Skip to main content
Fig. 1 | Trials

Fig. 1

From: Simple compared to covariate-constrained randomization methods in balancing baseline characteristics: a case study of randomly allocating 72 hemodialysis centers in a cluster trial

Fig. 1

Percentage of time each pair of centers were randomly allocated to the same group (i.e., Center 1 with Center 2, Center 1 with Center 3, Center 1 with Center 4, ..., Center 71 with Center 72). There were a total of 2556 unique center pairs. A Centers randomly allocated without constraints (i.e., simple randomization) would appear in the same arm approximately 50% of the time. B Constraining on a subset of 11 prognostic baseline characteristics. C Constraining on ten principal components from a Principal Component Analysis. The horizontal dashed lines show center pairs (if any) allocated to the same arm 25% or 75% of the time [41]

Back to article page