabulary and similarities/analogies to represent verbal comprehension. This is done by examining the difference, gap, or scatter between two subdomains of verbal abilities (vocabulary and similarities) and then comparing the difference score by variables such as gender, intelligence quotient (IQ), and age. Deidentified participant data was drawn from 77 studies in the National Database for Autism (NDA: National Institutes of Health, NIH USA) that met inclusionary criteria following a rigorous screening process. Data were categorized based on data patterns between scaled scores for vocabulary and similarities or analogy subtests, which were analyzed to assess within-group differences across gender. Differences (gaps) between scores were also analyzed by IQ level and by age to determine the roles of other common variables in any differences that were found. Statistical analysis involved a single linear regression for each variable to determine predictability, as well as an interaction effect analysis. The analysis resulted in three findings. First, females within this sample population had, on average, higher gaps between vocabulary and similarities than males. Second, individuals with IQs above the mean in the sample demonstrated wider gaps in vocabulary and similarities scores than those with IQs below the mean. These findings were significant in the presence of each other, however there was no significant interaction effect between the two variables. Finally, age was not a significant predictor of gap scores, even in the presence of IQ, and no interaction effect was found. This demonstrates that that the discovered patterns of more differences between verbal scores relating to both IQ and gender were visible across participant age range. Limitations include potential variance related to time, with studies utilizing measures with older and newer publication dates and revisions as well as differences across measures. The study was also limited in its ability to gather demographic data as well as data on other potential variables that may impact the participants' testing outcomes. Unknown bias toward females with more obvious autism and/or co-occurring conditions within the NDA are potential confounds, as the number of females was far fewer than males. Gaps identified can contribute to differential decision-making processes and may inform supports but are not intended as diagnostic indicators.

Degree

EdS

College and Department

David O. McKay School of Education; Counseling Psychology and Special Education

Rights

https://lib.byu.edu/about/copyright/