Stratification factors are the criteria that ensure treatment arms maintain an balanced variety of subjects across all treatment arms. For example, you may wish to study the role biological sex plays on a particular treatment. In this case, biological sex would be set up as the stratification factor, allowing this criteria to be selected during randomization. This would ensure that a balanced amount of men and women are present amongst the treatment arms, allowing observations to be made between these subgroups. A Case/Control study could benifit from properly chosen stratification factors.

The study may also be interested in the effects of the treatment based on weight as well. In this case a second stratification factor, for example Body Mass Index, may be added.

Both the Permuted-block and the Minimization algorithm require discrete choices to be present for all stratification factors. In the case of biological sex and body mass index, these groups could look as follow:

Biological Sex
  • Male
  • Female
Body Mass Index
  • Less than 20
  • 20 to 30
  • Greater than 30

To ensure an even distribution across sites, a "Stratify By Site", stratification option is also available.

If you are using Permuted-block algorithm, increasing the number of stratification factors increases the number of subjects that need to be randomized to ensure even distributions across treatment arms. A better option in this case may be to use the Minimization algorithm instead, which handles distributions by stratification factors dynamically instead of relying on precomputed lists.