Abstract: Clinicians have very few tools to objectively identify whether a child may be accurately reporting sexual abuse. A few clinicians believe that accurate testimony can be defined by a continuous and consistent retelling of the abuse description (Gardner, 1987). However, many clinicians have found that this fluency of narrative may be very rare, even in accurate reporters (Foa & Riggs, 1993; Amir, Stafford, Freshman, & Foa, 1998). Due to the nature of the event it has been shown that stress and trauma can lead to dysfluency in the abuse narrative and inconsistency during the recounting of the events for adults (Foa & Riggs, 1993; Van Der Kolk & Fisler, 1995; Jacobs & Nadel, 1998). This study examines speech dysfluency in children in hopes to provide an alternative adjunctive measure of possible trauma history that may be less influenced by demand than self-report. A factor analysis was conducted on the dysfluency variables. It was found that the variables of revision, single syllable repetition, phrase repetition, interjection, and abandoned utterance loaded together. A 2(phase)×2(severity) mixed analysis of variance yielded an interaction between severity of abuse and phase of interview. Severely abused children showed more severe deterioration of language. A regression analysis was conducted on severity, emotion, sensation words, and rated anxiety in the prediction of dysfluency. Severity did increment over the child's age and rapport phase dysfluency, while emotion and sensation words and anxiety levels did not. In exploratory analysis, disorganized thoughts added significant variance to the prediction of dysfluency.