How does *Qualtrics* randomly assign participants "under the hood"? | XM Community
Skip to main content

A reviewer of our manuscript wants to know precisely how participants were randomized in our surveys. We already say they were evenly randomized, so I think they want to know what Qualtrics’s algorithm does, under the hood, not merely how we programmed question blocks and in what proportion we randomized participants to conditions, etc. E.g., does it use the equivalent of a random number generator to assign each participant as they enter the survey, or alternate participants as they come (first person to condition A, second to condition B, third to A, fourth to B, etc.)? We haven’t immediately found any documentation about this. Does anyone know?

@Michelle Meyer Check out this comment by @TomG

“Qualtrics keeps an internal count for each randomized item and the counts are incremented when the items are displayed. Display is on a least-fill random basis (The items are sorted primarily by item count and secondarily by a random number with the top n being displayed).”


Leave a Reply