The evidence that art makes employees more creative is real but narrower than vendors claim. The strongest controlled result remains the University of Exeter work by Craig Knight and Alex Haslam (2010): offices enriched with art and plants lifted productivity about 17% over bare “lean” offices, and offices workers decorated themselves lifted it about 32%. A Drexel University trial (2016) found 45 minutes of art-making cut cortisol in 75% of participants. The WHO’s 2019 scoping review of more than 3,000 studies confirms arts engagement improves wellbeing, which is the channel through which creativity gains arrive. Against Gallup’s 2026 finding that only 20% of employees are engaged, costing roughly $10 trillion, art is a cheap lever: measurable, but only when employees get a say in it.
Art and workplace creativity is one of the most-cited and least-examined claims in corporate culture. Every design firm and art-rental company says paintings on the wall make teams more inventive; few show the trial that proves it. In this guide the Kurums Startup team separates the peer-reviewed research from the marketing, quantifies what the best studies actually found, and explains what HR and facilities leaders can expect if they spend on art, art-making or arts programmes. It is part of our Arts & Creative Economy hub, and pairs with our analyses of corporate art collections and employee engagement and arts-based leadership training.
What is the best evidence that art raises workplace output?
The Exeter experiments (Knight & Haslam, 2010), run with more than 2,000 office workers across surveys and lab trials, found art-and-plant “enriched” offices raised task productivity about 17% and worker-designed “empowered” offices about 32%, with no rise in errors, compared with bare lean offices.
Does making art, rather than looking at it, change anything measurable?
Yes. In Girija Kaimal’s Drexel University study (Art Therapy, 2016), 39 adults aged 18 to 59 spent 45 minutes with collage, clay and markers; 75% showed lower salivary cortisol afterwards, regardless of prior art experience. Lower stress is the best-documented precondition for divergent thinking.
How big is the problem art is supposed to fix?
Gallup’s State of the Global Workplace 2026 puts global engagement at 20% (down from a 23% peak in 2022), daily stress at 40% and the productivity cost of disengagement at roughly $10 trillion, about 9% of world GDP. Europe is lowest at 12% engaged.
What does “art makes employees more creative” actually claim?
The claim bundles three separate mechanisms: passive exposure to visual art in the office, active art-making by staff, and arts-based training. Each has its own evidence base, and the research supports them to different degrees. Exposure effects are modest; active and participatory effects are larger and better documented.
Passive exposure is the version most companies buy: a painting in the lobby, prints in meeting rooms. Active art-making, from lunchtime drawing to pottery workshops, is studied mostly through stress biomarkers and self-reported wellbeing. Arts-based training uses theatre, music or visual-observation exercises to teach noticing, listening and improvising, and is evaluated like any other training.
Creativity is measured in two ways: divergent-thinking tests (unusual uses for a brick in two minutes) and real-task output (valid ideas, errors, time to complete). Most “art at work” studies measure the second only indirectly. The Exeter work is the exception, which is why it remains the anchor citation 15 years later.
What did the Exeter experiments actually find?
Craig Knight and Alex Haslam at the University of Exeter tested four office set-ups with more than 2,000 participants across two surveys and two experiments. Relative to a lean office, an enriched office with art and plants raised productivity about 17%, and an office the worker designed raised it about 32%.
Participants completed information-processing and attention tasks in one of four conditions: lean (bare, functional desk), enriched (the same desk with plants and pictures), empowered (the worker arranged plants and pictures as they wished) and disempowered (the worker arranged the space, then a “manager” rearranged it). Productivity rose in the enriched condition and rose further in the empowered one, with no increase in error rates, according to the University of Exeter’s 2010 announcement of the research, which was funded by the Economic and Social Research Council and supported by workplace-services firm Ambius.
The disempowered condition is the part most companies ignore. When a manager undid the worker’s choices, productivity fell back towards the lean baseline and participants reported disliking the space. The lesson is not “hang art”; it is “let people shape the space they work in”. Art is one of the cheapest things to let them shape.
Two caveats matter. The lab tasks were short, so the figures describe acute effects, not a year of work; and the study tested plants plus pictures as a package, so it cannot isolate art from greenery. Knight has since quoted the enriched result as 15% to 17%; both numbers come from the same work.
Does making art reduce stress in a way that matters for creativity?
Yes, and the effect appears regardless of skill. Drexel University’s 2016 study found that 45 minutes of free art-making lowered cortisol in 75% of 39 adult participants. Because stress narrows attention and suppresses idea generation, lower cortisol is the mechanism by which art-making can plausibly raise creative output.
Girija Kaimal’s team at Drexel gave participants collage materials, clay and markers with no instruction beyond “make something”, and measured salivary cortisol before and after. Three-quarters had lower levels afterwards, with no significant difference between self-described artists and others, as reported in Drexel’s summary of the paper in the journal Art Therapy. Younger participants saw larger drops, which the authors attributed to older adults having other coping strategies.
