In short
The moral credential effect is the finding that people who have just done something that establishes them as fair or unprejudiced become more willing to express an attitude or make a choice that could be seen as prejudiced. It is looked for in hiring panels that pick one diverse candidate and then relax, in the phrase "I'm not racist, but," and in anyone who treats a past good act as a credential that covers the next judgment; its size, and whether it is reliable at all, is now contested.
How the moral credential effect works
The proposed mechanism is a change in what a person thinks an action will say about them. Many people hold views they would rather not voice, because voicing them risks being seen as prejudiced. Monin and Miller argued that a recent act which plainly contradicts that reading, such as choosing the best-qualified candidate when she happens to be a woman, works as a credential: it lowers the risk that the next statement will be taken as evidence of prejudice, and so lowers the cost of making it. The person's underlying attitude need not change. What changes is their estimate of how the attitude will be read.
Read the diagram as text
A balanced prior choice earns a check badge. The same badge becomes the crest of a dashed shield surrounding a later choice between people, with one accepted and one rejected. An incoming scrutiny arrow stops at the shield. The shield and transfer route are explicitly hypothetical; the effect is contested.
A fair act spent as cover
Conceptual sequence; the effect is contested
Picks the standout applicant: The best-qualified candidate happens to be a woman or a Black man
Credential: “Nothing I say next can be read as prejudice.”
Second job judged more freely: Rated as better suited to a man or a White person, more than controls (2001)
Does the past choice bear on this one?: Each decision has its own evidence; the question is skipped
Picks the standout applicant → lowers the expected cost of speaking → Credential
Credential → treated as permission → Second job judged more freely (error transition)
This is why the effect was framed as licensing rather than as hypocrisy. In Monin and Miller's third study the audience for the second judgment did not know about the credential, and participants expressed the view anyway, which suggests the license is granted by the person to themselves, not only performed for others. Later writers folded the finding into the broader idea of moral self-licensing, reviewed by Merritt, Effron and Monin in 2010: a good deed in one domain making a questionable one feel permissible. Effron, Cameron and Monin reported in 2009 that people who had been given the chance to express support for Barack Obama were afterwards more willing to favor White people in a hypothetical allocation of money, and that in the study which also measured racial prejudice, the allocation effect appeared only among participants who scored high on that measure.
The evidence has not held up well under larger tests. Blanken, van de Ven and Zeelenberg's 2015 meta-analysis of 91 moral licensing studies with 7,397 participants estimated the overall effect at a Cohen's d of 0.31 and found that published studies showed larger effects than unpublished ones, a sign of publication bias. In 2014 the same group had reported three failed attempts to replicate a different licensing result, Sachdeva, Iliev and Medin's 2009 finding that people who wrote about themselves in positive trait words then gave less to charity; that bears on the wider licensing literature, not on Monin and Miller's design. In 2024 Xiao, Li, Au, Tan, Chung and Feldman ran a preregistered replication of Monin and Miller's second study with 932 US participants: the effect was close to zero in the race version and ran in the opposite direction in the sex version. The Many Labs 3 project had earlier replicated the first study with a much smaller effect than the original.
The mechanism is plausible, the original demonstrations were real, and the effect is small or absent in the large preregistered tests that have followed. The demo above reproduces the design so that the live split can be compared with both the 2001 result and the 2024 one.
The original study
Benoit Monin and Dale Miller published "Moral credentials and the expression of prejudice" in 2001, reporting three experiments. In the first, participants marked five statements right or wrong. For one group the statements were blatantly sexist, such as the claim that most women are better off at home caring for children, which almost everyone rejected; for the other group they were milder versions that were harder to disagree with. Everyone then read about a construction company seeking a sales representative, a job described in stereotypically male terms, and rated whether it suited a man or a woman better on a seven-point scale with a neutral midpoint. Men who had rejected the blatant statements favored a man more than men who had read the mild ones; for women the difference was negligible.
