Ethics of publishing privately shared data
January 30, 2023 9:01 AM Subscribe
If I share human behavioral data with a rival lab, what are the expectations surrounding their use of that data?
I shared behavioral data collected in my lab with another research group. This is not publicly available data - it is available by request only. Since this is a rival lab, my expectation was that they were attempting to find a flaw in our analyses in order to force a retraction of a publication critical of their theoretical framework.
Some two years later, that lab is attempting to publish a manuscript based entirely on the data I shared. The good news is that our analyses replicated perfectly. However, I never explicitly granted permission for them to use the data for publication, although I did not explicitly request they not do so.
My sense is that this is unethical conduct on the part of the other lab, and I'm trying to find information about whether or not this is the case. The APA's ethics codes (pdf) seems to support this view (8.14(b): "Psychologists who request data from other psychologists to verify the substantive claims through reanalysis may use shared data only for the declared purpose. Requesting psychologists obtain prior written agreement for all other uses of the data.") However, I don't think the APA code applies to researchers who are not APA members, so I'm trying to find codes of conduct that do - either institution-specific or discipline-wide.
Unfortunately, most resources i've been able to find are mostly about how great it is to make your data freely available online for everyone to do with as they see fit. My interest is more in finding information regarding what rights are implicitly retained when sharing research data.
I shared behavioral data collected in my lab with another research group. This is not publicly available data - it is available by request only. Since this is a rival lab, my expectation was that they were attempting to find a flaw in our analyses in order to force a retraction of a publication critical of their theoretical framework.
Some two years later, that lab is attempting to publish a manuscript based entirely on the data I shared. The good news is that our analyses replicated perfectly. However, I never explicitly granted permission for them to use the data for publication, although I did not explicitly request they not do so.
My sense is that this is unethical conduct on the part of the other lab, and I'm trying to find information about whether or not this is the case. The APA's ethics codes (pdf) seems to support this view (8.14(b): "Psychologists who request data from other psychologists to verify the substantive claims through reanalysis may use shared data only for the declared purpose. Requesting psychologists obtain prior written agreement for all other uses of the data.") However, I don't think the APA code applies to researchers who are not APA members, so I'm trying to find codes of conduct that do - either institution-specific or discipline-wide.
Unfortunately, most resources i've been able to find are mostly about how great it is to make your data freely available online for everyone to do with as they see fit. My interest is more in finding information regarding what rights are implicitly retained when sharing research data.
Questions I think have some relevance to the answer here:
- Are you contractually obligated by a funding agency to make the data available to others for secondary use?
- Did your institutions sign a data use agreement?
In my field, this would not be a breach absent more information, and indeed this type of data reuse is an explicit goal of our funders. Breaches would be things like: attempting to reidentify research participants, or resharing the person-level data beyond their own group, including posting it in a public place.
posted by eirias at 9:58 AM on January 30, 2023 [6 favorites]
- Are you contractually obligated by a funding agency to make the data available to others for secondary use?
- Did your institutions sign a data use agreement?
In my field, this would not be a breach absent more information, and indeed this type of data reuse is an explicit goal of our funders. Breaches would be things like: attempting to reidentify research participants, or resharing the person-level data beyond their own group, including posting it in a public place.
posted by eirias at 9:58 AM on January 30, 2023 [6 favorites]
If the answers to eirias's excellent questions are "no" and "no:"
I see two ethical issues here: the human-subjects research ethics of this secondary publication (assuming, of course, that we're talking about human behavior here), and the publication ethics of publishing out of another lab's data.
I consider the first issue considerably more serious; protecting research subjects' dignity, privacy, and autonomy is vastly more important than a possible scooping. What exactly did the research subjects agree to when they participated in your lab's research? Was the rival lab made aware of this agreement? Did they violate it -- or did you by disclosing the data, or both? (Plus a crapton of additional data-management issues around data de/reidentification, data security, and similar that I don't have enough information from the question to opine about.)
The rival lab's IRB may or may not have had a chance to opine, and if it did have a chance, may or may not have taken it. IRBs tend to focus hard on data collection to the exclusion of any other source of harm to research subjects -- the rival lab's IRB may well have said "existing data? EXEMPT!" I've seen it happen. That doesn't mean the ethical issues here magically went away; it just means the IRB abdicated any responsibility to address them. As many IRBs do.
The second issue is why, when I teach research-data management, I lean hard on the idea that non-open data sharing needs to come with written agreements. "Implicitly-retained rights" are not really a thing; gotta use your words.
That said, I doubt you are without recourse, especially since you have a prior publication from the data. You can contact the editors of the journal where the rival lab published and request a retraction on grounds of misappropriated data. The Committee on Publication Ethics has a handy flowchart for how it thinks editors should resolve a similar situation. Your case is stronger if the rival lab did not attribute the data to you in its publication.
