A replication crisis plagues current psychological research. Recent findings suggest that most psychological findings are failing to replicate with increasing distrust in the literature. This applies to the developmental literature as well, where the difficulties associated with participant recruitment and the natural variability in participant behavior during early development causes additional problems in statistical analyses and interpretation. We proposed that Bayesian sequential testing (BST) may provide an alternative to standard testing paradigms which typically test more or fewer children than absolutely required by the analyses and even then end with inconclusive results in cases where a non-significant result is obtained. We tested three different tasks examining infant word learning, infant word recognition and infant reasoning and found that BST robustly led to the same conclusion as the original study with far fewer participants in one case, clearly showed a failure to replicate the original study in a second case, and replicated the original study with the same number of participants in a third case. Taken together, our findings highlight the real benefits of BST in developmental research, highlighting the possibility of robust replicable findings using sequential testing.