How selection bias skews a result
Launch library · evergreen read
Selection bias occurs when the people or cases included in a result are not representative of the wider group being discussed, often because of how participants were recruited or because only certain outcomes were visible enough to be counted in the first place at all, which quietly biases the whole exercise from the very start.
A common example is relying on people who volunteer to respond, since those willing to take part often differ systematically from those who do not, typically holding stronger or more unusual views than the wider population that the result is meant to represent to a reader.
Spotting selection bias means asking who was left out of a result and why, since the pattern of exclusion often matters just as much as the pattern found among those who were actually included in the final count that gets reported publicly to an audience, since a skewed sample can look perfectly ordinary on the surface.