Recent years have seen great changes in how laboratory and field experiments are conducted. These changes are intended to increase the replicability of results by reducing false positives. Practices such as preregistering the experiment (often including experimental designs, hypotheses and analysis plans), correctly specifying significance thresholds, increasing the number of independent observations, and clearly distinguishing between the testing of hypotheses and exploratory analysis, have increased in popularity. Some have argued that these practices also impose new costs on researchers, do not eliminate all sources of publication bias and can slow the publication of new results. The goal of this panel discussion is to discuss important questions related to these and other practices that aim to produce more replicable research. What are the merits of these practices for different types of experimental research? How widely have they become employed? What should be the role of preregistration in the reviewing process for publication? Are there other practices that might further increase the replicability of experimental research?