Procedure Comparison
This tool compares the predictability of results for both the asymptotic maximal procedure and variable-size permuted block randomization.
After clicking "Compare," two sample allocation sequences will be generated according the specified trial parameters, one using the asymptotic maximal procedure and one using permuted blocks. When the two simulations have been generated, a report comparing the two procedures will be displayed.
MTI can be from 2-10
Trial size can be from 500-1000
Number of arms can be from 2-4
Generate Comparison For:
Asymptotic Maximal Procedure:
- Trial Participants: 500
- Trial Arms: 2
- MTI: 3
Permuted Block Procedure:
- Trial Participants: 500
- Trial Arms: 2
- Block Sizes The shown block sizes provide an equivalent maximal tolerated imbalance to the asymptotic maximal procedure.: 6, 4
Results:
Correct guess percentage is the standard metric used to assess the predictability of a randomization procedure. Correct guess percentage is the percentage of trial participants allocated to the most probable (i.e. the least-represented) arm. Ideally, a randomized trial should have as low a correct guess percentage as possible.
The expected correct guess percentage for a totally random trial is one divided by the number of arms. Such a trial can be considered to have "Zero Predictability."
Metric | Asymptotic Maximal | Permuted Block |
---|---|---|
Correct Guesses | ||
Correct Guess Percentage | ||
Correct Guesses Above Zero Predictability | ||
Correct Guess Percentage Above Zero Predictability |