Minimum Detectable Effect

Revision as of 18:37, 9 August 2023 by Zkevala (talk | contribs)
Jump to: navigation, search

The minimum detectable effect is the effect size set by the researcher that an impact evaluation is designed to estimate for a given level of significance. The minimum detectable effect is a critical input for power calculations and is closely related to power, sample size, and survey and project budgets. The page provides an overview of minimum detectable effect and provides points to consider when choosing it.

Read First

  • The minimum detectable effect is an important input in power calculations.
  • The minimum detectable effect can be considered the level of impact below which a program will be considered unsuccessful.
  • When setting the minimum detectable effect, look at studies on similar programs to understand the potential impact size of the program.
  • Set the minimum detectable effect conservatively given sufficient resources.

Overview

The minimum detectable effect answers the question, “Given that I only have budget to sample x households, what is the minimum effect size that I will be able to distinguish from a null effect?” Any level of impact below the minimum detectable effect will not be detected. As such, the minimum detectable effect can also be considered the level of impact below which a program will be considered unsuccessful.

Considerations

To set the minimum detectable effect, consider the change in outcomes that would justify the investment in an intervention. What is a policy relevant-impact? What are the policy implications? Consider results from studies on similar programs: these may give insight into the potential impact size.

While it is ideal to set the minimum detectable effect conservatively at a small figure, consider that, all else held constant, detecting a smaller minimum detectable effect requires a larger sample size. Thus, factors like the survey budget may also influence the researcher’s choice for the minimum detectable effect.

Back to Parent

This article is part of the topic Sampling & Power Calculations

Additional Resources