Approval for Genetics of Personality Type

I’m posting a review request for the Genetics of Personality Type project which is applying for project approval. The project is part of a research study led by Dr. Denise Cook. The whole documentation of the study’s IRB approval is available here!

Should this project be visible and available for all Open Humans members to join?

Please vote Approve or Deny , and/or comment.

Quick links

Project info

  • Title : Genetics of Personality Type
  • Managed by: Dr. Denise Cook
  • Description: Current academic research looking for personality genes measure personality traits using questionnaires for the “Big 5”: Extraversion, Openness, Conscientiousness, Agreebleness, Neuroticism, traits that have been found to vary continuously in the human population. While these studies have successfully identified genetic variants associated with each of these traits, each polymorphism has only a small effect on the phenotype and explains only a tiny fraction of the genetic variance in each trait. Consequently, none of the SNPs or combinations of SNPs identified so far can predict someone’s personality traits. Instead of using the Big 5, this study will use the Myers-Briggs Type Indicator, which categorizes people into 16 different groups that share similar behavioural tendencies, and we hypothesize, similar patterns of genetic variants as well. The goal is to use this framework to isolate genetic variants that are strongly correlated with personality type.
  • Project website:
  • Connections: 13 members
  • Data received: 23andMe Upload, AncestryDNA Upload, FamilyTreeDNA integration, Imputer
  • Data added: This project will return a person’s personality type (i.e., INTJ, ESFJ) as determined using three personality questionnaires.

Approve - Nothing about the IRB or project are raising any flags for concern about the project content.

As a note: I sent their contact email a message outlining minor website server configuration improvements so https is more reliably accessed. Such as additional redirects and addition of HSTS header. Trivial to fix.

1 Like

Do Not Approve - The Myers Briggs is considered pretty much junk pop psychology these days. I can’t see this as doing anything but providing data to enable a commercial site to delivery data of questionable value. The authors are specifically rejecting the generally accepted instrument of personality measurement, the “Big 5”. This is akin to saying that they are choosing not to test for an infection with a white blood cell count, but instead measure the relative amounts of the 4 bodily humors.

1 Like

I just wanted to weigh in here as well, summarizing my response on the #project-approval slack channel.

I want to disagree on this being a ground for rejecting a project. My personal feeling is that Myers-Briggs isn’t a great scientific instrument because of its poor reproducibility as far as I can tell. Nevertheless I feel it’s a valid research goal to see whether it’s possible to find genetic association to those types.

I would agree that this might be something to keep in mind if this was some medical intervention with potential harm to participants for cases where we feel that the proposed alternative intervention seems useless (to pick up on your white blood cell count-example from the discourse thread). In this specific case I don’t see a potential harm for participants. And it seems that the IRB that approved the study feels the same way (this is not to say that IRB approval means that we can’t have our own review and views).

Generally I feel that Open Humans as a piece of infrastructure/platform should try to remain a rather neutral point of view with respect to the scientific utility of projects (within the limits outlined above).

In Slack there was also the question on whether the research that comes out of this project could be used to justify commercial services: I’m not sure whether we should evaluate this point during the project approval process. I don’t think we have judged other projects on their potential that researchers might commercialize their findings? And I don’t think there’s a risk of misuse of data in this case, as the whole project documentation (Terms of Service, IRB approval) is really good and I think even an example of how other projects can approach it.

And to be fully transparent about a potential conflict of interest: The project discussed here is partially funded by one of the Open Humans project grants. The decision to give this grant was made back in February by @madprime and me, no other Open Humans staff or board members were involved.


I’ve discussed the MBTI more with my psychologist wife, and she says it has some limited value. Her questions were: 1) If lots of people are already doing research using the Big 5, what’s the advantage of trying the MBTI, other than that it hasn’t been done? 2) Will there be published research out of this, or is it just to fuel a commercial enterprise? 3) Are there personality psychologists involved in the study? 4) What statistical methods will be used, especially given what might be a small sample size. Will this be a GWAS?

1 Like

COI up front: because OHF has supported this project with funding, I’d rather see a consensus from people other than me and @gedankenstuecke. (This project was awarded funding to connect to Open Humans; prior to this it was operating in Gencove.)

Currently “scientific merit” is not part of the considerations in the Project Review Guidebut that guide was a first pass, we can update it if people think it can be improved?

