Assessments of early cognitive and linguistic abilities typically involve picture stimuli. As these assessments spread worldwide, researchers make an implicit assumption: that children across contexts understand pictures in the same way, at the same developmental timepoint. What if this assumption does not hold for some or all kinds of pictures? In the present research, a preregistered sample of 128 3- to 7-year-olds from Kisumu, Kenya participated in a Swahili vocabulary assessment. Using a within-subjects design, each participant completed vocabulary trials in four formats (i.e., objects, photographs, cartoons, black-and-white line drawings). Preregistered analyses showed that children performed equally accurately across object, photograph, and cartoon formats, but less accurately in the line drawing format. However, exploratory analyses suggested that a subset of line drawings drove this difference. These findings suggest that caution is necessary in the use of picture stimuli and that assessments involving line drawings may sometimes underestimate children’s capacities.