Enhancing the internal validity of entrepreneurship experiments by assessing treatment effects at multiple levels across multiple trials

Treatment effects in entrepreneurship experiments could vary significantly due to environmental uncertainty and individual heterogeneity. Drawing on critical realism, experimental replication is useful for assessing the generalizability of the treatment effects or for confirming treatment effects across diverse entrepreneurship settings. Yet, experiments in entrepreneurship have traditionally relied on single experimental trials and have not considered potential variance in treatment effects across contexts. Accounting for variance in treatment effects is even more pertinent under higher levels of environmental uncertainty. To help assess the contingent effects of treatments across contexts, we propose a set of methodological and statistical approaches for entrepreneurship researchers. Using different types of experimental replications and proposed statistical methods, one may draw inferences about treatment effects at the trial-, the treatment-, and at the individual-level. We illustrate the proposed approach using experimental replications across three randomized, post-test, experiments.