The second study, which the demo follows, used a choice rather than a questionnaire. Participants picked one of five applicants for a starting position at a consulting firm. One applicant was made to stand out, with the best grades and the most prestigious college; in the credential conditions that applicant was a woman or a Black man, and in the control condition a White man, the other four being White men throughout. Most participants in every condition chose the standout, 110 of 132. They then judged a second job, the construction sales role in the sex version or a police position in a racially hostile town in the race version. Participants who had been given the chance to pick a woman or a Black candidate rated the second job as better suited to a man or a White person more strongly than the controls, and the pattern appeared both in the full sample of 132 and when the analysis was restricted to the 110 who had chosen the standout.
The third study tested whether the license depended on an audience. Participants with established credentials were more willing to express a politically incorrect opinion even when their audience knew nothing about the credential.
The 2024 replication by Xiao and colleagues reproduced the second study's materials closely, with the applicant task, the same two second-job scenarios and the same seven-point response scale, in an online sample of 932 after exclusions. The share choosing the standout applicant matched the original. The credential effect did not appear: near zero for the race scenario and reversed for the sex scenario, where participants who had met a female standout were, if anything, less likely to favor a man.
Everyday examples
The rows below are places where a past good act is used as cover for a present judgment. Given the replication record, they are risks to watch for, not demonstrated effects.
| Situation | What the bias does | What helps |
|---|---|---|
| A hiring panel that has just made one diverse appointment | Members feel the fairness question is settled and apply stereotyped criteria to the next role with less self-scrutiny | Score each role on fixed, written criteria, and review the whole year's appointments together |
| The sentence that begins "I'm not prejudiced, but" | A self-awarded credential is presented so that the clause after it can be said | Judge the clause on its own, as if the preamble were not there |
| A company that publishes a diversity statement | The statement is treated as evidence that individual decisions no longer need checking | Measure the decisions, not the statement |
| A donor who has just given to a cause | Feels entitled to a sharper remark about the people the cause serves | Separate the giving from the talking; neither licenses the other |
| Citing a friend from the group in question | The friendship is offered as proof that a generalization about the group is not prejudice | Ask whether the generalization would be acceptable from someone without the friend |
How to counter it
Whether or not the license is large, the habit of treating past fairness as present cover is one worth breaking.
- Judge each decision on its own evidence. A fair choice last week is not information about whether this week's choice is fair.
- Write the criteria before you see the candidates, and apply them to every role, not only the ones where you feel your fairness is being tested.
- Notice the preamble. When you hear yourself, or someone else, establishing credentials before making a point, treat the point as if the credentials had not been mentioned.
- Keep a record over time. Licensing, if it operates, works decision by decision; a year of decisions reviewed together shows a pattern that single choices hide.
- Do not lean on the effect in the other direction either. The replication record means you should not assume that a colleague's one fair act predicts a lapse, any more than it proves a virtue.
What it is not
Several entries describe how group membership shapes judgment, and one describes the broader family this effect belongs to.
- It is not implicit stereotyping or stereotyping as such. Those are the associations and generalizations themselves. The credential effect is about when a person feels free to act on or voice them; it presupposes an attitude and changes only its expression.
- It is not in-group favoritism, the tendency to treat members of one's own group better. The credential can be earned by favoring an out-group member and spent by disfavoring one; the direction is set by what looks fair, not by group loyalty.
- It is not self-consistency bias, the tendency to see one's past attitudes as matching the present ones. Licensing runs the other way: the past act is used to permit a present act that departs from it.
Tests and tools that use this
Related biases
Moral Credential Effect FAQ
What is the moral credential effect?
It is the finding, reported by Monin and Miller in 2001, that people who have just done something that marks them as unprejudiced, such as rejecting sexist statements or choosing a well-qualified woman or Black candidate, become more willing to express a view that could look prejudiced, such as saying a stereotypically male job suits a man better. The proposed mechanism is that the earlier act lowers the risk of being seen as prejudiced, so the view becomes cheaper to voice.
