Researchers typically plan evaluations with samples that are large enough to detect meaningful intervention effects. Minimum detectable effect (MDE) calculations estimate the “the smallest true impact that an experiment has a good chance of detecting.”^[1] A smaller MDE means a study is better positioned to identify intervention effects, even if they are modest. A large MDE may cause smaller—but important—effects to go undetected.

Several factors influence the MDE:

• study design, such as whether individuals or groups are randomized
• target of inference, such as whether the study aims to estimate the effect on the study colleges or a broader population of colleges
• design parameters, such as the correlation between baseline data and the target outcome

THE-RCT’s MDE calculator [forthcoming] is a postsecondary-specific tool that is designed for individual randomized controlled trial evaluations (RCTs), which are frequently used to evaluate postsecondary interventions. The tool is especially useful for outcomes such as enrollment, credits earned, and degree completion. It includes the following features:

• built-in design parameters based on data from past randomized controlled trials of postsecondary interventions
• built-in benchmarks comparing a user’s MDE with effects that were found in past postsecondary studies (mostly at community colleges), which helps assess whether the MDE is realistically achievable

PowerUp! is a transparent, easy-to-use tool that is available in Excel and R and is suitable for a broad range of study designs. The creators also provide related resources.^[2]

The Causal Evaluation website contains links to MDE tools for various study designs and parameters (for example, moderation, mediators) and a detailed reference list.^[3] The Abdul Latif Jameel Poverty Action Lab (J-PAL) website offers many useful guides and tools for those interested in learning more about power analyses and MDE calculations.^[4]