Selection Bias is a form of bias in research studies where the results of the study are biased because people who participate (or for whom we are able to retrieve data) differ systematically from those who don’t participate. Therefore, if we attempt to generalize the findings of the study to the general population, they will not be representative.
An example might be analyzing the effect of an optional math intervention on student math proficiency. If participation in the program is voluntary, then it’s likely that the students who choose to participate in the intervention are fundamentally different than those who do not (e.g. they are likely more motivated). Therefore, the results we obtain from analyzing the data would likely be biased. It would be impossible to determine whether increases in proficiency are attributable to the intervention itself, or due to the pre-existing motivation differences.
Addressing Selection Bias
There are a few ways to potentially address selection bias. One is to conduct a Randomized Controlled Trial (RCT). Another is via Propensity Score Matching.