Our faculty search so far

Without getting into identifiable details, here’s what’s happening so far in a search I’m co-chairing for the Harvard FAS Center for Systems Biology, together with Sharad Ramanathan, with some thoughts on how this process works.

We got 182 applications. The applicants are remarkably strong, which is great. I thought that in an open application process we’d get more of a range of qualifications. Apparently people self-select pretty strongly.

Too strongly, in fact. I’m worried about how much self-selection is going on, in two different respects. One is that our applicant pool starts out biased: it’s only 21% female, and it’s only 5% underrepresented minority. The other is that it’s striking how many applicants are either at Harvard already, or have past Harvard training. I think both observations are telling us the same thing: that there are highly qualified people who don’t think they’ll be comfortable or welcome here. These are hardly novel observations about seeking fairness in faculty recruitment. It’s not just about how we treat our applications once we get them, or how we treat our faculty once they’re here; it’s also about getting qualified people to apply in the first place. I’m looking for ways to do better next time.

One thing I’m skeptical about now is whether it’s a good idea for me to advertise positions on Twitter. My Twitter followers are over 80% male. (I think this is an interesting and damning statistic. It’s not like I control who follows me, so where is the bias coming from? I follow about 50/50 female/male, myself.)

The first step is affectionately, brutally dubbed “triage”. Every application gets read, albeit quickly, by three of us. I spend about 10 minutes on each application. (You can do the math: it works out to about 30 hours, total.) First thing I look for is what the research question is. I write a one-line summary for myself, to force myself to extract the relevant info. If, after due effort, I can’t get it, then I’m worried about clarity and focus — either mine (if I’m tired and it’s time for a break), or the candidate’s.  Second I’m looking at publication history. I’m looking for evidence of substantial, original, creative work, and a trajectory that I can understand that leads up to the research proposal. I also check the titles and abstracts of the 1-3 papers/manuscripts that are submitted in the application package, so make sure that what the candidate identifies as their major contributions agrees with what I got from the CV, and I’ll dip into those papers to see if I get the points. Third, I skim reference letters, where (hidden amongst the superlatives) I’m looking for evidence that the candidate developed original ideas. Fourth, I’m looking for other rigorous selections that the candidate has passed before – graduation with honors as an undergrad, competitive postdoc fellowships, substantive research awards. No one thing is disqualifying by itself, especially because I’m on the lookout for unconventional people with unique superpowers, who are going to open up totally new niches.

What I’m not concerned with: journal titles, H-index, citation counts, or impact factors. I think a lot of the angsty gnashing of teeth about needing Nature/Science/Cell papers is self-inflicted by the candidates. I can only speak for myself, not how other searches work — yes, I know, I have been told that some universities do require you to list impact factors and citation counts and whatnot. But it sure seems to me that it’s more often the applicants themselves who are emphasizing names of journals, H-indexes, and impact factors, even as search committees like ours are trying (and are explicitly instructed) to read past all that crap. When I see someone saying “submitted to Science” on their CV, I’m thinking this person is maybe too concerned with playing the game. And if that submitted manuscript is on the CV but wasn’t provided to us as one of the papers we can actually read and evaluate, then I’m especially cross that you seem to think we’re more swayed by surface than substance.

We want to identify the most highly qualified candidates. The literature (from Jo Handelsman, Mahzarin Banaji, and others) is unequivocal: equally qualified candidates will be disadvantaged in selection processes when a person doesn’t fit my implicit internal template for what a successful scientist looks like. I used to think that I don’t have such biases. My wife is a scientist, I have two daughters, some of my best friends are… you’ve heard it, all the usual stuff people say in their own defense. Now I know I have implicit biases, because I’ve taken (unsuccessfully fought against) Banaji’s implicit bias tests. Now I take conscious steps to doublecheck my initial judgements. When I’m recommending people for seminars or talks, or nominating for prizes, I stop and make myself think of female or minority candidates, and then think whether they’re just as qualified as the first people who came to my mind. And often they are. So at all stages of our process, we read and rank the female and underrepresented minority applications separately; then we interleave our rankings as objectively as possible, forcing ourselves to be conscious to judge people on substantive grounds. I also track statistics on our pool, not because of quotas, but because it’s one way to keep an eye on whether the process is inadvertently falling prey to implicit bias.  My thinking about diversity totally changed when I realized this is not about quotas at all. It’s about the fact that my implicit bias works against equally qualified candidates, and I must work consciously against that bias.

