At a glance
Faculty search committees face fewer lines and heavier scrutiny. What research and campus practice say about panel composition, scoring rubrics, training gates and keeping deans, HR and candidates on the same page.
Faculty search committees are walking into the 2026-27 hiring cycle with fewer lines to fill, wildly uneven applicant pools, and, in a growing number of university systems, a training certificate they have to produce before anyone will hand them the files.
The scarcity is measurable. The Computing Research Association's survey of its academic members found units expecting at least 27 percent fewer full-time faculty positions in 2025-26 than the year before. Craig Wills's long-running count of advertised tenure-track searches found roughly 33 percent fewer positions than in 2024 and the lowest volume of searches in nearly a decade, with only the pandemic-hit 2021 cycle worse. In chemistry, the tally of openings at PhD-granting institutions came in at 185 for the 2025-26 search year, against an average of 289 across the three cycles before it.
Then there is the other market. U.S. dental schools reported 522 open faculty positions in fiscal 2025, up from 383 before the pandemic, and named limited response to job announcements as their leading hiring barrier, displacing worries about candidate quality. Same profession, opposite problem.
A committee that runs a careless search in a flooded field ends up with a contested hire. In a thin field it ends up with no hire at all, and a line that gets recycled to the next budget round.
The panel is the first decision
A search committee is a small organisation with a hierarchy problem. The chair usually holds tenure. Junior members may be relying on that chair's vote for their own reappointment. The person from outside the department often has no standing in the room at all.
Written composition rules help more than good intentions. Minot State University's faculty search handbook, revised in summer 2025, sets a floor of three departmental faculty plus one external member, makes a student representative optional, and requires both a chair and a secretary who keeps minutes. The minutes matter more than they sound. A search that cannot show what was discussed at the shortlist meeting cannot defend the outcome to a dean, an HR office or a rejected candidate.
Some campuses add a role rather than a rule. Román Liera's study of a ten-month equity training programme at a religiously affiliated liberal arts university, published in Teachers College Record, followed 17 trained faculty, 10 of whom were designated as equity advocates on search committees. They used that position to change recruitment tactics and the way criteria were argued over, and they met resistance when they did. Liera's conclusion for administrators was blunt: an advocate needs seniority or explicit institutional backing, or the committee will simply outvote them.
Then comes the training gate, now the most contested piece of the process. The University of Washington's academic personnel office began requiring a 90-minute course for tenure-track and teaching-track search chairs in July 2026, and told Interfolio administrators not to give committee members access to application files until the requirement was met. Certification lasts three years. UC Santa Barbara extended its own requirement in June 2026 from chairs alone to every committee member, blocking search plans until each person completes it. UC Riverside went further in March 2026, telling committees already at work to pause until every member had finished a 30-minute online module. The same memo noted that two courses previously required of all members had been suspended in May 2025 following Executive Order 14173.
Texas Tech asks members to repeat its workshop within five years and chairs within two. North Dakota State requires chair training within the three years before a post is advertised, and encourages departments with no live search to train anyway, so the committee is ready when a line is approved.
The pattern worth copying is administrative rather than moral. Training that gates access to files gets done. Training that produces a certificate and nothing else never reaches the shortlist.
Five to eight criteria, defined before anyone reads a CV
Rubric guidance from the University of Washington School of Medicine puts the useful range at five to eight criteria, and recommends deciding in advance whether a search uses one rubric or a staged set, so the same instrument is applied to every applicant at the same point. Scaffolding is common: two or three criteria for the file screen, more added at interview.
The criteria need definitions a stranger could apply. A line like 'demonstrated commitment to undergraduate teaching' is not a criterion until the committee has written down what evidence counts. Gustavus Adolphus College's tenure-track search guide makes the committee answer that question in writing, and asks specifically how 'interest in a liberal arts setting' would be demonstrated by an applicant who has never worked in one.
Setup choices that keep coming up in campus guidance:
- At least two members read every file in the pool, not two readers per shortlisted file.
- Calibrate before the real review: everyone scores five to ten applications independently, then the committee argues about the gaps in those practice scores.
