Analysis of children's spoken narratives represents a potentially informative approach to language assessment within early childhood settings. Yet, narrative assessment is not readily amenable to at-scale use given the time needed to collect, transcribe, and analyze a child's narrative sample and the lack of consensus regarding what aspects of narrative expression ought to be examined (e.g., language form, language content). The purpose of this study was to describe a direct assessment of children's language abilities within a narrative context, the Narrative Assessment Protocol (NAP), which examines five aspects of language: sentence structure, phrase structure, modifiers, nouns, and verbs. In this study, we present findings regarding internal consistency, test-retest reliability, construct validity, and the concurrent and predictive validity of the NAP. NAP scores from 262 3-5-year-old children participating in preschool programs were assessed for these purposes. Findings indicated that the NAP exhibits reasonable psychometric properties across the areas addressed, to include significant concurrent and predictive relations with a norm-referenced measure of general language ability. Although more research is needed, preliminary findings indicate that the NAP provides professionals with a valid and informative assessment approach for examining children's language skills within a narrative context; such information may be useful for establishing and monitoring children's language growth within preschool programs or language interventions. (Contains 3 tables.)