||Previous research documents the tendency for listeners to assign negative attributes to children with communication disorders. However, variability within rater groups has been large across studies. Factors that influence ratings remain unaccounted for. In this study, 60 mothers rated 1-minute narratives produced by matched child speakers with specific language impairment (SLI), attention deficit hyperactivity disorder (ADHD), and typical development (TD). It was predicted that mothers who have a child with a disability (MWCD) would be more sensitive to the consequences of negative biases and would be less inclined to assign pejorative ratings to child speakers with communication disorders than mothers whose children only have typical development (MWCTD). Participant groups were matched for age, educational level, family size, and race/ethnicity. After hearing each sample, participants provided ratings in response to 20 questions about each speaker's narrative and a 16-item questionnaire providing details about their personal experiences with individuals with disabilities. Narrative questions included information about the actual narrative, child speaker attributes, and parent/family attributes. A 2 (MWCTD, MWCD) X 3 (SLI, ADHD, TD) mixed model ANOVA revealed statistically significant main effects within group (SLI, ADHD, TD). Analyses revealed that ratings for the TD speaker were significantly higher (i.e., "better") than the SLI and ADHD speaker ratings such that ADHD=SLI < TD or ADHD < SLI < TD. The pattern of ADHD < SLI < TD was found for questions related to narrative form and behavioral attributes. No significant effects were found for the between groups variable or for the group x speaker interaction. Overall age, education level, and exposure to disabilities predicted ratings but having a child with disability did not. Raters seemed to be very sensitive to features present in the ADHD speaker's narrative that marked him as having behavioral difficulties. Another surprising outcome was the lack of differences between ratings of narrative form assigned to the two clinical speakers, given that the narrative produced by the SLI speaker contained numerous grammatical errors. It appears that listeners do not differentiate grammatical problems from other problems with form (phonology, fluency). Clinical implications are discussed.