Back to: Assess for Risk of Bias
Remember earlier in the course, when we defined randomized controlled trials (RCTs)? In RCTs, you have at least two groups of study participants: those who receive the treatment and those who receive an alternative intervention (for example, a placebo, another kind of treatment, nothing at all).
Random sequence generation is the method researchers used to randomly assign participants into groups.
Protects against: Selection bias
If the process for assigning participants to groups wasn’t random, participants in the treatment group may differ in important ways from those in the control group.
Examples of random sequence generation:
Many RCTs use a computer algorithm to generate random numbers which are then used to assign participants into groups. Older RCTs, especially, may have used other methods to “randomize” study participants, such as by the day of the week. However, this isn’t truly random method. Patients who seek treatment on a weekend might be different in important ways from patients who seek treatment during the week.
How to assess this domain:
Whatever method researchers used, ask yourself whether the sequence generation method was truly random. If the process was truly random, then on average the two groups should be balanced on all characteristics (except for the type of treatment they receive during the study)