Accurate measurement of autistic children's social communication is critical for assessing skills, setting intervention goals, evaluating change over time, determining service eligibility, and determining classroom placement. There are various types of assessments, some of which use specific tasks to elicit social communication. Structured tasks may frustrate children, inadvertently elicit irritability, and have a cascading effect on their ability to communicate. To date, no studies have evaluated how differing types of social communication assessments may exacerbate children's irritability and impact assessment scores. We examined the extent to which (a) social communication assessment type (structured vs. naturalistic) impacts autistic children's irritability and (b) child irritability is associated with social communication scores. Autistic toddlers (n = 114, Mage = 33.09 months, SD = 6.15) completed the Communication and Symbolic Behavior Scales (CSBS; structured) and a 10-min play-based mother-child interaction (MCX; naturalistic). Child irritability was scored on both assessments using a global rating scale of 0-15. Child irritability during the CSBS was significantly higher than during the MCX (V = 4892, p < .001, r = .68). Higher irritability was associated with lower CSBS social communication scores (B = -0.05, p = .03), but not MCX scores (B = 0.04, p = .13; Theil's F = 6.92, p = .009). Our findings suggest that the CSBS may pose unique challenges for autistic children, as it led to higher rates of irritability and negatively affected children's social communication scores. Evaluating the association between assessment type and irritability supports the complete characterization of autistic children's experience during assessments and clinicians in obtaining a more representative measure of social communication.