Purpose: This study examined the empirical evidence for using the Structured Photographic Expressive Language Test: Third Edition (SPELT-3; Dawson, Stout, & Eyer, 2003) to diagnose language impairment in preschool children. The SPELT-3 is a revision of the SPELT-II (Werner & Kresheck, 1983), which has been proven in the past to have high levels of discriminant accuracy in identifying preschoolers with language impairment. Method: Forty-two 4- and 5-year-old children with a specific language impairment (SLI) and 43 children with typically developing (TD) language abilities were studied to determine the classification accuracy and other aspects of validity for the SPELT-3. Results: Results from both an exploratory and a confirmatory sample indicated 90% sensitivity and 100% specificity when a cutoff standard score of 95 was applied to the data. In addition, use of the SPELT-3 was supported by additional data on convergent and divergent aspects of validity. Implications: The data provide empirical support for the use of the SPELT-3 for the purpose of differentiating between children with normal language and those with impaired language.