Difference between revisions of "Minimum Detectable Effect"
Maria jones (talk | contribs) |
Maria jones (talk | contribs) |
||
| Line 1: | Line 1: | ||
== Read First == | == Read First == | ||
Minimum Detectable Effect Size (MDES) is a critical part of power calculations. | |||
== Guidelines == | == Guidelines == | ||
We do not know in advance the effect of our policy. We want to design a precise way of measuring it. | We do not know in advance the effect of our policy. We want to design a precise way of measuring it. But precision is not cheap: need cost-benefit analysis to decide. We need to identify the smallest program effect size that it would be useful to detect, i.e. the smallest effect for which we would be able to say with statistical confidence that the program effect is statistically different from zero. | ||
But precision is not cheap: need cost-benefit analysis to decide. | |||
We need to identify the smallest program effect size that it would be useful to detect | |||
===Subsection 1=== | ===Subsection 1=== | ||
| Line 24: | Line 15: | ||
== Additional Resources == | == Additional Resources == | ||
Please add here related articles, with a brief description and link. | |||
[[Category: Sampling & Power Calculations]] | [[Category: Sampling & Power Calculations]] | ||
Revision as of 22:29, 16 October 2017
Read First
Minimum Detectable Effect Size (MDES) is a critical part of power calculations.
Guidelines
We do not know in advance the effect of our policy. We want to design a precise way of measuring it. But precision is not cheap: need cost-benefit analysis to decide. We need to identify the smallest program effect size that it would be useful to detect, i.e. the smallest effect for which we would be able to say with statistical confidence that the program effect is statistically different from zero.
Subsection 1
Subsection 2
Subsection 3
Back to Parent
This article is part of the topic Sampling & Power Calculations
Additional Resources
Please add here related articles, with a brief description and link.