Research summary · Meta-analysis · 2017
Scent, music and colour in stores: a meta-analysis of 30 years of research
Pooling 66 studies with 15,621 participants, this meta-analysis found that stores and service settings with scent or music produced higher pleasure, satisfaction and intentions to buy or return than settings without them. The effects were small to medium, and scent’s effect on pleasure was stronger in samples with more women.
Reviewed by Simon Faure-Field · 8 October 2026 · Checked against the full text
At a glance
Key findings
- 66studies combined, covering 74 samples and 135 effects from 1982 to 2016.
- r = .18average effect of scent on customer satisfaction, the largest of its four scent effects.
- 2.5×larger effect of scent on pleasure in samples with more women (r = .18 against .07).
The study
Citation and study details
Citation: Roschk, H., Loureiro, S. M. C., & Breitsohl, J. (2017). Calibrating 30 years of experimental research: A meta-analysis of the atmospheric effects of music, scent, and color. Journal of Retailing, 93(2), 228–240.
Read the original: doi.org/10.1016/j.jretai.2016.10.001
- Study typeMeta-analysisPeer-reviewed, open access
- Sample66 studies, 15,621 participants135 effects, 1982–2016
- SettingRetail, service and onlineReal and simulated environments
- Evidence gradeD: review or meta-analysisQuantitative synthesis of experiments
Method
What the researchers did
The authors pooled the experiments they could find that tested music, scent or colour against a comparison condition.
The researchers searched the main business and psychology databases for experiments that manipulated music, scent or colour and measured how customers responded. They found 66 qualifying studies, 64 published in journals and 2 unpublished working papers, covering 74 independent samples and 15,621 people between 1982 and 2016. The median participant was 33 years old, and about 62% of participants were women.
For scent and music, each study compared a setting with the stimulus present against the same setting without it. Four outcomes were analysed: arousal, pleasure, satisfaction with the experience, and behavioural intentions such as intending to buy or return. The authors also tested whether results differed between students and real customers, between samples with more or fewer women, between retail, service and online settings, and between real and simulated environments.
Results
What they found
Scent had a positive effect on every outcome measured, with effects in the small-to-medium range.
Compared with unscented settings, scented settings produced higher satisfaction (r = .18), pleasure (r = .09), arousal (r = .08) and behavioural intentions (r = .05). The arousal result was less robust than the others. Music had similar effects on pleasure, satisfaction and intentions, but not on arousal. Warm colours raised arousal but lowered satisfaction compared with cool colours.
Scent’s effect on pleasure was significantly larger in samples with a higher proportion of women (r = .18) than in samples with fewer (r = .07).
Effects of scent and music tended to be stronger in service settings than in retail. For scent this difference was only marginal: the effect on intentions was r = .34 in service settings against .04 in retail, but it rested on just two service-setting results and did not reach conventional statistical significance.
Limitations
Limitations
A careful synthesis, limited by the studies available.
Results varied considerably between studies, and the authors could only partly explain why. Some comparisons, including scent in service settings, rested on very few studies, so the tests had little statistical power. Factors such as scent intensity and complexity could not be analysed, because too few studies measured them consistently.
The analysis combines real stores with simulated and laboratory settings. The authors found no significant difference between the two, but the individual studies vary widely in design and quality.
A note on the evidence
How strong is this evidence?
Strong: the most comprehensive synthesis of its kind when it was published.
By pooling three decades of experiments, the study turns a scattered and sometimes contradictory literature into average effect sizes. Its conclusion that scent reliably, if modestly, improves the shopping experience is more trustworthy than any single study. A later meta-analysis by Roschk and Hosseinpour (2020), focused on scent alone, builds on these findings.
For commercial spaces
What it means for commercial spaces
For a retailer or service business, the meta-analysis sets realistic expectations: a pleasant ambient scent is likely to improve how customers feel about the experience, with satisfaction showing the clearest benefit, but the effects are modest rather than dramatic. The authors advise matching scent to the setting and treating it as a long-term part of the environment rather than a short-term sales lever. The stronger effect on pleasure among women may matter for businesses whose customers are mostly female.
Scent Swirl supplies ambient scenting for retail spaces. See how scenting works in retail →
How to cite this summary
Scent Swirl. (2026). Summary of Roschk et al. (2017), Calibrating 30 years of experimental research: A meta-analysis of the atmospheric effects of music, scent, and color. Scent Research Library. https://scentswirl.com/research/roschk-2017-atmospherics-meta-analysis/ (last reviewed 8 October 2026 by Simon Faure-Field).
This summary has been checked against the full paper. If you spot an error or know of a study we should add, we would like to hear from you.
Keep reading