Retracted scientific research should be treated with caution. Often its conclusions are at least partially flawed. Yet many retracted papers continue to be referenced by scientists for years after being withdrawn. Why? Because retractions are not always clearly marked. RetractOMatic highlights retracted titles in Google Scholar and Web of Science searches. When RetractOMatic is confident that a paper is retracted it turns the text red, while lower confidence matches turn orange. It's a small first step toward notifying scientists of withdrawn papers when browsing the web. Try it out! Fortunately retracted papers are pretty rare in the wild, but if you've just downloaded RetractOMatic, you can test it by searching for one of the top 5 most cited retracted papers (according to retractionwatch.com): 1. Visfatin: A protein secreted by visceral fat that mimics the effects of insulin 2. Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children 3. An enhanced transient expression system in plants based on suppression of gene silencing by the p19 protein of tomato bushy stunt virus 4. Purification and ex vivo expansion of postnatal human marrow mesodermal progenitor cells 5. Viral pathogenicity determinants are suppressors of transgene silencing in Nicotiana benthamiana Of course, not all retracted science is completely wrong! Papers are withdrawn for all kinds of reasons, including licensing issues, duplication, and minor honest mistakes. By making you aware of retractions, RetractOMatic simply allows you to use your own good judgement when deciding what work to build on. FAQ: Where do you source your list of retracted papers from? Since there's currently no definitive database of retractions, RetractOMatic matches titles that are marked as "Retracted Publication" in PubMed. Version 1.0 uses a snapshot taken in March 2016, and we'll be updating the RetractOMatic title list regularly. PubMed's retraction list is of course only part of the picture. If you have a suggestion for additional sources of retraction data, please do get in touch. What's the difference between a high confidence match (red) and a low confidence match (orange)? A title is colored red only if it completely matches a title in the retracted paper list. If a Google Scholar title is truncated (i.e. it ends in "...") but the visible part of the title matches a retraction, we color it orange to indicate a lower confidence match. After all, we can't be sure what the full article title is. Any title containing the word "retracted" or "retraction" is also highlighted in orange. This helps to make existing retraction notices more obvious, but we still count the matches as low confidence, since the paper might simply be about retraction. Why is this retracted paper not highlighted? There are a number of reasons why a retracted paper might not be highlighted: - It's not labelled as retracted in the PubMed database - The retraction is recent and will show up in the next update - The paper was returned by a Google Scholar search and begins with "...". RectractOMatic does not currently support titles truncated at the beginning. - There's a typo in the retraction notice (we're looking at you "retration!") - There's a subtle difference between how the title is displayed in the search results, and how it appears in our retraction list. We've tried to handle many minor differences of this kind already. - Or, there's a genuine bug in our code :-( Why is this paper that isn't retracted highlighted? The only time this should happen is if the paper title contains the word "retraction" or "retracted", in which case RetractOMatic will color it orange as a low confidence match. We do this to make retraction notices stand out more, but no Chrome extension can replace your own good judgement! So please remember that low confidence matches are just that and do some digging. If you see a high confidence match that is incorrect, please get in touch so we can investigate further. Why does RetractOMatic require these permissions? The permissions we ask for sound extensive, but they're actually the minimum needed to be able to check the paper titles shown in your browser and change the appearance of some of them. RetractOMatic does not run on any websites other than Google Scholar and Web of Science searches, and we gather no data about our users' browsing habits! The code for RetractOMatic has not been obfuscated in any way, so please go ahead and take a look at how it works.