diff --git a/multiple-comparisons/multiple-comparisons.Rmd b/multiple-comparisons/multiple-comparisons.Rmd index abee39d..bfb95c6 100644 --- a/multiple-comparisons/multiple-comparisons.Rmd +++ b/multiple-comparisons/multiple-comparisons.Rmd @@ -459,7 +459,7 @@ knitr::kable(table, digits = 3) 10: Use Design Based approaches == -There are also two design based approaches for thinking about this problem. One is to use pre-analysis plans to describe in advance which comparisons will be made; the number of tests implicated in a Bonferroni or Bonferroni-style correction is specified before any data analysis is conducted. A pre-analysis plan, among its many benefits, helps to clarify the multiple comparisons problem. +There are also two design based approaches for thinking about this problem. One is to use pre-analysis plans to describe in advance which comparisons will be made; the number of tests implicated in a Bonferroni or Bonferroni-style correction is specified before any data analysis is conducted. A pre-analysis plan, among its many benefits, helps to clarify the multiple comparisons problem. (See our guide [10 Things to Know About Pre-Analysis Plans](https://egap.org/resource/10-things-to-know-about-pre-analysis-plans/) for general information on pre-analysis plans.) A good example is this pre-analysis plan by [Gwyneth McClendon and Rachel Beatty Riedl](https://egap.org/registration/651). The authors specify 24 individual tests they plan to conduct and state ahead of time that they will employ both the Bonferroni and Benjamini-Hochberg corrections. The authors do not state beforehand how they will handle a situation in which the corrections disagree; presumably it will be a matter of judgment for both the authors and the readers of their study. diff --git a/multiple-comparisons/multiple-comparisons.html b/multiple-comparisons/multiple-comparisons.html index a6f9fa4..7944368 100644 --- a/multiple-comparisons/multiple-comparisons.html +++ b/multiple-comparisons/multiple-comparisons.html @@ -266,6 +266,7 @@ +