This isn’t to say that all this striving is sufficient, either. There’s still plenty of questions about whether our academic culture is comfortable, supportive, and welcoming for all. Harvard’s child care options for faculty aren’t all that great, by all reports, and that’s a problem for scientists with children. And though I’m male and white, I come from a rural and non-affluent background, and even I sometimes feel like an uncomfortable outsider in Harvard’s pomp and circumstance, and I can only imagine what it’s like for others.

After triage, we assign each application to be read carefully by three randomly chosen search committee members. At this stage, we’re reading the research statement closely, and we’re also reading publications to some depth. Of course, it’s a compromise between the time we have versus the depth we desire. I aim to spend about an hour or so on each application. It’s not enough time to decide whether the papers are correct, but it is enough time to see how the candidate frames a question, what sort of approaches they take, and whether their writing is clear, engaging, and well-organized. I impose a discipline on myself, making sure I’ve read for substance: I write down a concise summary of the candidate’s major accomplishments and research aims.

Again, I don’t care where a paper was published, so long as it’s in a reasonably reputable place. Reputable, for me, includes the arXiv, PLOS ONE, and suchlike. I would rather see a substantive complete story in an excellent journal, as opposed to an overly telegraphic letter in Nature. I don’t want to get into specific examples that might become identifiable, but: yes, I do want to read your first-author arXiv paper, more than I want to read a hundred-authored consortium advertisement in Nature. The committee has instructions to pay attention to substance, to read your research summary and your publications in as much depth as time allows.

We have little illusion about how objective this process is — it’s just not. A lot of it comes down to intangibles, like whether people in the department get excited about a candidate’s research question.  When I applied for faculty positions, it was hard to deal with all the rejection letters. I took them all super personally, some more than others. (Looking at you, University of Vermont.) Now rejection letters will be going out on my watch, including letters to some of the candidates that I liked the best but couldn’t sway the rest of the committee, and of course I won’t say that.  Damn, we have a lot of superb applicants who we’re not going to invite. We’re going to look like dummies when they go off and be successful somewhere else. So it goes.

We’ll be making our six invitations for interviews in the coming week or so.

24 thoughts on “Our faculty search so far

  1. Thanks very much for sharing some insights into your selection process. I was wondering roughly how many of the 182 applications you found broadly competetive and how many you would say were at the same level as those selected for interviews? I also wondered whether you had considered other ways of selecting the candidates for the interviews?

    One method that has been tried elsewhere and received a lot of attention is to skype a larger pool of people and only then shortlist the best ones. There must have been many more than six excellent candidates and selection among those was presumably based on personal preferences of the search committee and maybe to some degree arbitrary. Looking at 20 or 30 candidates would maybe have helped to make a more informed choice. I would also argue that selection solely based on a written application misses some important information. Talking about science in a clear and engaging way is clearly an important aspect of the job, in particular at a teaching institution, but can’t be assessed from a written resume. While a 5-10 min online interview is of corse much too short to gain a lot of insight, I believe that it can bring accross some valuable additional information about the candidate. I have seen a number of recruitment symposia in my home institute and while the standard has generally been very high, there have been a few instances, where it was clear 5 minutes into the talk that this is not the one that will get an offer. I think in these instances a short pre-interview would probably have resulted in someone else getting the opportunity to present herself at the symposia.

    I also would like to mention that I applaud you for thinking so deeply about implicit bias and how to foster minorities in science.



    • We had a lot of discussion about skype interviews. The molecular & cellular biology search did skypes; our search for the center for systems biology did not.

      Your point about knowing in 5 minutes in an interview (skype or not) whether a candidate is going to work or not is widely shared. I realize I’m in the minority, but personally I’m wary of gut instinct calls like that. Partly because I worry about overly weighting entertainment over substance (I know of people who give great talks, about science that I have later developed serious concerns about when I studied it for real); I think the written record is a more reliable gauge of long-term substantive contributions. Also partly because I worry that the deck is already stacked enough in favor of confident extroverts and against cautious introverts; I think more introverts would be good for science, and I also believe there’s a male/female bias in (over)confidence. And partly because I think (and research apparently argues) that unstructured panel interviews are an excellent way to inject implicit bias into the process. We tend to like who we’re like. The concept of “fit” to a department is necessary, but also fraught.

      Nonetheless, MCB is reporting that they feel the skype interviews were useful – and your point about evaluating teaching ability through observing the candidate’s spoken communication skills is a very important one – so we’ll be thinking more about this. We do have some anecdata to consider now, because a number of candidates applied to both the MCB and the CSB search, and went through the two processes in parallel. (We’re now coordinating, at interview stage.)