Does the moral credential effect replicate?
Not reliably. A 2015 meta-analysis of 91 moral licensing studies put the average effect at a Cohen's d of 0.31 with signs of publication bias. The Many Labs 3 project found a much smaller version of the first 2001 study. A 2024 preregistered replication of the second study with 932 participants found an effect near zero for race and reversed for sex. The page reports the original and the replications side by side rather than choosing between them.
What is the difference between moral credentials and moral licensing?
Moral licensing is the general idea that a good act makes a later questionable act feel permissible, across domains such as charity, consumption and honesty. The moral credential effect is the specific version about prejudice: a past act establishes you as unprejudiced and so licenses a statement or choice that might otherwise look prejudiced. Merritt, Effron and Monin's 2010 review treats credentials as one route to licensing, alongside the feeling of having earned moral credit.
How does the demo on this page work?
When the page loads it assigns you one of two versions of the first hiring decision, using a value kept only in your browser's session storage, so that reloading shows the same version. In one version the standout applicant is a woman, in the other a man; the construction company question is identical. The page records only the version and your answer, and once each version has 50 or more responses it shows the two splits side by side next to the 2001 and 2024 results.
Does the demo tell me whether I was licensed?
No. You answered one question after one version, so there is nothing in your own result to compare. The effect, where it appears, is a difference between groups: a larger share saying a man suits the job among people who met the female standout than among people who met the male one. Your answer is one data point in that comparison. Saying a man suits the job is not in itself evidence of licensing, and saying both suit it equally is not proof of its absence.
Why would being fair once make someone less fair afterwards?
The 2001 account is about appearances rather than attitudes. People avoid saying things that could be read as prejudiced; an act that visibly contradicts that reading makes the next statement feel safer, so a view that was already held gets voiced. The third 2001 study found the effect even when the audience did not know about the credential, suggesting people grant the license to themselves. Whether the effect is large enough to matter outside the laboratory is what the replications call into question.
References
- Monin, B., & Miller, D. T. (2001). Moral credentials and the expression of prejudice. Journal of Personality and Social Psychology, 81(1), 33-43. doi.org
- Xiao, Q., Li, L. C., Au, Y. L., Tan, S. N., Chung, W. T., & Feldman, G. (2024). Licensing via credentials: Replication Registered Report of Monin and Miller (2001) with extensions investigating the domain-specificity of moral credentials and the association between the credential effect and trait reputational concern. International Review of Social Psychology, 37(1), 10. doi.org
- Blanken, I., van de Ven, N., & Zeelenberg, M. (2015). A meta-analytic review of moral licensing. Personality and Social Psychology Bulletin, 41(4), 540-558. doi.org
- Merritt, A. C., Effron, D. A., & Monin, B. (2010). Moral self-licensing: When being good frees us to be bad. Social and Personality Psychology Compass, 4(5), 344-357. doi.org
- Effron, D. A., Cameron, J. S., & Monin, B. (2009). Endorsing Obama licenses favoring Whites. Journal of Experimental Social Psychology, 45(3), 590-593. doi.org
- Ebersole, C. R., et al. (2016). Many Labs 3: Evaluating participant pool quality across the academic semester via replication. Journal of Experimental Social Psychology, 67, 68-82. doi.org
- Blanken, I., van de Ven, N., Zeelenberg, M., & Meijers, M. H. C. (2014). Three attempts to replicate the moral licensing effect. Social Psychology, 45(3), 232-238. doi.org
- Sachdeva, S., Iliev, R., & Medin, D. L. (2009). Sinning saints and saintly sinners: The paradox of moral self-regulation. Psychological Science, 20(4), 523-528. doi.org
Cite this source
Psychology.com. (2026, October 24). Moral Credential Effect. Psychology.com. https://psychology.com/biases/moral-credential-effect