search committee reviewing candidate CVs

Suggestions for running a faculty search

New faculty are the lifeblood of any department. Over the past 20+ years, I’ve seen different iterations of the faculty search process, including occasional forays into semi-dysfunctional chaos. One danger is when the process becomes unduly influenced by individuals or factions with a narrow agenda. Another is search fatigue or burnout, which can lead to hiring someone who isn’t a good fit just to fill the position. Both scenarios can be damaging to the department and are unfair to the person being recruited. Although there’s no guarantee that any search will result in an optimal outcome, the likelihood can be increased if the process is well-planned, fair, efficient, and transparent. With that in mind, here are some thoughts that may prove useful.

I) Establish a sound process and stick to it.

Changing the rules in the middle of a game is almost never a good idea. It undermines confidence in the process, creates bad feelings, and can lead to biased outcomes. So, always start with a sensible and detailed plan for how the search will proceed—one that is supported by a clear majority of the department. (For one such example, see below.) Having a specific plan doesn’t mean that some degree of flexibility won’t be required, but any changes should be minimized, should be clearly articulated, and should be supported by faculty consensus.

II) Divide and conquer.

Depending on the projected amount of work, it can be helpful and even essential to divvy up the tasks involved in a faculty search. This might include having co-chairs to organize and oversee the search as well as dividing up the initial screening of applicants among members of the search committee. For example, if there are 120 applications and six committee members, rather than having everyone attempt to read through 120 applications, divide the six members into teams of two based on expertise. Then have each team screen 40 candidates. This can be aided by dividing the applications into three thematic piles and assigning each pile to the team that is best qualified to do the evaluation, which provides “2x coverage” for each application. (Hint: sorting applications can be facilitated by asking candidates to provide five keywords relevant to their research.) This type of scenario balances workload with the need to build in an appropriate level of redundancy.

III) Generate tiers and rankings in a clear, quantitative manner.

Even though evaluating candidates as individuals can often be difficult, messy, and subjective, this doesn’t mean that the ultimate ranking system shouldn’t strive to be as quantitative, precise, and democratic as possible. Such rankings can be performed at multiple steps of the search process using a few commonsense variations. Here is one scenario, presented as 6–7 steps, that has proven effective and efficient.

Step #1: Establish several broad groupings from the candidate pool.

From the full applicant pool, have the search committee generate groupings of first-tier, second-tier, and third-tier candidates. From an initial pool of 120, this might result in 15 first-tier and 20 second-tier candidates. Second-tier candidates should include strong applicants with clear merit but who were not deemed to be in the top ~15 by the committee. The remainder are considered third-tier candidates. This will provide the rest of the faculty with a reasonably focused number of applications to evaluate. Depending on preferences and the size and strength of the applicant pool, these numbers might vary. Importantly, this system ultimately makes all members of the departmental faculty equal participants in deciding which applicants will be pursued at each stage, with the search committee providing informed recommendations only.

Step #2: Refine and rank the first-tier candidates in preparation for Zoom interviews.

What follows is a generic email (with example dates and times used, simply for clarity) that can then be sent to voting (e.g., tenured and tenure-track) faculty from the departmental chair. It details a ranking system, which can be adapted for subsequent steps in the selection process.

Dear Faculty,

I’d like to start off by thanking the faculty search committee for their efforts and hard work. The committee has supplied me with a list of 15 “first-tier” candidates along with a list of 20 candidates who were also in the running and considered competitive to highly competitive (“second-tier” candidates). The committee acknowledges that these divisions, though well-intended, discussed, and justifiable, are subjective. In addition to considering factors such as productivity, the committee tried to identify applicants who seemed to have a unique niche and a clear research plan and who would be a good fit for the department and overall environment at our university.

After conferring with the search committee, I’d suggest we use the following strategy for reviewing and ranking applicants in preparation for our faculty meeting on January 15 at 1 PM, the purpose of which will be to determine which applicants we will interview by Zoom.