Why does this matter for an HR budget? Gallup’s 2026 report records that 40% of employees worldwide experienced a lot of stress the previous day, rising to 45% among managers. The CIPD’s September 2025 survey puts UK sickness absence at 9.4 days per employee, the highest in more than 15 years, with mental ill health the leading cause of long-term absence. A 45-minute activity that reduces a stress biomarker in three out of four people is, at minimum, a cheap wellbeing intervention.
The limit is that cortisol is not creativity. The chain of inference runs from art-making to lower stress to wider attention to better divergent thinking; each link is separately supported, but the whole chain has not been tested in one company.
What does the WHO’s 3,000-study review say about arts and wellbeing?
The 2019 WHO Europe scoping review by Daisy Fancourt and Saoirse Finn, drawing on more than 3,000 studies, concluded that arts engagement supports mental health, reduces stress and strengthens social connection across the life course. It did not study offices directly, but it established the wellbeing channel on which every workplace claim rests.
The review, published as Health Evidence Network synthesis report 67 and summarised by the Jameel Arts & Health Lab, grouped findings into prevention and promotion (social determinants, healthy behaviours, child development) and management and treatment (mental illness, neurological and non-communicable conditions, end-of-life care). It described arts interventions as low-cost and multi-modal, and recommended that policymakers scale the better-evidenced ones.
For HR the relevant findings are those on stress, anxiety and social cohesion in adults, the outcomes Gallup and the CIPD track. The review also flagged a gap still open in 2026: few economic evaluations, so cost per unit of wellbeing gained is not well established.
The EU-backed CultureForHealth programme ran a Slovenian pilot from 2021 placing art installations, dance workshops and residencies in three companies with 23 employees; two asked for more, the third saw dropouts. The sample is tiny, but the pattern is the one practitioners keep reporting: participation is uneven, and those who opt in gain most.
How do the main studies compare?
Only a handful of studies carry real weight, and they differ sharply in design. The table below lists the ones HR teams most often cite, what each measured and how much confidence the method justifies. Controlled experiments with productivity measures rank highest; commissioned surveys rank lowest.
| Study | Design and sample | Headline result | Weight |
|---|---|---|---|
| Knight & Haslam, University of Exeter (2010) | Two surveys plus two controlled experiments; 2,000+ office workers; four office conditions | Enriched +17%, empowered +32% productivity vs lean; no error increase | High for acute effect; cannot separate art from plants |
| Kaimal et al., Drexel University, Art Therapy (2016) | Pre/post biomarker trial; 39 adults aged 18–59; 45-minute art-making | 75% had lower cortisol; no difference by art experience | Medium; small sample, no control group, stress not creativity |
| Fancourt & Finn, WHO Europe (2019) | Scoping review of 3,000+ studies on arts and health | Arts engagement improves mental health, lowers stress, builds social connection | High for wellbeing; not workplace-specific |
| Naghshineh et al., Harvard, J. Gen. Intern. Med. (2008) | Controlled course; 24 medical students vs 34 peers; eight art-observation sessions | 5.41 accurate observations vs 0.36 for controls (p<0.0001) | High for observation skill; medical, not corporate |
| Brookfield Properties / Perspectus Global (2022) | Survey; 3,250 UK office workers in ten cities | 69% say striking art at work supports wellbeing; 41% call their office poorly designed | Low; commissioned by a landlord, self-report only |
| CultureForHealth Slovenia pilot (2021–2023) | Pre/post surveys; 3 companies, 23 employees; installations, dance, residencies | Two firms asked for more; one saw dropouts; 47% had no prior creative experience | Low; tiny sample, results still being analysed |
The best-designed studies are old and the newest ones are weak. Nobody has repeated the Exeter design at scale inside a real company with modern engagement metrics.
Why does employee control over art matter more than the art itself?
Because the biggest measured gain, the 32% in the Exeter empowered condition, came from letting workers choose, not from the quality of the pictures. The same logic appears in Gallup’s engagement research, where autonomy and feeling heard are among the strongest predictors of performance.
Gallup’s Q12 meta-analysis, which pools 456 studies across 276 organisations, 54 industries and 96 countries covering about 2.7 million employees, finds that teams in the top engagement quartile show 23% higher profitability, 18% higher sales productivity and 81% lower absenteeism than bottom-quartile teams, according to Gallup’s summary. Engagement rises when employees have a say in their work and environment; art selection is one of the few environmental decisions a company can hand over without operational risk.
