# 190 | ResearchBox

ResearchBox # 190 - 'Generation and choice rely on distinct value forms'

Bingo Table
  Show file names
  Show file IDs
  Show timestamps










Studies 1-2


Studies 4-6




Studies 4-7, 9




Study 1






Study 2






Study 3







Study 4











Study 5

  Study 5 - AsPredicted #14220.pdf











Study 6











Study 7

  Study 7 - AsPredicted #45132.pdf











(23 Mb)


Study 8







Study 9












Previewing files
Files can be previewed by clicking on descriptions.
Codebooks can be previewed by clicking on


Tell us if something is wrong with this Box


Adam Morris; Jonathan Phillips; Karen Huang; Fiery Cushman, 'Generating options and choosing between them rely on distinct forms of value representation', Psychological Science

All content posted to ResearchBox is under a CC By 4.0 License (all use is allowed as long as authorship of the content is attributed). When using content from ResearchBox please cite the original work, and provide a link to the URL for this box (https://researchbox.org/190).

March 04, 2021   (files may not be changed, deleted, or added)

Adam Morris (adammorris@g.harvard.edu)
Jonathan Phillips (Jonathan.S.Phillips@dartmouth.edu)
Karen Huang (karen.huang@georgetown.edu)
Fiery Cushman (cushman@fas.harvard.edu)

Humans have a remarkable capacity for flexible decision making, deliberating among actions by modeling their likely outcomes. This capacity allows us to adapt to the specific features of diverse circumstances. In real-world decision making, however, people face an important challenge: There are often an enormous number of possibilities to choose among, far too many for exhaustive consideration. There is a crucial, understudied "pre-choice" step in which, among myriad possibilities, a few good candidates come quickly to mind. How do people accomplish this? We show across nine experiments (N = 3972, U.S. residents) that people use computationally-frugal "cached" value estimates to propose a few candidate actions based on their success in past contexts (even when irrelevant for the current context). Deliberative planning is then deployed just within this set, computing more accurate values based on context-specific criteria. This hybrid architecture illuminates how typically-valuable thoughts come quickly to mind during decision making.


The authors have written the following message for visitors to this box.
Please note that these messages can be modified or deleted at any point (even after a box is made permanent)

Dear Reader,

You must use R version 3 to run the analysis code (I ran it all with R 3.6.3); it will break with R version 4. (Sorry for this annoyance.) An easy way to switch between versions of R can be found here: http://groundhogr.com/many/. Also, the code uses groundhog to ensure reproducibility, but it's kind of finicky; if it fails, just restart your R session a few times and try again.

Studies 1-3 and 8 were made in Qualtrics; the PDF and QSF exports from Qualtrics are included here. (For the Qualtrics data, some of the variable names are sloppy.. E.g. "cs_13" is actually the 8th slot for entering a considered item. Sorry. It should all be clear in the analysis script.) Studies 4-7 and 9 were done in Javascript, using jsPsych. There are three scripts common to all of Studies 4-7 and 9 (utilities.js, savedata.php, and end.php), and then each study has its own index.html. The data for each of Studies 4-7 and 9 are split into four csv files: demographics, stage 1 training, stage 2 test, and month attributes. It should be clear from the codebooks and analysis scripts.

In addition to downloading the materials, you can also go through the experiments yourself with the following links: Pilot, Study 1, Study 2, Study 3, Study 4/5 (short version), Study 6 (short version), Study 7 (short version), Study 8, Study 9 (short version). (The short versions skip most of the long training, so that you can see the whole experiment.) The materials are modified to remove all identifying information, consent forms, database logins, etc.

Note that some analysis scripts (for Studies 1-2 and 4-6) are for multiple studies at once; they include a prompt, asking you which study you want to analyze. A special note for analyzing Study 7: ResearchBox automatically zips the data file "study7_s1.csv" when you download it. So before running "study7_code.R", unzip "study7_s1.csv.zip" and then rename the file that emerges to "study7_s1.csv".

The analysis scripts assume that all the data files are in the same folder; so put them all together before running them.

For the github repo version of this, see: https://github.com/adammmorris/consideration-sets.

For any questions (or if you find any errors), please reach out to me at thatadammorris@gmail.com! :)

This version: March 02, 2021