  1. Over the next few weeks, please look over the full applications, for which you will receive a link. For expediency, you might want to focus on the first-tier and possibly second-tier candidates, but everyone is obviously free to review any applications, including ones that may arrive late. Key words in the spreadsheet may help you identify applications of particular interest.
  2. If you believe that an applicant who was not placed in the first tier should be included in this group, please let me and the rest of the faculty know by 9 AM on January 12 at the latest, although the earlier the better. The candidate(s) you name will automatically be added to the first tier. The timing is important, as we will be voting to rank the first-tier applicants ONLY, and faculty will need time to look at their full applications to make their own judgements.
  3. Prior to the faculty meeting on January 15, each faculty member will choose and rank their top-10 candidates from the first tier and send their rankings to me. Note that you MUST rank your complete top 10. The #1 candidate on a given faculty’s list will get 10 points, the #2 candidate will get 9 points, and so forth through to #10, who will receive 1 point. All other candidates will receive a zero from that faculty member. These rankings must be submitted to me by 9 AM on Friday, January 15.
  4. Prior to the meeting a total point score will be assigned to each first-tier candidate. For example, if for a given candidate five faculty gave a rank of #1 (10 ⋅ 5) and five others a rank of #2 (9 ⋅ 5), then that candidate’s total points would be 95. By this system, the top candidates will acquire the highest numerical scores. The scores will be tallied so that everyone can view them prior to the meeting.
  5. The meeting will be an open discussion during which faculty can bring up any points they feel are relevant regarding the strengths and weaknesses of individual first-tier candidates who may or may not have scored highly in the initial round of voting. After the meeting, faculty will have until 5 PM on January 16 to send me an updated top-10 ranking list, if they want to make any changes after the discussion. The final scores will be tallied and sent out to faculty the next day.
  6. Based on precedent we will conduct Zoom interviews with the top ~12 candidates, which will be based on their total final point scores. There may also be natural cut-offs established by the numerical scores that will determine exactly who makes the first round of Zoom interviews. If needed, we can finalize the precise cutoff by email.

Thank you and let me know if you have any questions…Sincerely, The Chair.

Note that to simplify the voting process it can be helpful to send out an Excel file in which the ranking choices are restricted to the first tier. This will clarify what is required and make it easier to compile votes. If for some reason a faculty member doesn’t rank the full 10, the ranking sheet should be returned/rejected until they have ranked all 10 candidates (or otherwise their scores will not be included in the combined final tally).

In my experience, this leads to very clear cutoffs, with all faculty having equal say in the rankings. And regardless of what is discussed at the in-person meeting or how persuasive individual faculty may be in the moment, everyone will have the chance to go back to their offices, cogitate, and vote their conscience in private. In general, I prefer ranked voting to be anonymous, although this is something for individual departments to determine and faculty are often vocal about their preferences anyway. Don’t be surprised if rankings don’t change dramatically, if they change at all, between the first and second votes.

Step #3: Conduct Zoom interviews with top-ranking candidates.

This might involve 10–12 interviews depending on natural cutoffs. Generally, 30 minutes is sufficient for a Zoom interview and should include a scripted set of five or six questions that are relevant to the position along with providing time for the candidates to ask questions of their own.

Examples of such questions are shown below:

  1. What are your research goals as an independent investigator? Could you briefly outline your plans for obtaining research support?
  2. Do you have any specific/special lab equipment needs?
  3. How would you imagine contributing to undergraduate/graduate teaching in our department? (Prompt, if needed, to give specific courses they would/could teach.)
  4. What are the key factors you are looking for in deciding about where to go for a job? What aspects of our university and department were of specific interest to you?
  5. What are some of the ways that you have worked toward greater diversity and inclusion as it relates to your research, teaching, service, or some other aspect of your work?
  6. Is there any additional information you would like to provide us, and do you have any questions for us?

Invite all interested faculty to attend and participate in the Zoom interviews and record them so that those who couldn’t attend can still watch and obtain their own impressions.

Step #4: Establish rankings for in-person interviews.

This will follow the same pattern as was used to establish the initial rankings for Zoom interviews. For example, if 12 candidates were interviewed via Zoom, have faculty rank their top six. Their #1 candidate will receive six points, the #2 candidate will receive 5 points, and so on, down to their #6 candidate, who will receive 1 point. Other candidates will receive zero points from that faculty member. As for above, the results must be submitted prior to an additional in-person faculty meeting. Natural breakdowns are likely to occur, which can lead to a straightforward consensus. Again, faculty should be allowed to voice their opinions in a meeting, and, if needed, rankings can be recast 1 day after the meeting. At that point, the top three to five candidates (i.e., the new first-tier candidates) may be brought in for in-person interviews, in part based on natural cutoffs or other considerations.

One additional component to this step is to allow faculty to comment in their rankings if they think a candidate (following the Zoom interview) is not worth pursuing further. Namely, that they are no longer viewed as a viable candidate. Such information may come into play if one or more of the highest-scoring first-tier candidates fall through. Another relevant factor is timing. For example, it may be reasonable to skip over a candidate and potentially move down the list if a given applicant can’t schedule their first (or second) in-person interview within a reasonable time frame (e.g., within 4 weeks of being invited to visit). Admittedly, this can get complex, and some compromises may be necessary depending on the situation. Just don’t let your streamlined search get derailed because a candidate is simply too busy to interview in a timely manner—it may be indicative of their level of interest in the position.