Progressive Insurance has run this principle at scale since 1974. Its collection of about 10,000 works is rotated across more than 400 US offices, with no art kept in storage, and an art-education programme reaches all 39,000 employees. The company also stages a “People’s Biennial” of work made by its own staff. Progressive’s stated purpose is provocation: art in the workplace is meant to prompt employees to voice critical views on the art and on the times. That is a deliberate use of the empowered condition rather than a decorative one.
The disempowered condition has corporate analogues too: JPMorgan took down a Jason Seley bumper sculpture in 1960 by the end of the lunch hour after employee protest. Art imposed from above can cost goodwill faster than it earns it.
Does arts training transfer to business skills such as observation and listening?
The clearest transfer evidence comes from medicine. Harvard’s “Training the Eye” course, evaluated in the Journal of General Internal Medicine in 2008, raised medical students’ accurate clinical observations from 0.36 to 5.41 per case after eight sessions of guided art viewing. Skills built on paintings transferred to patients.
Sheila Naghshineh and colleagues compared 24 pre-clinical students who took the eight-session museum course with 34 classmates who did not. The intervention group produced a mean of 5.41 accurate observations per clinical image afterwards, against 0.36 for the controls (p<0.0001), and students who attended all eight sessions scored higher than those who missed one or more. The published paper is at Springer. The method, Visual Thinking Strategies, is now used by medical schools in several countries and is sold to corporate audiences by museums and consultancies.
The corporate version of the evidence is thinner. Providers such as Second City Works, which says it has worked with more than 600 Fortune 1000 companies, and Olivier Mythodrama, founded in 1997 and listing Microsoft, Rolls-Royce, Pfizer and Diageo among clients, sell improvisation and Shakespeare-based leadership programmes. Their sites describe the programmes as research-based but publish no controlled outcome data. We examine the full market in our guide to arts-based leadership training.
The defensible position for an L&D buyer is that observation and listening are trainable through art, that the medical trials prove transfer to a demanding professional task, and that corporate transfer is plausible but unmeasured. Training magazine’s 2024 Industry Report puts US corporate training spend at $98 billion, down 3.7%, with per-learner spend falling from $954 to $774. In that climate, arts-based courses need to show their own data.
How should a company measure whether art at work is working?
Treat it as an experiment with a control group. Pick two comparable teams or floors, change the art or art programme on one, hold the other steady, and track engagement pulse scores, absence and a task-output measure over at least a quarter. Without a comparison the result will be a story, not a number.
Three metrics are cheap and credible: a short engagement pulse using the items your company already uses; absence days, which the CIPD shows are rising; and something specific to the team’s work, such as tickets closed or defects found. The Exeter design worked because it measured the task, not the feeling.
Against a disengagement cost that Gallup values at roughly $10 trillion globally, or around 9% of GDP, the question is not whether art is affordable but whether it is measured.
In Türkiye, corporate art is mostly a foundation activity rather than an HR one: Borusan Contemporary in Istanbul’s Perili Köşk operates as Borusan Holding’s offices on weekdays and a public museum at weekends (adult admission 350 TL in 2026), while Koç’s Arter and Eczacıbaşı’s Istanbul Modern anchor the Istanbul scene. Few Turkish employers have published employee-outcome data, which leaves room for the first one that does. We cover the wider national picture in Türkiye’s creative economy playbook.
Frequently Asked Questions
Is there a single study that proves art increases creativity at work?
No. The closest is the University of Exeter research by Knight and Haslam (2010), which measured productivity, not creativity, and found gains of about 17% in art-and-plant enriched offices and 32% in worker-designed ones. Creativity-specific outcomes are supported indirectly, through lower stress (Drexel 2016) and better observation (Harvard 2008). Anyone quoting a single “art raises creativity by X%” figure is usually stretching one of these.
Does the type of art matter, abstract versus figurative?
The controlled evidence does not distinguish. The Exeter work used generic pictures and plants, and the gain came from enrichment and control rather than style. Survey data, such as the 2022 Brookfield Properties study of 3,250 UK workers, suggests “interesting and visually striking” work is what employees say supports wellbeing. Practically, letting teams choose from a curated set matters more than the curator’s taste.
Do remote and hybrid teams get any benefit?
Exposure effects require a shared space, so a half-empty office dilutes them. Art-making and arts training do not: Drexel’s cortisol result came from individual activity, and virtual observation or improvisation sessions are routinely delivered online. For hybrid companies the practical shift is from decorating the building to funding activities, which also makes outcomes easier to measure per participant.
How long before any effect shows up?
Acute effects are immediate: the Exeter productivity gains appeared within a single task session and the Drexel cortisol drop within 45 minutes. Engagement and absence effects, if they occur, take a quarter or more to show in company data, which is why pilots shorter than three months rarely produce anything a CFO will accept. Plan for a 90-day minimum with a comparison team.
Discover more from Kurums | Business Intelligence
Subscribe to get the latest posts sent to your email.