      Thanks for your comment about fostering minorities, but I’ll just add to that that it’s easy to make the right noises about these issues — it’s far harder to make real change. The number of underrepresented minority candidates who apply is way too small. At least with female/male representation issues, we feel like we have some traction, and I think that’s why we’re talking more about that than minorities.


  2. Thank you very much for your reply. I share your concerns about skype interviews favoring the entertainers and extroverts. And I agree that carefully vetting a written proposal and associated publications is more informative. I would not advocate to use skype to rank applicants who could be separated by looking at their proposal. But I do think that there are probably applicants who end up not being selected, but who have a proposal as strong as those that get shortlisted for interview and in these cases skype could make sense. Not only can talking to people give valuable information that is not included in a written application. But it also sends a signal to everyone: In this search you go the extra mile to get it right. Skyping many applicants will take some time and doing it goes against the feeling that selecting the candidates in the end is based on some lazy metric. Far too many searches are “for the best candidate and solely focused on scientific excellence”, but somewhat magically end up only inviting people that have SCN papers. So I think taking visible steps to show that you are not just looking at the journal titles in the publication lists makes a positive contribution in general. Your decision to post here your selection process and thoughts about it does something similar and certainly leaves me with the impression that here is somebody who really cares. And not only about getting an excellent candidate for the position, but also about the applicants that are not selected and the academic job market in general.

    As for skype interviews injecting more implicit bias into the search: I agree that this is a real danger, but can be somewhat counteracted by questioning every decision in that regard (which you seem to be doing already). And also having a search committee that is as diverse as possible is of course an important factor.

    Best of luck with your search. Hope there will be an update once you have filled the position.


  3. Thank you for this post. I’m currently on the market myself and I appreciate the transparency you’ve brought to this process.

    I have one question about how you handle the applicant’s demographic information. What do you do when the applicant declines to identify his/her sex, ethnicity, disability status, and/or veteran status? Do you have figures on the frequency with which applicants declined to self-identify? I was under the impression that the search committee members were blind to the supplied demographic information. Rather, it appears that the demographic information figures prominently, or is at least a significant factor, in how the committee approaches the evaluation of each application. Do non-self-identifiers put themselves at a disadvantage? Does the committee assume they are white males (or at least consider them in that batch)?


    • Detailed reporting on people’s self-reported race/ethnic status (apparently called the EEO report) is only available to our committee co-chairs, not the rest of the committee. What we take through as info for the full committee is only female vs. male and URM vs. not URM. If someone declines to identify their male/female status, we still try to figure that out as best we can. If someone declines to identify their URM status, we can’t do anything, and we leave them in the not identified as URM list. The only way this information really factors in is in the way we ask the committee to evaluate the female and URM candidates first and separately, then interleave objectively with the other candidates in a last step.

      I don’t think a non-self-identifier would be putting themselves at a disadvantage. (And I respect non-self-identifiers. The idea of tagging people with this info is, in a sense, exactly what we’re aiming *not* to do, in a just society.) They wouldn’t get considered in that separate first process, but really, I think the main effect of all our chatter and worry about implicit bias is to generate a healthy continuous conversation and reexamination in light of possible bias of all forms — are we judging each candidate fairly on the actual substance of their work and their ideas, did we actually read the proposal and the papers and understand them, or are we reacting to stereotypical ways that people “look like” top candidates, where that includes not only demographics, but also other peripheral stuff like what lab they trained in, do we know the people who wrote their letters, where they published, or how confident and grandiose they make things sound. From what I wrote, I’ve probably overemphasized the demographic issues, just because they’ve been particularly high on my mind for other reasons too, because they’re so fraught with difficulty, and because my thinking has so drastically changed over the past few years because of work like that of Handelsman and Banaji.


  4. On self-selection: I’m on the job market this year, and I didn’t apply – despite being a reasonably good fit on paper. I heard similar comments from a few of my friends on the market. In case you’re curious, the reasons we had were 1) Perception that tenure was unlikely (this may be false, but is common belief), 2) Perception that Harvard hires Harvard and Princeton people only. With #2 in mind, how does the percentage of Harvard-trained candidates compare to the percentage of women and URMs?


    • Thanks for this, the insight is helpful.

      I’d have to go back and count but it was too high. High enough that I noticed (underlining and !!’ing the Harvard connections as I saw them), and worried about it. It’s pretty scary because it’s a self-fulfilling cycle: we can’t hire people who don’t apply. (We can, however, directly encourage people to apply, and I did do some of that.)

