So, you’re reading a research paper and come across an iffy looking piece of data. What to do? Maybe email the journal? What if the author finds out and trashes your next grant proposal? What if someone gets fired? The key to all these issues is ANONYMITY. The following guide lays out basic practices of how to report suspected scientific research fraud…
Is it fraud?
The first step is to establish whether what you think is fraud, actually is fraud. There are several excellent resources available, such as the US Office of Research Integrity’s forensic tools. A simple work-flow is as follows:
- Get the PDF of the paper and, if possible, high-resolution copies of the Figures. If no PDF is available, save the Figures (usually .JPG) from the HTML version of the paper.
- Use the snapshot tool in Adobe Acrobat to capture the Figure (or the relevant piece of it), and paste into Microsoft PowerPoint. Note that the resolution of the captured image will vary depending on how far the PDF is zoomed in. Alternatively in Windows use the Snipping Tool (Start menu > All Programs > Accesories) or the Prt-Scr key to grab a screenshot of the Figure.
- If looking for seams in images, adjust the brightness/contrast settings of the pasted image using the Picture toolbar in PowerPoint. On an LCD monitor, tilting the screen vertically can sometimes reveal subtle break-lines in images.
- If looking for similarities between 2 images, crop and size each image so they can be placed adjacent to each other. If necessary use the flip/rotate tools to orient the images to reveal duplications.
- Is it fraud? For this you should consult the various guidelines on image integrity for what is acceptable. The Committee on Publication Ethics, and the International Committee of Medical Journal Editors both have guides. Also this editorial from the Journal of Cell Biology is viewed as a landmark statement. In deciding whether something is a genuine mistake or fraud with intent to deceive, you should consider the intent of the authors… For example if an image is duplicated in its intact form, maybe it was just a labeling or copy/pasting error. However, if an image is duplicated but has been flipped/rotated/stretched or altered in any way, this is more indicative of intent to deceive.
What files to submit?
Once you’ve compiled your evidence file, there are many important considerations to maintain integrity and anonymity in materials you submit:
- Notate the images, indicating the source of the Figure, describing exactly what you think has been manipulated. Use a new page for each Figure. Be sure to cite properly, using PubMed ID #s if possible. Never include hyper-links, as these may contain identifiers that could reveal the origin of the file.
- Save the entire file as a PDF or other compact format. This is especially important when submitting large evidence files for which the originals may be several megabytes in size.
- REMOVE META-DATA. I cannot stress strongly enough how important this is! Every file you generate on your computer will contain unique information which could be used to determine your identity. The simple way to remove this in Windows… Right Click > Properties > Details Tab > Remove Properties and Personal Information link.
- If possible, submit original PDFs of the suspected manuscripts, again stripped of meta-data. Be aware that journal PDFs often contain “Downloaded by name on date” text in the margins.
- Draft a cover letter outlining your case, and describing the fraudulent activity and any important background information (e.g., how did you come across the fraud?). Be as professional as possible and include any conflict of interest information.
Where to Submit?
So, after many sleepless nights you have a lovely meta-data free PDF waiting to go. Where to send it? There are several options, and you should probably pursue more than one…
- Send it here. I will do my best to publish the findings in a timely manner, but please be patient. I have a real job and this site is a part-time hobby (a pretty dangerous one too). I cannot guarantee to publish everything.
- Notify the journal(s) involved. The editor-in-chief is a good start. Do not use threatening language, but by all means feel free to let the journal know you have made the evidence available to this website (it might nudge them to act).
- Notify the appropriate regulatory bodies…
- Be careful to frame your accusations as suspected fraud. There may be a perfectly innocent explanation for what you think is fraud. We are lucky to live in a society that counts “Ei incumbit probatio qui dicit, non qui negat” (innocent until proven guilty) among its founding values.
Note… YOU MUST REQUEST ANONYMITY when submitting information. It seems counterintuitive, but some of the above-mentioned bodies will simply pass on your email to the authors or their institution! Anonymity is not the default setting, as I once found out the hard way.
How to Submit?
I have no interest whatsoever in knowing who you are, and it’s probably best if the authorities don’t know either. Thus, it is advisable to register a new gmail account and use it for all communications regarding research fraud. Choose a sensible user name (i.e., not one including your initials) and be sure to use a strong password. Please try to limit PDFs to less than 5MB when sending here. PLEASE indicate in your email how you would like the tip attibuted. i.e., do you want to be named? If you do not say, I will just use whatever name is on the email.
**It is strongly advised to pick a user name/email for this website that you will not use anywhere else. I have had a couple of requests recently for attributions to be removed by the people who reported misconduct. This can only mean one thing – those user names or email addresses got compromised because they were being used for something else (e.g. commenting on another website). You should also be aware (doh!) that when you make a comment to this site, your IP address is logged. While this doesn’t give away your ID, a simple reverse IP lookup does tell me where in the world you are from, sometimes even revealing the University you may be posting from. For this reason, again, please use DIFFERENT IDENTITIES for commenting here, sending information here, and for all your other InterNet business.
What happens next?
- If you submit here, as mentioned above I will do my best to publish the findings in a timely manner, but please be patient. I have a real job, and this site is a part-time hobby.
- If you notify the journals, do not expect a reply to your anonymous gmail account. Many journals have policies of not responding to such whistleblowers, and will often demand you identify yourself. All I can advise here is persistence, and a gentle reminder that we live in an age of social media
- If you notify the regulatory authorities (e.g., ORI), they often do not actually conduct an investigation. They simply pass the accusations to the author’s institution, and provide oversight to an investigation by the local ethics office. The findings of the local office may then be used in compiling the ORI’s final report, which often takes several years to emerge. Do not expect ORI or the institution to be responsive to inquiries at any point during the investigation.
- Sanctions are highly variable, but usually can bans from journals, from applying for federal funding, and from serving in an advisory capacity (e.g., grant review panel).
- Naturally, if the data in a paper are eventually proven to be fraudulent, a retraction should be forthcoming, although increasingly corrigenda or errata are used by journals to save face (i.e., to avoid admitting their peer-review system failed).