HTTP/2 301 location: https://learnmoore.org/BDE/ content-length: 275 content-type: text/html; charset=iso-8859-1 date: Tue, 27 Jan 2026 23:03:39 GMT server: Apache HTTP/2 200 last-modified: Fri, 03 Mar 2017 05:04:39 GMT accept-ranges: bytes content-length: 13161 content-type: text/html date: Tue, 27 Jan 2026 23:03:39 GMT server: Apache

Research project:

Wide of the Mark: Evidence on the Underlying Causes of Overprecision in Judgment 

by Don Moore, Ashli Carter, and Heather Yang

Download the current version of the paper here

Study 1 (BDE16)

Qualtrics survey, viewed as a participant:  https://berkeley.qualtrics.com/SE/?SID=SV_29uAG3m5Mrt8KnH

The full code to the Qualtrics survey

The 20 photos of lottery outcomes we used

Data file

Study 2 (BDE13)

Qualtrics survey, viewed as a participant:  https://berkeley.qualtrics.com/SE/?SID=SV_d72dpbrC7U6mii1

The full code to the Qualtrics survey

The trivia quiz

Data file

Study 3 (BDE3)

Qualtrics survey, viewed as a participant:  https://berkeley.qualtrics.com/SE/?SID=SV_8zU7993dx7eCAEk

The full code to the Qualtrics survey

The slow Quincunx machine (log in with a number between 111 and 999)

The fast Quincunx machine (now appears not to work any more, but used to be like this one)

Data file

Study 4 (BDE6)

Qualtrics survey, viewed as a participant:  https://berkeley.qualtrics.com/SE/?SID=SV_abM7ZWJ2vShWIxn

The full code to the Qualtrics survey

Data file

Study 5 (BDE17)

Paper survey, viewed as a participant

Data file

.

.

.

Note that we ran 12 additional studies that did not make it in the paper:

BDE1 - Demonstrated the basic effect
Animated ball drop gif
Results write up
Data file
.
BDE2 - Varied the height of the Quincunx (the number of events)
The Qualtrics survey
Results write up
Data file
.
BDE4 - Frequentistic vs. probabilistic description of uncertainty
Take the survey as a participant
The Qualtrics survey code
Results write up (pathetically incomplete)
Data file
.
BDE5 - Varying the cover story: Quincunx vs. coin flips
Take the survey as a participant
The Qualtrics survey code
Results write up
Data file
.
BDE7 - Varying the cover story: Quincunx vs. jumping bean
Take the survey as a participant
The Qualtrics survey code
Picture of jumping bean
Results write up
Data file
.
BDE8 - Varying the cover story: Quincunx vs. jumping bean vs. dollar lottery
Take the survey as a participant
The Qualtrics survey code
Data file
.
BDE9 - Varying the cover story: Quincunx vs. jumping bean vs. dollar lottery
Take the survey as a participant
The Qualtrics survey code
Data file
.
BDE10 - SPIES vs. Confidence intervals with Quincunx vs. jumping bean vs. dollar lottery
Take the survey as a participant
The Qualtrics survey code
Data file
.
BDE11 - Jumping bean vs. dollar lottery with vs. dollar lottery without picture
Take the survey as a participant
The Qualtrics survey code
Data file
.
BDE12 - Multiple choice vs. fill-in-the-blank questions
Take the survey as a participant
The Qualtrics survey code
Data file
.
BDE14 - Multiple choice vs. fill-in-the-blank questions
Take the survey as a participant
The Qualtrics survey code
Results write up
Data file
..
BDE15 - Multiple choice vs. fill-in-the-blank questions
Take the survey as a participant
The Qualtrics survey code
Results write up
Data file
.

Why didn't these studies make it in the paper?  The paper is long enough as it is, and we believe the results of these 12 studies are less interesting and useful than those of the experiments that made it into the paper.  Generally, results from these 12 unpublished studies are compatible with the basic conclusions offered by the paper: reported probability distributions are too wide relative to the true probability distribution, but are so often centered on the wrong answer that they look overconfident when measured by hit rates.  Many of these 12 studies examined potential moderators of the effect, where the moderator appears to have failed.