A pair of perfume researchers in England have lost a 2019 paper on what makes a scent appealing because, ahem, something about the data didn’t smell quite right.
The article was titled “Social success of perfumes,” and it appeared in July in PLOS ONE. There was a press release and a university writeup about the paper — but not, we should note, about the retraction.
The authors were Vaiva Vasiliauskaite and Tim S. Evans, of the Theoretical Physics Group and Centre for Complexity Science at Imperial College London.
The abstract of the study stated that:
We study data on perfumes and their odour descriptors—notes—to understand how note compositions, called accords, influence successful fragrance formulas. We obtain accords which tend to be present in perfumes that receive significantly more customer ratings. Our findings show that the most popular notes and the most over-represented accords are different to those that have the strongest effect to the perfume ratings. We also used network centrality to understand which notes have the highest potential to enhance note compositions. We find that large degree notes, such as musk and vanilla as well as generically-named notes, e.g. floral notes, are amongst the notes that enhance accords the most. This work presents a framework which would be a timely tool for perfumers to explore a multidimensional space of scent compositions.
But according to the retraction notice:
After this article [1] was published, questions were raised about the dataset used in the study. In following up on these questions it came to light that the dataset was obtained from a third-party commercial entity whose identity cannot be shared due to a nondisclosure agreement, and that the authors cannot share the raw data or provide clarifications about how the data were collected or processed. The authors posted anonymized summary data on Figshare, as noted in the article’s Data Availability Statement. However, the reported Methods are not sufficient to enable other researchers to reproduce the study and the data provided do not meet PLOS ONE’s requirements as outlined in our Data Availability policy. The authors noted that they cannot reproduce the analyses using another public dataset as no comparable dataset is currently available.
In light of these issues, the PLOS ONE Editors retract this article due to concerns about the reproducibility of the study and noncompliance with the journal’s Data Availability policy. We regret that these issues were not identified prior to the article’s publication.
VV and TSE agreed with retraction.
The disclosure statement on the paper says that “V.V. acknowledges support from EPSRC, grant number EP-R512540-1,” but says nothing about the source of the data. Indeed, all we could find about that was this line:
We have information on 1047 different notes present in 10,599 perfumes.
Vasiliauskaite, a PhD student at Imperial College London and the corresponding author, told us:
the data is owned by a third party and we had to agree to very tight restrictions in order to use the data. For instance we no longer have access to the original data. So we were very aware of the restrictions when writing the paper. As we want to be as open as possible, we made as much of the data available as we could and this has always been accessible in the repository listed in the references. This was explained to the referees and to the journal before publication. The journal reviewed the situation again after publication and at that point decided the paper did not comply with their open data policy.
Vasiliauskaite added that she and Evans aren’t planning to repeat their study:
I think it would be straightforward to perform the same type of analysis as discussed in the paper (and to my eyes the methodology is the most interesting part). However, I am not aware of a free open-source dataset of such nature available, thus we do not plan to attempt the same analysis again.
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up for an email every time there’s a new post (look for the “follow” button at the lower right part of your screen), or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
Good to see a publisher actually enforce their data availability policy rather than just give it lip lip service, as do many journals. Still, one wonders why this is done after publication, rather than in the review process.
When I have reviewed for journals recently, the instructions for reviewers have not outlined the journal’s open data policy. I think they need to do so. It is too much to ask reviewers to research the journal’s policies; the journal should state them up front.