How “phylogenetically conserved” are views (and study results) on controversial scientific questions?

Everybody’s familiar with the idea of an academic lineage–you’re the “academic offspring” of your PhD supervisor. And perhaps of other mentors who also had a big influence on you. So you can imagine arranging academics into a sort of “phylogenetic tree” tracking lines of academic descent.

Question: how “phylogenetically conserved” are scientists’ views on controversial scientific questions? To what extent do mentees tend to share the views of their former mentors on controversial matters?

I’m particularly interested in cases in which there’s high conservatism not just of opinions, but of results. Are there any cases in which members of one “academic lineage” tend to obtain one sort of result, and members of another “lineage” do similar studies but yet tend to obtain different results than members of the first lineage? The first possible case that comes to mind is fluctuating asymmetry. Rich Palmer (1999) did a meta-analysis of the literature which asked (among other things) whether empirical studies by certain prominent authors tended to obtain different results than studies by other authors.

Sociologists of science must have studied this. Anyone know of any references (he asked lazily)?

17 thoughts on “How “phylogenetically conserved” are views (and study results) on controversial scientific questions?

  1. How would you be able to tell inheritance of views from homophily of views (mentors attract people with similar views)? I know from my own career that I definitely did not seek out mentors with wildly different views from my own.

    I think you’d have to at least look at the postdoc stage (since there’s generally no easy way to measure someone’s views before they started grad school), and see how someone’s views changed before and after interacting with an important mentor.

    • Interesting question. But I think it’s useful to start with the first order question “Do people tend to hold the same opinions as, and obtain the same results as, their mentors?” before asking the follow up question “Why do they hold the same opinions as, and obtain the same results as, their mentors?”

      As to whether mentors attract students holding similar views on controversial matters, I don’t know but FWIW (not much) I doubt it. Isn’t it fairly rare for prospective grad students to have firmly established views on any scientific matter? I mean, is there really lots of variation among grad students in their views on (say) the dilution effect, with those who believe it’s general going to work in the labs of dilution effect advocates and those who don’t believe it’s general going to work in the labs of dilution effect opponents? In my experience, prospective grad students choose mentors not for their specific views on controversial matters, but because of some combination of (i) the uni at which they work, (ii) their fame, (iii) the species/study system in which they work, and (iv) the broad topic on which they work (but not their specific opinions on that topic, if it’s a controversial one).

      Ok, I’m sure there are scientific topics on which prospective grad students do have firm views. But aren’t those topics ones that are widely taught at the undergrad level, and on which most/all prospective ecology grad students have similar views? e.g., “Earth is in the midst of an extinction crisis”.

      A final thought: the difficulty of distinguishing inheritance vs. homophily of views is one reason why it’s useful to look at “phylogenetic conservation” of study results. Maybe some prospective grad students do seek out supervisors who share their views on controversial matters. But nobody enters grad school with their study results already in hand. So if people in one academic lineage tend to obtain one sort of result, and those in a different lineage do similar studies but yet tend to obtain different results, well, that’s potentially worrisome. It means that *something* you learned from your supervisor–your study system, your methods, your preferred way to operationalize some vague concept, the way you spend your “researcher degrees of freedom” or navigate Gelman’s “garden of forking paths”–is dictating your study results.

      • I agree with all your points here; I do suspect that homophily is a much stronger driver at the postdoc stage though. Which likely would actually magnify the effect of any phylogenetic of profs on students, assuming students are more likely to do postdocs with their advisor’s mentors.

  2. The question of whether regional biotas become saturated or not seems to generate answers that are predictable by who produces the paper.

    On the more general topic, it seems worth distinguishing two underlying reasons one might find conservatism (not the only two reasons, and not mutually exclusive, but anyway…):

    (1) Preferred conclusion drawn despite the data. This is obviously not good, and can be due to various conscious and unconscious biases during the process of analysis (inclusion/elimination of data; choice of test). The existence of confirmation bias seems fairly well established.

