Think twice and then: combining or choosing in dialectical bootstrapping?

Area of Science:

• Cognitive Psychology
• Decision Making
• Social Cognition

Background:

• The "wisdom of crowds" effect demonstrates that aggregating independent judgments can yield accurate collective estimates.
• Individuals can simulate this effect internally, creating a "crowd within" by combining their own non-redundant estimates.
• Prior research introduced "dialectical bootstrapping" to enhance accuracy by encouraging divergent thinking between estimates.

Purpose of the Study:

• To investigate the conditions under which individuals spontaneously utilize the "crowd within" strategy for improved accuracy.
• To understand the factors influencing the decision to combine multiple self-generated estimates versus selecting a single best estimate.
• To explore whether alternative strategies can outperform the "crowd within" approach.

Main Methods:

• Participants generated multiple estimates for a given task, with some instructed to actively contradict their initial judgments.
• The study analyzed the relationship between instructions, the degree of disagreement between estimates, and the choice to combine estimates.
• Performance was compared between participants who combined estimates and those who employed other strategies.

Main Results:

• Participants were more inclined to combine their estimates when explicitly instructed to think differently or contradict themselves.
• The likelihood of combining estimates increased with the magnitude of disagreement between the first and second estimates.
• Participants who did not combine their estimates failed to achieve accuracy superior to their internal "crowd within" performance.

Conclusions:

• Spontaneous adoption of the "crowd within" strategy is influenced by cognitive prompting and perceived error risk.
• Individuals appear to use estimate combination as a heuristic to hedge against potential large errors.
• The "crowd within" strategy, particularly when employing dialectical bootstrapping, offers a robust method for enhancing individual judgment accuracy.