Lillard et al. (2013) concluded that pretend play is not causally related to child outcomes and charged that the field is subject to a "play ethos", whereby research is tainted by a bias to find positive effects of play on child development. In this commentary, we embrace their call for a more solidly scientific approach to questions in this important area of study while offering 2 critiques of their analysis. First, we urge researchers to take a more holistic approach to the body of evidence on play and learning, rather than relying on piecemeal criticisms of individual studies, since positive effects of play on learning emerge despite the use of a variety of methods, contents, and experimental conditions. Second, we consider how best to study this topic in the future and propose moving away from traditional empirical approaches to more complicated statistical models and methods that will allow us to embrace the full variety and complexity of playful learning.