    (2) Choice of study system for which particular results seem likely. We all study what we find interesting, and the judgement of what’s interesting often includes what seems to be “going on” in the system. Advisors and students most often share a sense of what’s interesting. At the level of a scientific discipline this can be problematic, even if there are no biases of any kind in conducting the research itself.

    • Good point. And tricky to separate these in practice, since as you say they’re not mutually exclusive. I guess a big cross-system preregistered collaborative study could separate them in some cases. The preregistration prevents your (1) from coming into play, and conducting the study across systems gets at your (2).

      An interesting debate could be had about whether your (2) is problematic at the level of an entire discipline. I have mixed feelings myself.

  3. My broad sense is that there is a fairly high degree of heritability of perspectives and ideas and interests. It was about half way through grad school that I realized knowing where somebody fit in the tree (basically their advisers and grand advisers) explained a lot about what they were doing. I still find that true today A great deal clicks into place when you know the academic trees.

    That said I think results is a whole other question. Once you control for being interested in similar questions and approaching it similar ways (and as Mark notes often with similar study systems). I am not sure there is a strong proclivity to conform the results to ones academic heritage. I could give any number of examples even in macroecology where academic offspring got results that contradicted their current or former supervisor and went ahead with that interpretation.

    • Yes, at the broad-brush level I’m sure there’s high heritability of perspectives/ideas/interests, due both to students choosing to work with supervisors who share their interests, and learning from their supervisors. I think that’s inevitable and fine. I’m thinking of heritability of specific controversial opinions, and of results.

      Off the top of your head, what are some of the macroecological examples of academic offspring publishing results contradicting their current or former supervisors?

  4. Agreed that students are unlikely to pick their advisers for the specific views they hold on a specific topic, but I surely chose my supervisors because I admired the way they think, which I think is closely linked. It seems to me that people often draw different conclusions (even with the exact same data) depending on whether they focus on, say, statistical significance vs effect size, aspects that are consistent with the hypothesis versus those that are inconsistent with it, residual variance after having accounted for known effects versus main effects, where one thinks the burden of proof lays (on demonstrating there is an effect vs on demonstrating that an effect is not there), etc. I admired and generally agreed with the views of my supervisors about these things, which further molded my own views. So, yes, I do think I hold views, and am likely to reach conclusions, close to those of my supervisors, mostly because we think in similar ways and focus on similar things when interpreting results. But, I certainly hold views that are different from my supervisors and would feel good about publishing results that contradict those of my supervisors (or my own!). My feeling is that this last point would unfortunately not be true in all labs… I suspect some students are implicitly discouraged from disagreeing with their supervisors.

  5. “Sociologists of science must have studied this.”

    From a variety of different perspectives for quite a while. I can’t think of anything on Advisor-Student Inertia off the top of my head, but it’s been put forward that biases in the membership of editorial boards can result in a similar problem (intellectual conservatism) and there is some evidence for nepotism on editorial boards, including “institutional nepotism”.

    The ideas go way back, but this is as good a place as any to start:

    Crane D. The gatekeepers of science: Some factors affecting the selection of articles for scientific journals. The American Sociologist. 1967;2(4):195–201.

    • Forward-citing this article, there seems to be a lot of excellent research in the general area of academic tribes and territories (at departmental, field of specialization, age cohort, etc levels). Skimming through this book (the first result on Google forward-citing) there isn’t anything that cries out Advisor-Student relationship, but it discusses supervisors at a few points.

      Becher, T., & Trowler, P. (2001). Academic tribes and territories: Intellectual enquiry and the culture of disciplines. McGraw-Hill Education (UK).

  6. Are there notable examples where students came to disagree with major ideas promoted by their former advisors? I’m sure there must be.

    • Oh sure. Brian noted that there are various examples in macroecology. And I vaguely recall a former student of David Tilman’s who went on to disagree with him about some things.

    • 🙂

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.