Critical differences can be thought of as critical regions for a priori and post hoc comparisons of pairs of means and of linear combinations of means. Critical differences can be transformed into confidence intervals. First, this entry discusses critical differences in the context of multiple comparison tests for means. Second, this entry addresses confusion surrounding applying critical differences for statistical significance and for the special case of consequential or practical significance.

Means Model

Multiple comparison tests arise from parametric and nonparametric tests of means, medians, and ranks corresponding to different groups. The parametric case for modeling means can be described as

where i = 1 … p (number of treatments) and j = 1 … ni (sample size of the ith treatment). The null hypothesis is that ...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles