A question we sometimes get as People Scientist is how we arrived at the suggested minimum reporting size. The explanation below helps to provide further clarity on this topic.
The US Census Bureau recommends a minimum degrees of freedom of two to eliminate the possibility that if you know one response/value, you can deduce the others. In other words, if you have a sample of two people and an average score of three, and it is revealed that Person 1 gave a score of five, I now know Person 2 gave a score of one (because five and one average to three). If you bump the sample size to three people, knowing Person 1's scores doesn't automatically tell me Person 2 and 3's scores.
Viva Glint uses a minimum of five because we provide multiple flavors of score that could be used in tandem to deduce scores. We provide a mean (Viva Glint Score), % Favorable, and % Unfavorable for every subgroup. So we need to create more room for 'uncertainty' about people's scores. The three unique data points + two recommend degrees of freedom = five people minimum sample.
What minimum reporting threshold does your organization set for survey results? Do you think the threshold is adequate? Should it be lower or higher? Share your thoughts below!
