Background: Numerous studies show clear evidence that children display typicality effects during early word learning. However, little is known of the typicality of stimuli used by standardized language tests to assess children's language development. Aims: To examine the typicality of stimuli used by the Reynell Developmental Language Scales--III (RDLS-III), a standardized language assessment that measures children's language abilities. Methods & Procedures: Two sets of items were compared: RDLS-III items and a set of Matching items that had previously been judged to be highly typical exemplars. Experiment 1 asked adults to rate images of RDLS-III items and the set of Matching items for typicality; while Experiment 2 asked adults to choose the "best exemplar" from these sets of items. Experiment 3 used Intermodal Preferential Looking (IPL) to examine 4-year-olds' looking behaviour towards both the RDLS-III items and Matching stimuli. In supplement of Experiment 3, Experiment 4 asked children to indicate the better exemplar using a forced-choice pointing method. Outcomes & Results: Both adults and children preferred the Matching, more typical items. Conclusions & Implications: This is the first study to demonstrate that stimuli used for assessing early language development in the RDLS-III are not judged to be typical by children or adults. It highlights the differences in stimuli and discusses the potential consequences of stimulus choice in the assessment early language. (Contains 4 figures and 2 tables.)