This page describes the output files generated from the Trial Randomization Tool.
Please note that after your trial's output is generated, you will have 16 days to download it. After 16 days, the output generated for your trial will be deleted from this website and will no longer be accessible.
Your randomized trial download will contain the following files:
Instructions File (.txt):
Your download will include an instructions file which contains the same information that you see on this page.
Parameters File (.txt):
A parameters file is included which shows which parameters you used to create the trial (such as trial size, arm counts, etc).
Assignments File (.xlsx):
An "assignments" file will be created in Excel (.xlsx) format. If your file was stratified, there will be one worksheet for each stratum. Each worksheet will have the following fields for each subject:
The label of the stratification group for the spreadsheet. If stratification is not used, each value for this field will be "All."
An integer id for each successive trial participant. Each participant can be uniquely identified by his or her Group ID and Subject Number.
The label of the trial arm the trial participant has been assigned to.
Assigned Arm Index:
The index of the trial arm the trial participant has been assigned to.
Here is a sample output of what the first few rows of the assignments file might look like:
Group Subject Assignment Assigned Arm Index All 1 Experimental 2 All 2 Control 1 All 3 Experimental 2 All 4 Experimental 2 All 5 Control 1
Assignments File (.csv):
A CSV version of the Excel assignments file will also be generated. The data in this file is identical to the Excel version. If your file was stratified, there will be a separate CSV file for each stratum. Each CSV file will correspond to one worksheet in the Excel version.
Metadata File (.xlsx):
This file contains the same columns and data as the assignments file, plus several additional columns. This extra information can be used to test for selection bias, but is not necessary to make assignments. The additional columns are:
Arm X:Arm Y Prior Imbalance
The imbalance of Arm X compared to Arm Y before the assignment is made (# Arm Y - # Arm X). A positive number signifies Arm Y has been assigned more applicants than Arm X in 1:1 allocations or more than the designated allocation ratio if the ratio is non-uniform.
Arm X Probability
The probability, prior to generating the assignment, that the participant would be assigned to Arm X. This is used to test for selection bias.
Arm X:Arm Y Posterior Imbalance
The imbalance of Arm X compared to Arm Y after the assignment is made (# Arm Y - # Arm X).
Here is a sample output of what the first few rows of the metadata file might look like:
Group Subject Assignment Arm 1:Arm 2 Prior Imbalance Arm 1 Probability Arm 2 Probability Assigned Arm Index Arm 1:Arm 2 Posterior Imbalance All 1 Experimental 0 0.5 0.5 2 1 All 2 Control 1 0.585786438 0.414213562 1 0 All 3 Experimental 0 0.5 0.5 2 1 All 4 Experimental 1 0.585786438 0.414213562 2 2 All 5 Control 2 0.707106781 0.292893219 1 1
- Arm X:Arm Y Prior Imbalance
Metadata File (.csv):
A CSV version of the medatata file is also generated. Just like the CSV assignments file, stratified trials will have one CSV per stratum.
In order to prevent selection bias, principal researchers must remain blinded to the allocation sequence. Therefore, it is essential that the researchers DO NOT have access to the output of this program themselves. Access to the output sequence must be restricted to a third party who would reveal each successive trial participant's allocation only after that participant has been accepted into the trial.