The Bias Against Creativity at Work
A 2018 gift book prints fifty-six creative habits as instructions and never mentions anybody judging the work. The evaluation research fills that gap, and the shape it finds is an arch rather than an aversion.
Answer to the question
What is the bias against creativity at work?
A tendency to mark novel ideas down while saying you want them. Jennifer Mueller and two co-authors proposed in 2012 that a motive to reduce uncertainty triggers it, and found that people primed to believe problems have one correct solution rated the same new product as less creative than people primed the other way.
Key takeaways
- In a randomised grant round producing 2,130 evaluator-proposal pairs, led by Kevin Boudreau, highly novel proposals scored lower, and so did proposals close to an evaluator's own expertise.
- Jennifer Mueller and two co-authors ran two experiments, with 73 and 140 participants, inducing uncertainty once by a random payment lottery and once by a one-right-answer essay, and the primed group rated the same running shoe as less creative.
- Paola Criscuolo and three co-authors found funding rising with novelty and then falling inside a professional services firm, so some novelty raised the share of requested funds received and too much was selected against.
- No direct replication of the 2012 experiments could be located, either way, for a paper cited over five hundred times, and that is not the same as a failed replication.
2,130 evaluator-proposal pairs in a real grant round, reviewers randomly assigned. The highly novel proposals scored lower.
That is the cleanest evidence for a bias against creativity.
It is not the whole shape of it.
Change the room, not the pitch
The useful finding about the bias against creativity is not about how to sell an idea. It is about who is in the room and what state they are in. The features that track the outcome in Paola Criscuolo's study of one firm belong to the panel, not to the proposal or the person holding it, so the lever sits with whoever books the meeting. Four moves follow, all cheap, and all our inference from correlations rather than anything trialled:
- Do not schedule the decision for a week when the panel is buried, because heavier workload went with less appetite for novelty.
- Put people with genuinely different expertise in the room, because a mixed panel went with a stronger one.
- Keep remote reviewers closer to the work than a document allows, because a shared location went with a stronger one too.
- Settle what the decision requires before anyone sees the idea, since an unsettled process is its own source of unease.
None of that is advice about pitching. The reason it works, if it works, is that a bias against creativity appears to run on uncertainty rather than on taste. Change the uncertainty and you change the reception.
Follow the uncertainty, not the dislike
Jennifer Mueller and two co-authors named the paradox in 2012: people say they want creative ideas and then mark them down. Their proposal was that nobody has to dislike novelty for this to happen. What is needed is a motive to get the question settled, and novelty keeps a question open.
The authors credit Barry Staw's 1995 chapter "Why No One Really Wants Creativity" as the paper's inspiration, and that credit is worth passing on.
What makes their account persuasive is how little it took to move people. In the first experiment, the uncertainty group were told an extra payment would be decided by a random draw rather than by how well they did. In the second, one group wrote a short essay arguing that every problem has more than one correct solution, the other that it has exactly one.
A lottery over a few dollars, and an essay. Nothing more. Moritz Reis and Wilfried Kunde at the University of Würzburg brought a different instrument in 2024, asking 67 and then 68 adults to choose between traditional and creative uses for everyday objects, reading the pull towards the familiar option off how the choices moved rather than off an association test. Repeating the objects until they were familiar did not remove that pull, which rules out the fix most innovation workshops sell.
Their third result is the one worth keeping. The pull towards the ordinary option was much weaker when the creative option was one the chooser had generated themselves. That is the best available explanation for why the person who had the idea cannot read the room.
Read the experiments before repeating them
The 2012 experiments were small, and the figures here come from the authors' manuscript rather than the journal's paywalled version of record. Seventy-three people took part in the first and 140 undergraduates in the second.
The finding the article rests on is in the second. Both groups rated the same pre-tested object, a running shoe using nanotechnology to vary its fabric thickness, cooling the foot and reducing blisters. The group primed to believe that problems have one correct solution rated it as less creative than the group primed the other way. Same shoe, different rating, and the bias against creativity is the authors' name for that gap.