In future: written agreement, written agreement, written agreement before another researcher sees so much as a single datapoint.
posted by humbug at 10:02 AM on January 30, 2023 [10 favorites]
I see two ethical issues here: the human-subjects research ethics of this secondary publication (assuming, of course, that we're talking about human behavior here), and the publication ethics of publishing out of another lab's data.
I consider the first issue considerably more serious; protecting research subjects' dignity, privacy, and autonomy is vastly more important than a possible scooping. What exactly did the research subjects agree to when they participated in your lab's research? Was the rival lab made aware of this agreement? Did they violate it -- or did you by disclosing the data, or both? (Plus a crapton of additional data-management issues around data de/reidentification, data security, and similar that I don't have enough information from the question to opine about.)
The rival lab's IRB may or may not have had a chance to opine, and if it did have a chance, may or may not have taken it. IRBs tend to focus hard on data collection to the exclusion of any other source of harm to research subjects -- the rival lab's IRB may well have said "existing data? EXEMPT!" I've seen it happen. That doesn't mean the ethical issues here magically went away; it just means the IRB abdicated any responsibility to address them. As many IRBs do.
The second issue is why, when I teach research-data management, I lean hard on the idea that non-open data sharing needs to come with written agreements. "Implicitly-retained rights" are not really a thing; gotta use your words.
That said, I doubt you are without recourse, especially since you have a prior publication from the data. You can contact the editors of the journal where the rival lab published and request a retraction on grounds of misappropriated data. The Committee on Publication Ethics has a handy flowchart for how it thinks editors should resolve a similar situation. Your case is stronger if the rival lab did not attribute the data to you in its publication.
In future: written agreement, written agreement, written agreement before another researcher sees so much as a single datapoint.
posted by humbug at 10:02 AM on January 30, 2023 [10 favorites]
If your data was collected with federal funding, there is probably an expectation written into the grant that you will make your data public(with the exception of privacy constraints for human subjects) alongside your publication. Open data policies are increasingly common with private funding organizations as well.
The days of hoarding data just in case you can wring one more paper out of it (my old advisor had 40 years worth of jealously guarded data) are coming to an end.
posted by rockindata at 10:04 AM on January 30, 2023 [1 favorite]
The days of hoarding data just in case you can wring one more paper out of it (my old advisor had 40 years worth of jealously guarded data) are coming to an end.
posted by rockindata at 10:04 AM on January 30, 2023 [1 favorite]
Yep, your first step is to figure out what your contracts say. The funding contract and the data use agreement, if you had one. (If you didn't, that's part of the lesson here - if you want to put restrictions around how data you share can be used, you need a DUA.). Also take a look at the requirements of the journal you published for what obligations you have to share data and what restrictions may exist. (And yes, there could be IRB issues - but that's much less likely if you deidentified the data appropriately before sharing it.)
Broadly, in many fields, this is the way the world is going. There's nothing inherently unethical about this, if done with appropriate attribution. But the details always matter and there could be something about your specific data set, contractual obligations, etc. that make this a problem you have some recourse for.
posted by Stacey at 10:21 AM on January 30, 2023 [1 favorite]
Broadly, in many fields, this is the way the world is going. There's nothing inherently unethical about this, if done with appropriate attribution. But the details always matter and there could be something about your specific data set, contractual obligations, etc. that make this a problem you have some recourse for.
posted by Stacey at 10:21 AM on January 30, 2023 [1 favorite]
In regards to research ethics, my opinion (as a PI and an AE) is: gating data is unethical and a barrier to science, progress, and ultimately, human health.
Re-use of data, and repositories for enabling that, is common in my field. NIH sure as heck expects us to manage and share, and seeing as they paid for it, they get to make the policy.
How do they attribute or cite the data in this new work? You don't mention that, and it's crucial that they correctly attribute the data, of course.
What is your preferred outcome? Do you want authorship?
posted by Dashy at 10:57 AM on January 30, 2023 [8 favorites]
Re-use of data, and repositories for enabling that, is common in my field. NIH sure as heck expects us to manage and share, and seeing as they paid for it, they get to make the policy.
How do they attribute or cite the data in this new work? You don't mention that, and it's crucial that they correctly attribute the data, of course.
What is your preferred outcome? Do you want authorship?
posted by Dashy at 10:57 AM on January 30, 2023 [8 favorites]
« Older Conversation starters with introverted adult child... | How hard can it be to choose a bathroom vanity? Newer »
This thread is closed to new comments.
posted by twelve cent archie at 9:39 AM on January 30, 2023 [5 favorites]