      Getting tenure at Harvard has been a problem in the past. It has had that reputation for a long time – it was the reputation when I was on the market in the 90’s. I wouldn’t say the university is entirely clear of that past, but policy here does seem to have changed significantly. In my new faculty orientations, senior leaders talked refreshingly frankly about Harvard’s need to support, mentor, and promote. Certainly our department and the systems biology center intends that anyone we hire at assistant level, we’re intending them to get tenure.


  5. Faculty searches are exciting opportunities. One thing I do not see emphasized is a question that is always on my mind: What kind of colleague would this person be as part of our faculty? It is not an objective question, but it is something that can be assessed directly and indirectly throughout an application packet. I would encourage everyone on a search committee to be asking this question as they screen applicants. Scholarship, research and teaching potential are critical, but inter-personal, collaborative, leadership, and mentorship potential are things I want to think about throughout the process. They are things I always try to address when writing letters of recommendation. I think the goal of hiring anyone into an academic position is to make the whole stronger in every way possible. Over the years, I have been impressed that the skills possessed by the faculty I most respect frequently go beyond their pure research skills and publications.


  6. Pingback: Entry into the Arcanum (or: on admissions committees and faculty searches) | neuroecology

  7. As I mentioned on Twitter, I saw multiple Harvard faculty positions advertised this year and also chose not to apply. This decision was complex, but the reasons outlined by Rheophile were definitely there in the back of my mind. However, I did apply to several Yale/Princeton jobs which were more closely aligned with my CV and research interests (these ads were recruiting evolution, ecology, and microbiology faculty)

    From the candidate perspective, we’re absolutely inundated with jobs ads that come out from July until December (which is quite a long period of time to scan RSS feeds, websites, email, etc. for new postings). I scanned the internet every few days and complied a running long list of job ads that fit my research interests. I would then narrow this down into a short list of “jobs I will definitely apply for” – based on geographic criteria, closer reading of the ad itself, the type of institute, institutional facilities and overall faculty research interests. The short list was then organized by deadline; unfortunately sometimes you end up with >5 applications due on the same day, and so time then becomes an additional factor in making late-stage cuts to your short list of jobs. I think this happened to one of the Harvard ads – I removed it thinking it would reduce my stress levels (one less job to apply for, and I probably wouldn’t even make the short list anyway). No matter how organized you are, the job search is an exhausting, time consuming process for candidates – it completely takes over your life for several months. I put in ~60 applications for faculty jobs this year, which is probably a bit on the high side – but on the Biology Jobs Wikis, it seems quite normal for people to put in anywhere from 20 – 100 applications. Even with template documents for research/teaching statements, every job ad requires – at minimum – several hours of browsing websites, taking notes, and tailoring the application documents. That’s not event taking into account the monotonous online submission systems, and checking up with letter writers to make sure they submit the needed references.

    As an interdisciplinary female candidate, I also worry that my broad interests and experience negatively affect how my applications are perceived. That is one primary reason for self-selecting where I do (or don’t) apply. I work on non-model, non-parasitic organisms (free-living nematodes) but I routinely leverage software tools, methods, and bioinformatics workflows from biomedical and model organisms research (especially C. elegans and human microbiome studies, fields that heavily drive methods development). My research statement is succinct and focused, but my publication list comprises diverse topics (driven by what projects my postdoc labs were working on). So in my mind I am a “Systems Biologist/Ecologist” – I’m interested in how natural ecosystems work and how microbial communities are organized – and since my research is heavily computational, I would benefit by working somewhere like the FAS Center for Systems Biology, since it would enable me to leverage tools/methods from disparate scientific fields. But perhaps the committee members don’t see this link as a selling point of my application, even though I have tried to make it as clear as possible across my cover letter, research, and teaching statements (or they understand the goal, but aren’t excited by it – which you mention as one form of bias).

    As far as moving forward? I get a lot of my information from the Biology Jobs Wikis I mentioned – it’s a cheat sheet for the faculty job search season, and there’s a lot of anonymous, candid information posted there: https://t.co/QAR2LbwZYx – there’s no reason why you couldn’t leave a comment encouraging more female/minority candidates to apply for Harvard jobs. I have relied on that quite heavily this year.

    I would also definitely encourage you to advertise jobs on Twitter – the ads get retweeted by other people, and get propagated across the online scientific community. So even if I don’t follow you directly (which I now do), its likely that your male followers will share the ad, and it will eventually get disseminated to a lot of females down the line.


