05 - Motivated Beliefs about Ability (Zimmermann)

Question 1

motivation

Answer 1

• many people are overconfident about their skills
• key: question: people get feedback all the time – how do they manage to maintain the belief that they are great?
• hypothesis: the forget bad news

Question 2

experimental design

Answer 2

• Subjects completed an IQ test. Specifically, subjects solved a total of 10 Raven matrices, which are frequently used as a non-verbal test of intelligence.
• Subjects are randomly matched in groups of 10; elicit beliefs about whether performed in top half of group
• Noisy feedback about performance: randomly pick three other participants from own group, and truthfully inform subjects about whether he performed better than these three other subjects – this introduces variation infeedback that is random conditional on one’s own score: because the sample of comparisons is small (three), it will happen that people with the same score get different amounts of positive/negative news

Question 3

Treatments

Answer 3

Immediate: elicit posterior belief about whether performed in top half immediately after feedback
1 month: elicit posterior belief about whether performed in top half after one month.

Question 4

Results (Experiment 1)

Answer 4

Directly after the feedback, subjects update in the appropriate directions, both for positive and negative feedback. One month after the feedback, beliefs still reflect positive feedback, but belief adjustments after negative feedback are substantially diminished.

Question 5

Experimental Design 2:

Answer 5

Instead of measuring beliefs one month after the feedback, subjects’ recall accuracy is measured. Accurary with which subjects recalled the feedback they had received during the first session is elicited.

Question 6

Results (Experiment 2)

Answer 6

• The estimated negative coefficients of the feedback dummy reveal that subjects that obtained negative feedback recall that feedback with signficantly less accurary one month later, compared to subjects that received positive feedback.
• Subjects that obtained negative feedback state “I don’t recall” more frequently, compared to subjects that received positive feedback

Question 7

Results 3

Answer 7

• Both treatments reduce motivated reasoning
• Perhaps most surprisingly, subjects seem able to dig out the true feedback from memory when the incentives are really high
• Suggest that they don’t actually forget, but that it requires effort to unearth negative feedback.