An experimental investigation of recruitment bias in eating pathology research

Previous, uncontrolled research has suggested a bias may exist in recruiting participants for eating disorder research. Recruitment biases may affect sample representativeness and generalizability of findings.

This experiment investigated whether revealing that a study's topic was related to eating disorders created a self-selection bias.

Young women at a university responded to advertisements containing contrasting information about the nature of a single study. We recruited one group by advertising the study under the title "Disordered Eating in Young Women" (n = 251) and another group using the title "Consumer Preferences" (n = 259).

Results indicated similar levels of eating pathology in both groups, so the different recruitment techniques did not engender self-selection. However, the consumer preferences group scored higher in self-reported social desirability.

The level of information conveyed in study advertising does not impact reporting of eating disturbances among nonclinical samples, although there is evidence social desirability might.