  8. Pingback: Insider’s view of faculty search kicks off discussion online – Jin's list of interesting papers

  9. Pingback: Recommended reads #68 | Small Pond Science

  10. Thanks for this thoughtful and useful post.

    Can you state what percentage of the candidates you are seriously considering vs. those who didn’t make it past triage have funding through NIH K-awards like the K99/R00 or some other source?


  11. Pingback: Behind the scenes: Harvard faculty search | yeast culture club

  12. Pingback: Zeta the Betta Demands a 40 Hour Work Week |

  13. Pingback: So You Want a Tenure Track Job | Avasthi Lab

  14. Pingback: Links 1/19/16 | Mike the Mad Biologist

  15. Having studied the academic job market for years (in my previous post), I admire your approach but would question one aspect of it. You wrote, “Fourth, I’m looking for other rigorous selections that the candidate has passed before – graduation with honors as an undergrad, competitive postdoc fellowships, substantive research awards.” I understand why you might want to do this to simplify and quicken the triage, but you’re letting other committees do your work for you, and they may not have been as open-minded or as rigorous as your committee about questioning implicit bias. Of course I can’t say how this worked out in your search, but in general this would seem to be a good strategy for overlooking non-obvious candidates.

    Jim Austin


  16. Jim Austin makes a very strong and valid point. I haven’t been on the job market for PI positions yet, but my husband, who is also a Biologist, has. It’s hard to get high impact publications in his field of research since it is very much a concept that people have just started considering and most of the big journals don’t want to take chances. What we typically observe in the response(=rejection) letters so far, is that most search committees are not as open minded about completely new areas of research- when a person has to set up experimental approaches to answer basic, albeit unanswered or worse, wrongly assumed, scientific questions, it is not possible to have the spotlight that gets one nominated for major research awards (the topic is new and people question everthing because it isn’t a high impact factor publication) or in most cases, get the application through with popular funding programs where reviewers want to see significant literature to support the concept and approaches detailed. What happened to bring excited about science- where people once tried absurd experiments to get unexpected and brilliant data, or when science was just about thinking out of the box? Why should one try to publish yet another paper that supports the claims of a popular Science paper and design approaches that are typically anticipated as a follow-up project? So with several flaws in the foundation work that goes into what you use as your basis for selection, aren’t you introducing a major indirect bias, even though you personally aren’t looking into H-index or impact factors? I understand wanting to give back the tax payers’ money back in the form of data, but how can new concepts be trashed without giving people the opportunity? Sorry, if this is a naïve question, but I truly am confounded by this.


  17. Keep in mind the distinction between the triage stage and our later and deeper evaluation. Jim’s comment is correct from an idealistic perspective, but at triage stage, as a matter of practicality, we do need to rely somewhat on imperfect indicators in that first pass. We don’t make the mistake of assuming those indicators are perfect. As I said, no one thing is disqualifying, and I most definitely have my eye out for unusual candidates – and in fact prefer people who are working to open up new questions in new systems – and I’m expecting those candidates to have a different look on their CV, if they’ve been trying to do something radical and different. I’ll pass such people through for a harder and more substantive look. As we get to a shorter candidate list, then we rely more on our own reading of the science.


  18. I am a female bioinformaticist married to a male biologist who is on the market and who did apply for this job and many many others. I came across this blog post in desperation as we try to work out a next step – seminars are getting posted for the jobs in his search and at all the top-fit schools the job talks are going to women. School A invited a woman in the same domain with 1/10th as many publications and a similarly modest background, School B scrapped the ad completely and invited women from completely different fields. Both schools invited 100% women.

    As a woman in science I still want to scream from the rooftops that implicit bias is NOT fixed with explicit bias. I thought we’d already worked out that separate is not equal?

    We have a young son. I cannot imagine encouraging him to follow his parents into biology. My mother has a degree in human biology from Stanford, because that’s what interested her; but works as a seamstress because that’s what worked out in her life. If my son finds his own dreams similarly curtailed by bias, what will then have won?


      • I know it’s not you (we’ve met IRL). I believe the audience of your blog is larger than yourself, and it is for that audience I posted the comment.

        There are explicit quotas (everyone’s opposed to those) and there are implicit quotas, which I believe to be biting my family in the pants. We feel we need more women in the department yesterday, so let’s only interview women… Sadly, my Harvard-trained extremely hard-working husband is probably going to follow similar trainees we know from my own institution into quantitative finance – it’s distressing to lose these good minds in science purely based on their gender, but if one can’t get work what’s a scientist to do?

        Wall Street, despicable as we may find it, is a pure system. The best human will certainly succeed. But it’s a sad day for science, and very sad for my little family.


Leave a Reply

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s