They should have known: Hindsight and outcome biases in child abduction cases. Issue 7 (8th May 2020)

• Record Type: Journal Article
• Authors: McDermott, Christine M.; Miller, Monica K.; DeVault, Alicia
• Is Part Of: Journal of applied social psychology. Volume 50:Issue 7(2020)
• Page Start: 430
• Page End: 438
• Publication Date: 2020-05-08
• Subjects: AMBER Alert system -- child abduction -- cognitive errors -- crime control theater (CCT) -- hindsight bias -- outcome bias; Social psychology -- Periodicals; Behavior -- Periodicals; Psychology, Social -- Periodicals; 302
• DOI: 10.1111/jasp.12671
• Languages: English
• ISSNs: 0021-9029

Abstract

Abstract: Two cognitive biases might partially account for public support of the ineffective AMBER Alert system. Hindsight bias is a cognitive error in which people with outcome knowledge overestimate the likelihood that this particular outcome would occur; outcome bias is an error made in evaluating the quality of a decision once the outcome is known. Two experiments assessed whether hindsight and outcome bias occur in child abduction scenarios. Study 1 was a pre/posttest experiment that examined whether hindsight bias occurs in situations in which the identity of the abductor (stranger or parent) is manipulated between groups, and all participants are told the child was killed. Study 2, a between‐subjects experiment, examined whether hindsight and outcome biases occur in situations in which no AMBER Alert was issued (because the situation did not meet the legal requirements to issue an Alert), and manipulated the identity of the abductor and the outcome (child safely returned, killed, or not outcome provided). Hindsight and outcome biases occurred in both studies, given the correct set of circumstances. Abductor identity also impacted outcome estimates. Results from the two studies indicate that hindsight and outcome bias occur, but this is dependent on the outcome (child killed, child returned safely, no outcome provided) and the identity of the abductor (stranger, dangerous parent, non‐dangerous parent). Limitations and future directions are discussed.