Based on prior work, we predicted that traditional, 2-D counting books would be better than tactile counting books at promoting young children's numeracy. However, the first author suspected that the effect of tactility on learning would differ for Black versus non-Black children. To test this, we examined data from an existing project on preschoolers' learning from shared counting book reading. Participants included 325 preschoolers, ages 2 to 6, 41% of whom were Black. Findings suggest that race moderates the effect of tactility on numeracy. Non-Black children conformed to the original hypothesis that non-tactile counting books would be best for promoting children's numeracy, but Black children did not. This finding is important because much of the research on children's early numeracy is conducted with homogeneous, convenience samples, so theories and predictions are being built on incomplete data. Without studying diverse samples, the field risks making inaccurate conclusions about how children learn.