Nothing about the idea changed. The only thing that changed was how badly the people judging it wanted the question settled.
Their own framing is more careful than the way the finding travels. The claim is that a motive to reduce uncertainty brings negative associations with novelty to mind, which lowers the evaluation, however open-minded the person is. Two cautions belong beside it.
The first is narrower than it is usually made to sound. What showed nothing the study could detect was the explicit measure of generic attitudes, creativity set against practicality. The headline result is itself explicit, a seven-point rating of how creative the shoe was. The implicit association test, which times how readily two ideas pair in the mind, was a partial mediator, and the direct path from the manipulation to the rating still stood.
The second is that nobody appears to have run the original again. Two targeted searches turned up follow-ups and press coverage, no registered replication, no multi-lab attempt and no published failure. For a paper cited over five hundred times that absence is worth reporting, and is not the same as a failed replication.
Which is why the field evidence carries the weight in any claim about a bias against creativity. Kevin Boudreau and three co-authors ran a real grant round at a leading research university, randomising which evaluators saw which proposals and producing 2,130 evaluator-proposal pairs. Highly novel proposals scored lower. So, unexpectedly, did proposals close to an evaluator's own expertise.
Then the complication, which is the study the practical section came from. Criscuolo and three co-authors examined project selection inside a large professional services firm, where funding rose with novelty and then fell. Projects carrying some novelty received a larger share of the funds they asked for. The most novel received less, being harder to appreciate.
That shape is more useful than an aversion, and it is the one we would defend, with the design stated plainly. Criscuolo's is archive data from a single firm with nothing randomised, so it cannot settle cause and is not the correction to a randomised experiment. We keep it because it is the only study here that watched real money being allocated, and read it as complicating the slogan rather than refuting Mueller: evaluators are not hostile to new ideas, they are priced out of the ones that would take longest to understand.
Keep the two literatures apart
Nothing above is evidence about shoppers, and the consumer research points the other way.
David Alexander and two co-authors ran four field studies on new products, meaning products whose usefulness a buyer cannot estimate from anything used before. Positive purchase intentions for those products were followed through less often than intentions for incrementally new ones, and the shortfall grew over time. Novelty did not test badly. It tested fine and then failed.
Steve Hoeffler had shown in 2003 that buyers face more uncertainty estimating the usefulness of such products, and that standard preference measurement handles them unreliably. No study in reach takes a novel concept through ordinary pre-launch testing and shows the novelty itself depressed the score. The bridge between the literatures has not been built.
The seed shows the same gap from the other side. Emily Gosling's Great Minds Don't Think Alike, a 2018 gift book, gives fifty-six creative figures a spread each under a monospace running head that reads as an instruction: GET OUTSIDE, EMBRACE CHAOS, PURSUE THE NEW. That last sits over page 123, quoting its subject's own manifesto, that nothing new comes from a situation without suffering.
In 125 pages about people who made original work, there is not one page about anybody judging it. No editor, no panel, no audience.
The question the research is about does not arise in the book, and the reason is instructive. A habit becomes advice once its owner is famous, and a bias against creativity is what the same habit meets beforehand. Nobody prints a running head over a failure.
A habit becomes advice once its owner is famous. The bias against creativity is what the same habit meets beforehand.
What this evidence cannot carry
So the limits are real, and so is the residue.
This research measures ratings and funding shares rather than the moment an idea is actually killed, the samples behind the laboratory half were small and young, the field half is observational and drawn from one firm, and the consumer version of the claim, that genuinely new products test badly before launch, is the part nobody has established, which is worth knowing before you use any of it to explain why your own best idea was turned down.
It is the case for not showing unfinished work to a hurried room, and part of the reason the best song does not win, and it is worth holding on to, because this research measures ratings and funding shares rather than the moment an idea is actually killed, the laboratory samples were small and young, the field half is observational and drawn from one firm, and the consumer version of the claim, that genuinely new products test badly before launch, is the part nobody has established.
