On the evaluation of one-sided evidence
Lyle Brenner, Derek Koehler and Amos Tversky studied predictions and confidence based on one-sided evidence in mock legal disputes. In the first experiment, students at San Jose State University read adapted civil, criminal and child-custody cases and saw only the plaintiff's side, only the defendant's side or both, then predicted how many of 20 jurors favoured the plaintiff and gave a 90 per cent range. Mean guesses were 12.3 for the plaintiff-only group, 9.5 for the defendant-only group and 11.3 for the group that saw both. The one-sided groups gave narrower ranges and were less accurate than the group that saw both sides. A later experiment in which participants first rated the strength of the unseen side's case cut the average gap between the two one-sided groups from 7.1 jurors to 2.7. The authors conclude that people do not compensate sufficiently for missing information.
Lyle Brenner, Derek Koehler and Amos Tversky studied predictions and confidence based on one-sided evidence in mock legal disputes. In the first experiment, students at San Jose State University read adapted civil, criminal and child-custody cases and saw only the plaintiff's side, only the defendant's side or both, then predicted how many of 20 jurors favoured the plaintiff and gave a 90 per cent range. Mean guesses were 12.3 for the plaintiff-only group, 9.5 for the defendant-only group and 11.3 for the group that saw both. The one-sided groups gave narrower ranges and were less accurate than the group that saw both sides. A later experiment in which participants first rated the strength of the unseen side's case cut the average gap between the two one-sided groups from 7.1 jurors to 2.7. The authors conclude that people do not compensate sufficiently for missing information.
Written from the full text of the paper read 2026-10-27 at the first author's university page.
What our sources record
Recorded when this source was checked, figures included, with page references and quotation marks removed.
Full text at bear.warrington.ufl.edu/brenner/papers/brenner-jbdm96.pdf