Decision Making
Preference Change Task
Also called: free-choice paradigm, choice-induced preference change task, spreading of alternatives task, rate-choose-rate task
Measure how choosing changes liking: participants rate foods, choose between pairs they rated alike, and rate again, with pairs chosen last as the control.
7-9 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 27
- Evidence grade
- D
- Registered clinical studies
- 0
How much do you like this food?
What it measures
Measures how choosing between two foods changes how much people like them. Participants rate pictured foods, choose between pairs they rated alike, and rate the foods again; half the pairs are chosen only after the second ratings, as a control for what choices reveal. No food is offered.
Constructs
- Choice-induced preference change
- Cognitive dissonance
- Preference stability
Research uses
- Judgment and decision-making research
- Cognitive dissonance and attitude research at the group level
- Within-study comparison of preference change
How the paradigm works
In the free-choice paradigm, people rate items, choose between two they rated alike, and rate them again; afterwards the chosen item tends to be rated higher and the rejected one lower, the spreading of alternatives (Brehm, 1956). Chen and Risen (2010) showed that spreading appears even when preferences do not change: first ratings measure preferences with error, and a choice reveals which item was really liked more. In their first study, art prints chosen only after a second ranking spread about as much (1.75 ranks) as prints chosen between the two rankings (2.20). Their control, rate, rate, then choose, measures this artifact, and Izuma and colleagues (2010) ran it within each person with pictures of foods. A meta-analysis of 43 studies that controlled for the artifact found an effect of choosing of d = 0.40 (Enisman, Shpitzer and Kleiman, 2021).
More about the paradigm
Participants mark how much they like each of 32 pictured foods on a bar from 0 to 100, with no numbers shown and no starting point. Each food is then paired with the food next to it in those ratings, so each pair was rated alike, and the session's seed gives one of each two neighbouring pairs to each condition. Participants choose the food they would rather eat from 8 pairs, mark all 32 foods again in a new order, and then choose from the other 8 pairs. No food is offered. A practice rating and a practice choice with two other foods come first. The pictures are Noto Emoji drawings, shown without names.
The main score, preference change, is the mean spreading of the pairs chosen between the two ratings minus that of the pairs chosen after both, in points. The spreading of a pair is the chosen food's change from the first rating to the second minus the rejected food's change. Spreading in the pairs chosen last cannot come from choosing, so it estimates the artifact, and the difference estimates the change that followed the choice. The task also reports both spreadings, the changes of chosen and rejected foods, how often choices followed the first and second ratings, how consistent the two ratings were and rating and choice times. One person's score is noisy, so compare groups. Scores support research and do not diagnose.
Scores
- Preference Changepoints
- Mean spreading of the pairs chosen between the two ratings minus that of the pairs chosen after both, on the 0 to 100 bar.
- Positive values mean ratings moved toward the choice beyond what the choice revealed; not better or worse
- Spreading, Chosen Before Rating Againpoints
- Mean spreading of the 8 pairs chosen between the two ratings: the chosen food's change minus the rejected food's change.
- Includes what the choice revealed
- Spreading, Chosen After Rating Againpoints
- Mean spreading of the 8 pairs chosen after both ratings, computed the same way.
- What the choice reveals, with no effect of choosing
- Change, Chosen Beforepoints
- Mean change from the first rating to the second of the foods chosen between the ratings.
- Informational
- Change, Rejected Beforepoints
- Mean change from the first rating to the second of the foods rejected between the ratings.
- Informational
- Change, Chosen Afterpoints
- Mean change from the first rating to the second of the foods chosen after both ratings.
- Informational
- Change, Rejected Afterpoints
- Mean change from the first rating to the second of the foods rejected after both ratings.
- Informational
- Choices Following First Ratings, Before%
- Share of the pairs chosen between the ratings in which the food rated higher the first time was chosen, among pairs rated differently.
- Compare with the same share for the pairs chosen after
- Choices Following First Ratings, After%
- Share of the pairs chosen after both ratings in which the food rated higher the first time was chosen, among pairs rated differently.
- Compare with the same share for the pairs chosen before
- Choices Following Second Ratings, After%
- Share of the pairs chosen after both ratings in which the food rated higher the second time was chosen, among pairs rated differently.
- Informational
- Rating Consistencyscore
- Correlation of the first and second ratings of the 16 foods in the pairs chosen after both ratings, rated twice before any choice about them.
- Higher means more stable ratings
- Mean Pair Gappoints
- Mean difference between the first ratings of the two foods in a pair.
- Small values mean the pairs were rated alike
- Left Choices%
- Share of the choices that went to the food on the left.
- Quality check
- Median Rating Timems
- Median time from the start of a rating screen to Confirm, leaving out times under 100 ms.
- Informational
- Median Choice Timems
- Median time from the start of a choice screen to the choice, leaving out times under 100 ms.
- Informational
- Ratings Givenitems
- Ratings in the scored run, out of 64.
- Quality check
- Choices Madeitems
- Choices in the scored run, out of 16.
- Quality check
- Longest Run of One Ratingitems
- Most foods in a row given exactly the same rating.
- Quality check
- Anticipationscount
- Ratings and choices made within 100 ms of the start of their screen.
