Advice taking in decision making: Egocentric discounting and reputation formation

The question

When a decision-maker’s own estimate conflicts with an advisor’s, how much weight do they actually give the advice — and does that change with the advisor’s track record?

The method

In four studies, Hebrew University undergraduates gave an initial estimate for general-knowledge questions (historical dates), then saw another person’s estimate and gave a final, possibly revised, answer, with a real cash bonus for accuracy [1].

The findings

People discount advice relative to their own opinion — on average, the final estimate stayed about 71% of the way toward the respondent’s own initial guess and only 29% of the way toward the advisor’s [1], even though the advisors were, on average, at least as accurate as the respondents themselves [1]. An advisor’s reputation also forms fast and asymmetrically: when an advisor’s advice quality dropped from good to average, respondents’ discounting jumped sharply (from a weight of 0.41 to 0.72 on their own opinion) [1]; when a poor advisor’s advice improved to average, respondents barely gave them credit (0.81 down to only 0.74) [1] — a good reputation is easier to lose than to earn. And it doesn’t take long to form: about three trials of experience were enough to guide whether people would even pay to buy an advisor’s next estimate [1].

The limits

The tasks were general-knowledge estimation questions with a single piece of advice per trial, run in an anonymous lab setting with real money at stake [1]. The authors note that real advisory relationships — a journal editor and reviewer, a patient and physician — may show reputation forming just as fast [1], but that’s inference from the lab pattern, not something the studies test directly.

• When an advisor pushes back on your read of a situation and you feel confident in your own judgment: check that confidence against the numbers here. People overweighted their own opinion (a weight of 0.71) even when the advisor’s estimates were, on average, no less accurate than their own [1] — noticing you disagree isn’t evidence you’re right.
• When someone gives you one bad piece of advice early in a relationship: watch the urge to write them off completely. Reputation formed within about three encounters here [1], and a single miss cost an advisor’s credibility far more than a hit built it [1] — that’s a bias worth naming before you act on it, not a signal to trust.
• When you’re the one being second-guessed after a wrong call: expect to be discounted more than the situation strictly warrants. The asymmetry runs against advisors, not just for them — a single decline in quality overturned nearly a third of an advisor’s earned trust [1].