Cluster Randomization Trials

5. Pair-Matching

Exercise 2

The goal of this exercise is to assist users in understanding and applying the proper implementation of pair-matching.

Question: What factors should a researcher consider when deciding whether to match clusters prior to randomization?

The user selects the appropriate options and the ‘How did I do’ button.

The options are:

  • The likely magnitude of the matching correlation.
  • Whether the primary trial outcome is binary or continous.
  • The interest of the investigators in evaluating the effect of individual level risk factors on outcome.
  • The total number of clusters available to be randomized.
  • Whether the intervention is to be offered at the cluster or individual level.
  • The relative sizes of the clusters to be randomized.

Answers:

  • The likely magnitude of the matching correlation. Correct!
  • Whether the primary trial outcome is binary or continous. Poor Choice
  • The interest of the investigators in evaluating the effect of individual level risk factors on outcome. Correct!
  • The total number of clusters available to be randomized. Correct!
  • Whether the intervention is to be offered at the cluster or individual level. Poor Choice.
  • The relative sizes of the clusters to be randomized. Possible.