- Quality check
Use and evidence
27 papers from 2021 to 2025 name this task in their title or abstract, down 13% compared with 2016 to 2020. That ranks it 372 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 7 |
| 2024 | 4 |
| 2023 | 4 |
| 2022 | 4 |
| 2021 | 8 |
| 2020 | 9 |
| 2019 | 5 |
| 2018 | 10 |
| 2017 | 5 |
| 2016 | 2 |
| 2015 | 10 |
| 2014 | 5 |
| 2013 | 4 |
| 2012 | 6 |
| 2011 | 2 |
| 2010 | 4 |
| 2009 | 4 |
| 2008 | 1 |
| 2007 | 1 |
| 2006 | 1 |
| 2005 | 2 |
| 2004 | 1 |
| 2003 | 0 |
| 2002 | 1 |
| 2001 | 0 |
| 2000 | 0 |
Use grade E
Fewer than 30 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 7
- Registered clinical studies using it as an outcome
- 0
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Foods
- 32 foods shown as pictures without names, from meals, snacks, sweets, fruit and nuts, a list written for this test.
- Ratings
- How much do you like this food? A 0 to 100 bar with no numbers shown and no starting point, from Not at all to Very much, with no time limit. Every food twice, in two seeded orders.
- Pairs
- After the first ratings, each food is paired with the food next to it in those ratings, ties in a seeded order: 16 pairs, each rated alike. Of each two neighbouring pairs, the seed gives one to each condition.
- Choices
- Which would you rather eat? The two foods side by side, a click or a tap on one. 8 pairs between the two ratings (rate, choose, rate) and 8 after both (rate, rate, choose), with the side of each food from the seed.
- Score
- The spreading of a pair is the chosen food's change from the first rating to the second minus the rejected food's change. The score is the mean spreading of the pairs chosen between the ratings minus that of the pairs chosen after both.
- Practice
- One rating and one choice with two other foods.
- No food
- The foods are pictures and nothing is given; participants are asked to choose as if they would get the food they choose.
Spreading after a choice can appear even when preferences do not change, because a choice reveals which of two items was really liked more (Chen and Risen, 2010). The pairs chosen after both ratings carry that artifact without any effect of choosing, so the score subtracts their spreading. If a choice follows a second rating just given more than a first one, the control's spreading grows and the score leans toward zero. One person's score is noisy, so compare groups. The scores support research and do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: preference_change_rating and preference_change_choice from the scored run, two ratings per food and one choice per pair, with no correct answer. Fewer than all 64 ratings or all 16 choices invalidate. The same rating on 32 or more foods in a row invalidates, and choices on one side of the screen 90% of the time or more flag the session for review. Answers under 200 ms on at least half the ratings and choices invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. The score compares the spreading of pairs chosen before and after the second ratings, so it needs every rating and every choice. A run as long as a whole pass over the foods separates nothing (long-string checks; Curran 2016), and choices that follow one side of the screen do not reflect the foods. An answer under 200 ms is faster than a simple visual reaction, so it cannot reflect the item. Following response-time effort screening for untimed tests (Wise and Kong 2005), the session is invalid only when at least half the answers are that fast.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
This test runs with fixed settings.
Need a different version of this test?
Devices and languages
- Untimed test: runs on desktops, laptops, tablets and phones. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Preference Change Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Brehm, 1956; Chen & Risen, 2010). The main score was preference change (mean spreading of the pairs chosen between the two ratings minus that of the pairs chosen after both, on the 0 to 100 bar).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Brehm, J. W. (1956). Postdecision changes in the desirability of alternatives. The Journal of Abnormal and Social Psychology, 52(3), 384-389. https://doi.org/10.1037/h0041006
- Chen, M. K., & Risen, J. L. (2010). How choice affects and reflects preferences: Revisiting the free-choice paradigm. Journal of Personality and Social Psychology, 99(4), 573-594. https://doi.org/10.1037/a0020217The rate, rate, then choose control.
- Izuma, K., Matsumoto, M., Murayama, K., Samejima, K., Sadato, N., & Matsumoto, K. (2010). Neural correlates of cognitive dissonance and choice-induced preference change. Proceedings of the National Academy of Sciences, 107(51), 22014-22019. https://doi.org/10.1073/pnas.1011879108The control within each person, with pairs of rated foods.
- Izuma, K., & Murayama, K. (2013). Choice-induced preference change in the free-choice paradigm: A critical methodological review. Frontiers in Psychology, 4, 41. https://doi.org/10.3389/fpsyg.2013.00041
- Enisman, M., Shpitzer, H., & Kleiman, T. (2021). Choice changes preferences, not merely reflects them: A meta-analysis of the artifact-free free-choice paradigm. Journal of Personality and Social Psychology, 120(1), 16-29. https://doi.org/10.1037/pspa0000263An effect of choosing of d = 0.40 across 43 studies that controlled for the artifact.
Source papers
- Chen MK, Risen JL (2010). How choice affects and reflects preferences: Revisiting the free-choice paradigm. Journal of Personality and Social Psychology, 99(4):573-594.Paradigm source, Adults
- Enisman M, Shpitzer H, Kleiman T (2021). Choice changes preferences, not merely reflects them: A meta-analysis of the artifact-free free-choice paradigm. Journal of Personality and Social Psychology, 120(1):16-29.Paradigm source, Adults
- No normative data for this task.Paradigm source, N/A
Questions
How long does the Preference Change Task take?
About 7 to 9 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Preference Change, Spreading, Chosen Before Rating Again, Spreading, Chosen After Rating Again, Change, Chosen Before, Change, Rejected Before, Change, Chosen After, Change, Rejected After, Choices Following First Ratings, Before, Choices Following First Ratings, After, Choices Following Second Ratings, After, Rating Consistency, Mean Pair Gap, Left Choices, Median Rating Time, Median Choice Time, Ratings Given, Choices Made, Longest Run of One Rating and Anticipations. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Preference Change Task used in research?
27 papers from 2021 to 2025 name it in their title or abstract, down 13% compared with 2016 to 2020. That ranks it 372 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Preference Change Task have?
In the published research, 0 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.