As far as structuring in-person interviews, I have just a few general comments. One is that the candidate should meet with as many departmental faculty as possible, including the departmental chair. As such, relatively short meetings (20–30 minutes) are customary. Semi-informal meals with faculty represent another opportunity to interact in a more relaxed atmosphere. The candidate should also have an opportunity to meet with students, postdocs, and other staff. In some cases, it may be advantageous to have candidates meet with at least one college administrator. The seminar should precede the chalk talk and, if possible, these should be scheduled for separate days. Lastly, candidates should be shown potential lab space(s), even if these spaces may require renovations. First visits should generally not exceed 1½–2 days total. Avoid overlap between two visiting candidates.

Step #5: Establish rankings for extending second-visit invitations and job offers.

This can be done in a similar manner to what is described above, with each faculty member ranking all candidates, including possible designations of ‘unacceptable’. Suffice it to say that if enough faculty provide cogent arguments as to why a candidate who was interviewed in person is unacceptable, this should be taken very seriously. This may be because a candidate was viewed as a poor research fit, was unprepared, failed to communicate well, or came across as noncollegial. Unanticipated red flags can come up during the interview stages and should not be ignored.

Step #6: Invite the top candidate back for a second visit, hopefully leading to an offer/hire.

Obviously, departments should go all out on recruitment. The first visit was for the department to determine if the candidate was a good fit. The second is for the candidate to determine if the department and locale is a good fit. Activities may include additional meetings with faculty, particularly those who may have missed the first interview. Of course, further discussions with the chair are expected, which can be aided by prior communications and preliminary negotiations to get the parties as close to alignment as possible on salary, startup and space. Candidates may also wish to tour their potential lab space again and should be given free rein to the greatest extent possible. Targeted meetings with administrators or other university officers, including relevant core directors and program coordinators, may also be beneficial. Candidates are likely to bring family, such as spouses, and should be given sufficient private time to explore the area. Tours by a trusted real-estate agent are also quite standard. If possible, host a final-evening informal dinner or pre-dinner party, ideally at a faculty member’s house, and minimally include the departmental faculty and their immediate families. Second visits typically run in the 2–3-day range.

Step #7: When/if the top candidate falls through.

If the top-ranked candidate declines the position in a timely manner after their second visit, move on to the second-ranked candidate (if still available), etc. If none of the acceptable candidates who were interviewed in person take the job, move down the list of acceptable candidates from the Zoom interviews in numerical order for additional in-person interviews. Here it’s preferrable to invite two or more candidates within a short time frame, thereby increasing the probability of success while avoiding awkward up-down votes on individual candidates. If you run out of acceptable candidates from the initial Zoom interviews, take a serious look at the second-tier group and rank worthy candidates as above. Also, don’t be surprised if someone from the “bottom” of the first tier or even the second tier interviews better than any of the highest-scoring first-tier candidates! Evaluating human beings on paper is an inexact science, even if it’s essential and generally instructive.

IV) Balance added breadth and research diversity with a scientifically good fit.

My department has always valued scientific breadth. As such, we have faculty working in many different systems and on different types of questions. Other departments may be much more focused, and this model can work well too. But what unites my department is a shared interest in discovering fundamental molecular and cellular mechanisms. So, hiring someone who is purely theoretical, or perhaps has a primary focus in evolutionary or population biology, could be a recipe for incompatibility. Remember, it’s no fun being a fish out of water and is deeply unfair to the potential hire. New hires can be naïve about what lies ahead (I was!) or even semi-desperate to get a faculty position (I was too!). So, the department needs to look out for the applicants as well as for itself. For example, if the idea of meeting with the candidate’s seminar guests or serving on their students’ committees isn’t appealing, that could be cause for concern. Obviously, this will depend on the size and diversity of your department. But the bottom line is to avoid hiring someone whose work won’t be understood or appreciated by a sufficient number of the department’s faculty. At the same time, hiring virtual clones of current faculty can lead to stagnation.

V) Ultimately, take a holistic view of candidates and what they bring to the table.

Everyone will differ somewhat in how they weigh the strengths and weaknesses of each candidate. And that’s a good thing. For some, the potential for scientific stardom may trump everything else. For others, the perception of collegiality and added diversity may take a front seat. In the end, everyone gets to vote their conscience. I find that almost every candidate who reaches the in-person interview stage is likely to be doing excellent science. So, for me it often boils down to other factors involving soft skills. For example, does the candidate seem like they’d be a giving, generous colleague—one who would be willing and competent to share in the departmental work burden, such as serving on committees and generally lending a hand where needed? More than anything else, I want a good colleague. This includes being a strong scientist, but it also means someone who will have a positive impact on the atmosphere in the department beyond their own lab.

VI) Look gift horses in the mouth!