Common questions
What is the bias against creativity at work?
A tendency to mark novel ideas down while saying you want them. Jennifer Mueller and two co-authors proposed in 2012 that a motive to reduce uncertainty triggers it, and found that people primed to believe problems have one correct solution rated the same new product as less creative than people primed the other way.
Do organisations reject creative ideas?
They discount the most novel ones. In a randomised grant round producing 2,130 evaluator-proposal pairs, highly novel proposals scored lower. In archive data from one professional services firm, which cannot settle cause, Paola Criscuolo and colleagues found funding rising with novelty then falling, so some novelty won a larger share of the funds requested and the most novel won less.
Has the bias against creativity been replicated?
Not directly, as far as we can establish. Two targeted searches found no registered replication, no multi-lab attempt and no published failure of the 2012 experiments, which is not the same as a failed replication. The nearest thing is a 2024 study by Moritz Reis and Wilfried Kunde finding a similar pull towards traditional options with a different measure.
Does the bias against creativity apply to consumers?
The evidence does not support that yet. Nothing in reach shows a novel concept scoring worse in standard pre-launch testing because of its novelty. What exists is the opposite failure: David Alexander and two co-authors found warm stated intentions towards new products followed through less often than intentions towards incremental ones.
Sources
Mueller, J. S., Melwani, S., & Goncalo, J. A. (2012). 'The Bias Against Creativity: Why People Desire but Reject Creative Ideas.' Psychological Science, 23(1), 13-17. doi:10.1177/0956797611421018. [FULL TEXT READ, Cornell ILR DigitalCommons AUTHOR VERSION headed 'In press at Psychological Science', copy hosted at harrt.ucla.edu; the Sage version of record is paywalled and was NOT read, so every figure is from the author version and is attributed that way in the article. Author order confirmed at the Crossref item record: Mueller first, Wharton; Melwani second, University of North Carolina at Chapel Hill; Goncalo third, Cornell. Published online 29 November 2011, print issue January 2012. WARNING: OpenAlex also holds a Carolina Digital Repository record, doi:10.17615/8h1c-s497, which lists Goncalo first; that is a repository artefact and Crossref is the record]. Experiment 1: 73 participants, uncertainty induced by telling one group an additional payment would be decided by a random lottery rather than by performance, with a pilot of 82 verifying the manipulation; the dependent measure was an Implicit Association Test plus a seven-point explicit measure, and the explicit difference was nothing the study could detect. Experiment 2: 140 undergraduates, uncertainty induced by an essay arguing that every problem has one correct solution against one arguing it has more than one; the low-tolerance group rated the same pre-tested running shoe as less creative. The acknowledgements credit Barry Staw's 1995 chapter 'Why No One Really Wants Creativity' as the paper's inspiration.
Reis, M., & Kunde, W. (2024). 'How to overcome biases against creativity: The role of familiarity with and confidence in original solutions.' Cognition, 245, 105741. doi:10.1016/j.cognition.2024.105741. [Abstract read in full at OpenAlex. Author order confirmed at OpenAlex: Reis first and corresponding, Kunde second, both Institute of Psychology, University of Wuerzburg]. The closest thing in reach to a conceptual replication, using a different instrument: two studies, with 67 and 68 international adults, choosing between traditional and creative uses for given objects. Spatial and temporal measures of response selection indicated an implicit bias towards the traditional approach independent of repetition type, and the pull was considerably reduced for self-generated creative ideas, so building confidence rather than familiarity is the authors' suggested route.
