The purpose of this study was to examine the reliability of measures from conversational language samples in the school-aged population. Two 10-minute conversational language samples were collected from 20 at-risk kindergarten children to determine their reliability and feasibility as an assessment and progress monitoring tool. All samples were collected and transcribed by one examiner using an outlined elicitation protocol in which the children were asked open ended questions about school and home activities in various time segments. Test-retest reliability was determined across eleven language measures for each child. Significant reliability correlation coefficients were observed for each language sample measure. The amount of variance due to sample length and topic were analyzed utilizing generalizability theory, which observes various facets accounting for measurement error. Variance was calculated for each language measure. Conversational language samples were found to have strong test-retest reliability across all language measures. The greatest amount of variance in measures was attributed to the child in most language measures with the interaction between the child, session, and length/topic accounting for the second largest amount of variance in most measures. Length and topic accounted for negligible amounts of variance in language measures. These findings demonstrated that shorter conversational language samples are reliable across time and that the length and topic of the sample do not have a substantial impact on the reliability of language measures.