My take: I’d rather have fairly minimal barriers for projects beyond stuff like safety and “informed decisions”. There are things that are popular and questionable – MBTI and genome association efforts both! – but that’s the sort of thing a lot of people like to do… and I want to enable projects.

My hope is that people (project creators and community) learn by doing, inspired by each other. Thus – if someone is dissatisfied with something like MBTI, maybe they’ll be inspired to add big 5 profiling. :wink: I’m also taken by Denise’s openness about MBTI vs Big 5 on her about page. My impression is that this project is in large part a labor of love (as are so many!), even if there’s that potential to commercialize.

@blackbearnh any chance you might feel differently, on reflection?

– Mad, INFP.

Or INTP. Or something else. See, I got different answers – I like how Denise’s study has people to use at least two different tools, teaching us firsthand that they’re inconsistent!

@blackbearnh seems we posted about the same time, oops! I think it’s best for Denise to answer directly… but your #4 gets at an arguably more substantial “scientific merit” issue – but this community loves genetics projects…!

I love to see things grow, not die, e.g. if Denise shifted the interest into other correlations… and maybe she can share more about what’s already been done with Gencove. :slight_smile:

@blackbearnh @madprime I will add more to this as time permits. As mad mentioned, this is mostly a passion-project with the imagined best-case-scenario goal to be able to introduce a DNA-based personality test that can predict a person’s personality type. But that’s a lofty goal, especially given current Big 5 GWAS results and the tiny tiny tiny effect sizes of each individual SNP.

  1. I actually think that one of the biggest complaints of the MBTI, even amongst MBTI enthusiasts-- that people are categorized into “types” rather than given a percentage on some continuum-- is one of its strengths for something like a genetic study. Since the 4 dimensions (thinking-feeling, sensing-intuition, introversion-extraversion, judging-perceiving) are meant to be independent, you can group people by one dimension ("introversion vs extraversion) and run a case-control GWAS on that data.

The MBTI is meant to get at someone’s innate, preferred, way of perceiving the world and making decisions on it. Now, lots of people deviate from their preferred way of doing things, and being in certain environments can definitely alter our preferred way of doing things (to the point that after a while you don’t even notice you have “adapted” to your surroundings). Animals/plants do this all the time as well. But I think that one of the most beneficial concepts from MBTI is that when we are “adapting” there is a stress on the body/nervous system that can eventually wear a person down (in animal literature this would be measured as lower fecundity, lower ability to maintain resources, less survival rates). In human literature this would be known as “burnout”. In MBTI speak this is known as “Falsification of Type”. We see this a lot in humans that are disengaged from their work, low energy, etc. And then you find people that are at the top of their games or “self-actualized”. What are the genetic and environmental conditions that lead to these two alternative outcomes (assuming average, non-traumatic upbringing)?

Getting back to the “this” or “that” behaviors: in the animal literature you have males of the same species that prefer to guard the nest and males that prefer to venture out/hunt/collect food. Both come with a series of risk and rewards and obviously under certain conditions the same male could show both behaviors. But the idea is that on average, the male shows predominantly one or the other behavior (the strategies oppose one another). There are many more examples of these divergent “personality traits” or “behavioral syndromes” in animals. Although I concede that in the animal literature most of the talk is about these traits being on continuums as well…

So I agree that the MBTI test itself (and all the clones out there) is not, by itself, a very accurate way of getting at someone’s “Personality Type” (assuming such a thing even exists). However, this is more due to the fact 1) personality is inherently difficult to measure 2) people are not very self-aware/they lie 3) people tend to compare themselves to other’s in the immediate surroundings, so a person’s definition of extravert or introvert can vary widely 4) people don’t always have all the information necessary to make accurate decisions about themselves (either because they’ve never encountered certain situations or because they tend to see only the good/bad side of themselves, etc), and not really a flaw of Jung’s system itself… The MBTI is actually always supposed to be accompanied by a session with a trained MBTI specialist, for these exact reasons (they would ask questions to probe for natural responses/behaviors). Now this would be impossible to do for a genetic study and I actually switched from the MBTI to a clone for financial reasons and other more technical issues (I’ll correct this in the description).

To summarize, I think the theory of personality types has some great potential, if only as a way to bring together very disparate fields (humanities: ie, search for meaning/self-actualization/human needs/religion), the soft sciences (psychology, economics/decision-making, social psychology, sociology, etc) and more hard core sciences (behavioral genetics, neuroscience, animal behavior, etc). I think one of the first steps towards getting personality types to be recognized as a legitimate thing to study by academics would be some genetic support for their existence.

