Self-Selection Bias

The distortion that occurs when people choose whether to participate in a group, study, or activity—making the participants systematically different from the non-participants in ways that affect the outcome. People who attend a self-help seminar are different from those who don't (more motivated, more open to change)—so the seminar's 'effectiveness' partially reflects who showed up, not what was taught. Customers who respond to premium pricing are different from those who don't. Employees who volunteer for stretch assignments are different from those who don't. Self-selection means the group you're studying was filtered by choice before you started observing.

When to use it

When program evaluations show strong results from volunteer participants; when early adopter success doesn't predict mainstream success; when 'what works for our best customers/employees' isn't working for the broader population; when any analysis involves people who chose to participate.

How it can help

For any program, product, or intervention: identify who self-selected IN and who self-selected OUT, and consider how that self-selection affects your conclusions. A weight-loss program that works for volunteers may fail for mandated participants because the volunteers were already motivated. A product that delights early adopters may disappoint mainstream users because early adopters are more forgiving. The correction: when possible, compare self-selected participants to randomly assigned participants. When randomization isn't possible, explicitly account for the self-selection filter in your conclusions.