In the wacky world of academia, situations can arise in which a specific candidate or type of position is unexpectedly presented to us on what appears to be a silver platter. And sometimes these can represent wonderful opportunities to add to our departments. But I’ve become cynical, or at least cautious, when faced with such “gifts”. Perhaps the hire comes through a spousal accommodation, which is potentially great, but, as stated above, if it’s a poor fit then it won’t benefit anyone and may lead to long-term negative consequences. In some cases, administrators may want to demonstrate that they possess vision and therefore know what’s best for your department even if you don’t, so they’ll try to force a certain type of hire. Or maybe a pile of money got earmarked at some point and this leads to your department being offered a hire in a very niche-like area. My advice is to walk away from the table if it seems like there is a high potential for being pressured into hiring someone who is a poor fit. Yes, it’s always nice to get a hire that you weren’t expecting. And yes, saying “no thank you” to a forceful administrator might have short-term repercussions. But do yourself, the department, and the potential applicant a favor by passing if the prospective colleague stands a good chance of not thriving or if there are clear red flags associated with the hire. Your future colleagues will thank you.

VII) Even the best-laid plans…

Searches can toss up many curve balls and may, at times, become slightly mired. Most often this happens during the recruitment and negotiation stages. Spousal accommodations are one example, which are real, valid, and must simply be dealt with. Protracted negotiations and candidates who hold out for other offers represent additional pitfalls. Some of this likely can’t be avoided, and I don’t blame the candidates in most cases; it’s a big decision for them. But having some established ground rules can help. For example, giving candidates no longer than 2 weeks to accept or reject a final offer can prevent searches from collapsing in late spring when other top candidates are no longer available. At the same time, it isn’t fair or productive to unduly pressure candidates into “signing on the dotted line” too quickly. These considerations must be balanced, but it’s appropriate for a department to act in a manner that protects itself from getting strung out and having a search fail.

That said, even well-designed, properly executed searches can and do fail with some regularity. A search may be reasonably terminated when there’s no longer a clear path forward, either with respect to the interviewed applicants or those remaining in the pool. Search fatigue can also set in, and good decisions are rarely made under conditions of burnout. Better in such circumstances to try again the following year, possibly getting out ahead of the competition with an early-fall search schedule—provided, of course, that the administration will support this course of action. If not, then faculty may face a difficult situation, but know that it’s almost always better to leave a position unfilled than to hire a potentially poor fit.

VIII) Other considerations.

There are many, but here are a few. The preceding sections bring up some practical issues, some of which may not be official HR criteria for making hiring decisions. Still, these factors are often discussed and debated among faculty when trying to figure out whom to interview or make offers to. One important question boils down to recruitability: “If we make this applicant a reasonable offer, how likely are they to come?” It would be false to pretend this isn’t a real consideration because searches can and do fail for this very reason. During the interview process, faculty will have their ears pricked to see how interested candidates might be in the position. In fact, this starts with the cover letter and continues through the Zoom and in-person interviews. Personally, if I think that two candidates are roughly equivalent but perceive that one is more likely to take the job, that candidate will always get my vote. (And if you’re a prospective candidate reading this, make sure to signal your level of interest—and please be honest!) At the same time, there is a danger in making too many assumptions about whether a candidate will accept an offer, and I’ve gotten it wrong sometimes.

Admittedly, foremost in everyone’s mind will be the candidate’s perceived research potential. Keep in mind that no one fresh out of their postdoc can possibly have full command over everything that might be relevant to their projects. (I know I didn’t—and still don’t!). During events like the chalk talk, I’m most interested in seeing how well a candidate thinks. Are they logical? Are their conclusions well founded and statements appropriately measured? Do they handle suggestions or even tough questions well? All this isn’t to say that a concerning level of naivete or persistent shallow thinking shouldn’t be a major red flag or even a disqualifier. But I’ve seen issues blown somewhat out of proportion just because a candidate lacked some specific knowledge about a particular pathway or technique (often one that is worked on by a faculty member in attendance). The bottom line is to keep things in perspective. And even if you don’t buy every aspect of their research plan, did you come away with confidence that they’ll think clearly and self-critically, seek appropriate input, and ultimately figure things out—just like we did?

A final factor boils down to diversity, something we are all hopefully striving for. So, if I’m choosing between several good candidates, I’ll likely lean toward those who will lead to increased diversity with respect to gender, ethnicity, or personal background, because this can make for a stronger, more-balanced department. Of course, everyone will weigh these factors differently. In the end, after all the reading, evaluating, voting, tabulation, and handwringing, as an individual you’ll vote with your gut. But having a clear process in place, such as the one outlined above, will minimize the messiness, streamline the debate, and will be fairest to all concerned.