A Stability Bias Effect Among Lie-Tellers: Testing the "Miscalibration" and "Strategic" Hypotheses

General Audience Summary It is well-established that truth-tellers report fewer details when interviewed after a delay (compared to when interviewed immediately) due to forgetting. In contrast, lie-tellers tend to report similar amounts of detail when interviewed immediately or after a delay. Lie-tellers' failure to correctly simulate forgetting is what's known as a stability bias effect. The current experiment examines the processes underlying this stability bias effect amongst deceivers. We found that participants were sensitive to delay and predicted forgetting when estimating how much detail could be remembered. In contrast, participants estimated that similar amounts of detail would be needed to convince others that a statement was genuine. This was the case when the statement was given immediately following an event, and after a three-week delay. Results suggest that the stability bias effect amongst lie-tellers is underpinned by their motivation to appear convincing to others. Unlike truth-tellers' statements that show forgetting, lie-tellers' statements appear less sensitive to delay. For lie-tellers, this failure to correctly simulate forgetting has been referred to as a stability bias. This experiment tests two explanations for this stability bias: the "miscalibration" hypothesis and the "strategic" hypothesis. Using a 2 (Task Type: Recall Estimate vs. Strategic Estimate) x 2 (Delay: Immediate vs. 3-week delay) design, participants (n = 142) either estimated how much detail a truth-teller might remember from an intelligence briefing (testing the miscalibration hypothesis), or how much detail was necessary to make a fabricated statement about the same intelligence briefing appear convincing to others (testing the strategic hypothesis). Before making these estimates, participants were informed that the briefing occurred immediately beforehand, or 3 weeks beforehand. Recall estimates correctly predicted forgetting would occur after a 3-week delay. Strategic estimates did not vary as a function of statement-time. No differences in subjective beliefs emerged.