- No rolling review. Reading files as they arrive rewards early applicants, and splitting the pool so one reviewer starts at file one while another starts at file fifty is a cheap fix.
- Build in a tie-break. Lehigh's candidate evaluation tool asks for a third read whenever two reviewers diverge sharply, and pushes reviewers to score against career stage rather than against the department's last hire.
| Stage | What gets scored | Who scores |
|---|---|---|
| Minimum-qualification screen | Only the required qualifications as written in the ad | Chair plus one other member, all files |
| Long-list review | Four to six weighted criteria; every file read twice | Whole committee |
| Job talk and teaching demonstration | The same criteria, plus a separate form for attendees who are not on the committee | Committee and a fixed list of observers |
| Campus visit | No new criteria; the visit is where existing scores get tested | Committee, with notes filed |
Where rubrics quietly stop working
The best evidence on this comes from a study of five faculty search committees in The Journal of Higher Education by Dawn Culpepper, Damani White-Lewis, KerryAnn O'Meara, Lindsey Templeton and Julia Anderson. Rubrics, they found, did improve searches: they clarified criteria, got those criteria actually used, calibrated how evidence was weighed, and in some cases pulled diversity work into view. They were also used in ways that appeared to entrench bias.
The detail is what stings. One committee in environmental engineering gave every candidate a zero on service because nobody had defined what good service looked like, and a member's joke about it went unchallenged. A chemical engineering committee treated one prestigious research award as sufficient grounds to advance a candidate regardless of scores in other areas. In a plant biology search, each member submitted a personal top-five list that was never checked against the rubric scores. A microbiology committee, by contrast, stated plainly that teaching was the heaviest weighted criterion, built the rubric around it, and could explain why the finalists won.
Two habits separated the stronger searches. They set cut-off scores and then deliberately looked past them for candidates who were outstanding in a single area. And a developmental psychology committee re-read every file from a minoritised candidate regardless of score, treating it as a bias check. Both moves widened the pool rather than narrowing it.
The scale on which people score also matters. Rivera and Tilesik's study of teaching evaluations found that a six-point scale produced more consistent ratings across genders than a 10-point scale, which seemed to activate stereotypes about brilliance. And Blair-Loy and colleagues, auditing one engineering department's own rubric scores, found women still received a 0.36-point penalty on research productivity after controlling for seniority and publication count. At the bottom of the citation distribution the gap reached 0.7 points, roughly a 35 percent penalty. Women did not draw level until an H-index of 17.5, against an average of 12.8 in that pool. Their higher scores on contributions to diversity offset the shortfall, which is a strange way to run a merit system.
A rubric is a structured argument, not a verdict machine. Used as a verdict machine, it launders the committee's preferences and attaches numbers to them. The full study is worth reading at Taylor & Francis.
Photo by Brett Jordan on Unsplash
What fit actually measures
Most committees say they hired for fit. White-Lewis's earlier study of four search committees at a single university, based on 23 interviews with faculty, chairs and deans, found that fit as an organisational concept barely operated. Basic qualification matching weeded out roughly a quarter of applicants. The rest of the winnowing ran on what he calls idiosyncratic preferences.
Research on U.S. immigration and marriage patterns was described as topically 'narrow' in one social science search, even though the same reviewer called the dataset novel. A candidate studying antigay bias was judged not to be 'as much of a scientist' as applicants doing basic science. The pattern held across disciplines: work touching a specific population, region or identity was discounted as insufficiently general. As Inside Higher Ed reported at the time, fit functioned as coded language more often than as a measurable standard.
The fix isn't to abolish fit. It's to drag it into the document. If a department needs someone who can teach a specific core course, carry a specific research group or supervise a specific method, that belongs in the ad as a preferred qualification with a definition attached, and in the rubric as a line item with evidence attached. What can't be written down shouldn't decide a career.
Alignment means signatures, not consensus
The most effective alignment device I've seen is also the least glamorous. Gustavus Adolphus College requires the committee's evaluation rubric and its draft position announcement to be submitted together to the provost's office for approval. The two documents cannot drift apart, because they arrive in the same envelope.