  1. Goal is to publish.

  2. I have Dr. Dario Nardi as a collaborator, but he is more an MBTI enthusiasts and runs EEG studies on personality type. If you know anyone interested please let me know, since I have been mostly shunned by academics by email and on twitter, lol. I am not a personality psychologist or psychologist, this is an interest I developed during my postdoc which was hard-core single-cell electrophysiology in rat brains (studying neuronal responses to noradrenaline in the prefrontal cortex; which is kind of related to personality/decision-making in a circuitous way; I believe there are lots of fMRI studies showing how people with different genetic variants in noradrenaline receptors respond differently to certain tasks).

  3. It’s a GWAS. I have a very small sample set of 340 people from the previous iteration of this project on Gencove (ran for 8 weeks) that is analyzed using PLINK all the way to the association data… there are hits at 1 x 10-7/1 x 10-6 which is supposed to be in the interesting range. Now I’m just not sure what to do with it. If there are any GWAS specialists interested in taking a look at the data/helping me figure out fact from fiction I am looking. All data is housed on DNANexus/can’t be copied to personal computers/etc.

My power analysis suggested I would need 200 people/type, so about 3000-6000 people in total. Any more than that and personality type will probably not be easily predictable (perhaps with polygenic analysis but that is above my understanding at the moment). Goal is to find a pattern of SNPs that could predict a person’s personality type.

The big issue I have with my previous dataset is that it was HEAVILY biased towards INTPs/INTJs/ENFPs/ENTPs and introverts in general. So would probably need help recruiting outside of the internet.

Ok that was more than I intended to write. If there are more questions I am happy to answer.


To be blunt: I don’t think you’re going to find much, with a relatively small sample size and likely low effect power. GWAS needs huge sample sizes. If you had a target gene, you might have better luck with a few thousand people. And I’m also concerned about not being able to interest traditional researchers in your ideas (being a rebel/outsider isn’t necessarily a bad thing, but when folks are shaking their heads and walking away, it’s at least incumbent on you to explain why they’re wrong (or, prove them wrong, which I guess this is what this could theoretically do.))

That said, I don’t think there’s any harm to be done in letting this run, save wasting your time. If you want to proceed with this, I have no objection.


Thank you. I believe this is what Open Humans was meant to do, help facilitate research that may not be possible within traditional academic boundaries and thus excluded from traditional funding sources.

I think you raise a good point – how does one interest researchers in something that they have already closed their minds to? My original plan was to start writing about Jung’s theories, MBTI, how it connects to all these different fields of research, how it explains so much, blah blah blah and in that way start convincing people… then I was like “well, without any scientific backing no one is going to take me seriously”. So this research project was born. And a small study may be sufficient to get people interested…

@madprime @gedankenstuecke any way to reach out to others to seek approval and/or how long does the review period usually last?

1 Like

@dencook I’ll try to push the thread once again on our Slack and see whether there’s someone else who wants to weigh in. If there’s no more feedback I’d say we can go ahead and approve next week some time!

Approve. The IRB and terms of service are clear. I think members would appreciate the intention/objective to publish, clearly stated, on the landing page. Links to reference material before signing up for the study would also be helpful for those not familiar with the research which this study will build upon.

I was left with open questions about the members’ reports:
Will members reports’ be from the personality tests only (no genetic component)?
Will members receive an updated, genetic association report once the study is published?
For all their shortcomings, I think 23andMe does a good job of conveying important points in their health reports (intended uses, limitations, non-genetic factors…). If this study plans on issuing a genetic association report, does it plan to addresses similar points?

Thank you! I can definitely make it clearer on the landing page that the intention is to publish and add some links to some studies.


  1. The reports are just personality reports for now, however, as mentioned in the IRB, I do have approval to report any associations that are found as the study progresses. I.e., we found that you are genetically more similar to the introverts in this study than extraverts, based on X, Y, Z.
  2. Yes.
  3. A 23andMe style report would be the ultimate goal, yes!

Thanks for the contributions to the discussion to @blackbearnh @wolfgang8741 and @arvkevi! And thanks for being so responsive to @dencook! I think the discussion was really useful one that provided a great opportunity to improve some of the documentation etc.!

Given that we seem to have reached a consensus for approval, I’ve marked the project as approved.

1 Like