The meaning, mechanism, and function of imitation in early infancy have been actively discussed since Meltzoff and Moore's (1977) report of facial and manual imitation by human neonates. Oostenbroek et al. (2016) claim to challenge the existence of early imitation and to counter all interpretations so far offered. Such claims, if true, would have implications for theories of social‐cognitive development. Here we identify 11 flaws in Oostenbroek et al.'s experimental design that biased the results toward null effects. We requested and obtained the authors’ raw data. Contrary to the authors’ conclusions, new analyses reveal significant tongue‐protrusion imitation at all four ages tested (1, 3, 6, and 9 weeks old). We explain how the authors missed this pattern and offer five recommendations for designing future experiments. Infant imitation raises fundamental issues about action representation, social learning, and brain–behavior relations. The debate about the origins and development of imitation reflects its importance to theories of developmental science. Re‐analyses of Oostenbroek et al.’s (2016) data show significant neonatal imitation. Infants produced significantly more tongue protrusions (TP) in response to the TP demonstration than to controls at all four ages tested, despite a weak design biased toward null effects.