Boudreau, K. J., Guinan, E. C., Lakhani, K. R., & Riedl, C. (2016). 'Looking Across and Looking Beyond the Knowledge Frontier: Intellectual Distance, Novelty, and Resource Allocation in Science.' Management Science, 62(10), 2765-2783. doi:10.1287/mnsc.2015.2285. [Abstract read in full at OpenAlex. Author order confirmed at OpenAlex: Boudreau first, London Business School and Harvard Business School; Guinan second, Dana-Farber/Harvard Cancer Center; Lakhani third, Harvard Business School; Riedl last, Northeastern]. A designed and executed grant proposal process at a leading research university with randomised assignment of evaluators, generating 2,130 evaluator-proposal pairs: evaluators systematically gave lower scores to proposals closer to their own areas of expertise and to those that were highly novel. NOT VERIFIED: the number of individual evaluators and proposals, the discipline mix, and the size of the novelty penalty; '2,130 evaluator-proposal pairs' is the only number printed.
Criscuolo, P., Dahlander, L., Grohsjean, T., & Salter, A. (2017). 'Evaluating Novelty: The Role of Panels in the Selection of R&D Projects.' Academy of Management Journal, 60(2), 433-460. doi:10.5465/amj.2014.0861. [Abstract read in full at OpenAlex. Author order confirmed at OpenAlex: Criscuolo first, Imperial College London; Dahlander second, ESMT Berlin; Grohsjean third, Bocconi; Salter last, University of Bath. OpenAlex gives 2016 while volume 60 issue 2 is April 2017; 2017 is printed and the discrepancy noted]. A multi-source, multi-method study of research and development projects in a leading professional services firm: organisations were more likely to fund projects with intermediate levels of novelty, so some novelty increased the share of requested funds received while too much was difficult to appreciate and was selected against, with high panel workload reducing the preference for novelty and diversity of panel expertise and shared location increasing it. NOT VERIFIED: project count, panel count, sector beyond 'professional services', period covered, and how novelty was measured.
Alexander, D. L., Lynch, J. G., & Wang, Q. (2008). 'As Time Goes By: Do Cold Feet Follow Warm Intentions for Really New versus Incrementally New Products?' Journal of Marketing Research, 45(3), 307-319. doi:10.1509/jmkr.45.3.307. [Abstract read in full at OpenAlex. Author order confirmed at OpenAlex: Alexander first, Opus College of Business, University of St Thomas; Lynch second, Fuqua School of Business, Duke; Wang last, Warwick Business School]. Four field studies: consumers followed through less often on positive purchase intentions for really new products than for incrementally new ones, and the decrement grew over time, with buyers of really new products thinking less concretely about the circumstances of use and being more poorly calibrated about initial use. NOT VERIFIED: the sample sizes of the four field studies, the product categories, and the magnitude of the gap.
Hoeffler, S. (2003). 'Measuring Preferences for Really New Products.' Journal of Marketing Research, 40(4), 406-420. doi:10.1509/jmkr.40.4.406.19394. [Abstract read in full at OpenAlex. Sole author, corresponding, University of North Carolina]. Consumers have greater uncertainty estimating the usefulness of really new products than of incremental ones, and cope with inferential techniques that standard preference measurement such as conjoint does not capture well; incorporating mental simulation and analogies showed some methods enhancing and others hindering predictive accuracy. It does NOT show that novel products test badly, which the article states. NOT VERIFIED: sample sizes, number of studies, products used, and which methods helped or hurt.
Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J., & Tetlock, P. E. (2013). 'Predicting ethnic and racial discrimination: A meta-analysis of IAT criterion studies.' Journal of Personality and Social Psychology, 105(2), 171-192. doi:10.1037/a0032734. [Abstract read in full at OpenAlex. Author order confirmed at OpenAlex: Oswald first, Rice University; Mitchell second, University of Virginia School of Law; Blanton third; Jaccard fourth; Tetlock last, Wharton]. A meta-analysis of the predictive validity of the Implicit Association Test and of explicit measures across two intergroup domains and six criterion categories: IATs were poor predictors of every criterion category other than brain activity, and performed no better than simple explicit measures. SCOPE LIMIT stated prominently in the article: this concerns racial and ethnic IATs predicting discrimination and does not test the creativity-versus-practicality instrument, and it bounds how the 2012 Experiment 1 may be reported rather than refuting it.