Below that level, alignment is mostly about who says what to whom. Minot State requires interview questions to be cleared with HR before they are used, asks that every candidate get the same questions in the same order, and keeps search records for three years. Its handbook also specifies that the formal offer and salary negotiation stay with one named administrator, which stops a well-meaning committee member from improvising a promise nobody can keep.
Texas Christian University's AddRan College adds two practices worth stealing. The department chair writes down why the chosen candidate was chosen, and writes a separate justification for why each other finalist was not. And every candidate gets a short exit meeting before leaving campus, which lets the chair explain next steps and hear how the visit actually felt. AddRan also recommends scheduling the same independent person, someone outside the search, to meet every candidate, so their impressions are comparable rather than scattered.
Department votes are part of this. White-Lewis found votes can turn on departmental politics or on who was hired the year before rather than on the candidates in front of the committee. A written rationale tied to the rubric is the only thing that makes a vote reviewable later.
Diversity criteria after the retreat
Heterodox Academy's analysis of more than 16,500 full-time faculty advertisements, drawn from postings on HigherEdJobs.com between August and December of two successive years, found requests for diversity, equity and inclusion material falling from about 25 percent in 2024 to 11 percent in 2025. Standalone DEI statements dropped from 13.4 percent of ads to 2.5 percent. Seventeen states have banned the use of such statements in public university hiring.
The decline is not uniform, and the signalling has not disappeared. Roughly 37 percent of the fall 2025 ads required no DEI material but still told applicants that a commitment to it would be valued. About 40 percent said DEI would be used as a framework for evaluating candidates, almost exactly the share in 2024. Rates ran from 69 percent of ads in Vermont and 44 percent in Maine to 0.9 percent in Kansas and zero in the Dakotas. The University of California dropped its requirement entirely, with provost Katherine Newman arguing that mandating the essay pushed applicants to dwell on something outside their expertise. The wider pattern is covered in our report on the decline of DEI statements.
For a committee, the operational rule is simple. If the ad promised a criterion, the rubric has to score it, and it has to be defined in behaviour a reviewer can see: mentoring record, curriculum design, work with specific student populations, evidence of widening access to a field. If the ad dropped the criterion, drop it from the rubric as well. A criterion that survives in the committee's heads after leaving the document is the one that gets a search challenged, and the one that lets a reviewer's private preference do the deciding.
The numbers on who gets hired haven't moved much either way. The Computing Research Association's Taulbee Survey of new faculty hires found Black and Hispanic new hires combined still in the low single digits across tenure-track and teaching-track roles. Whatever the paperwork says, the composition of the shortlist is the number that matters.
What a candidate can reasonably ask for
Applicants have more standing here than they usually use. Ask the chair which criteria the committee is scoring and whether the file review is structured. A committee that has done the work answers in two sentences. One that hasn't will tell you the process is 'holistic' and change the subject.
Other questions worth putting in writing: who decides, and at which stage; whether feedback from job talk attendees is scored or advisory; when shortlisted and unsuccessful candidates will be contacted. Minot State's guidance tells chairs to notify non-selected applicants promptly and to give finalists a personal call. If a search promises that and doesn't deliver, it tells you something about how the place treats people.
Keep your own notes after each conversation. If the chair's account of the timeline and the dean's don't match, that's information, and it's better to have it before you accept an offer than after.
One more thing, for anyone reading a thin applicant pool as a verdict on their field: dental schools named limited response to advertisements as their top hiring barrier in 2025. A low number of applications is usually a statement about the ad, the listing and the timing, not about the discipline.
Photo by Zulfugar Karimov on Unsplash
The one change that does most of the work
Write the rubric before the ad. Send both to whoever signs off, as one document, and don't open the files until the committee has scored practice applications together and every member has cleared whatever training the institution requires.
That's an afternoon of work. It gives the chair a defensible answer at every stage, gives the dean something to approve rather than something to arbitrate, and gives candidates a process they can predict. Committees that do it don't end up with fewer arguments. They end up with better ones.
- higher-education
- academic-recruitment
- faculty-hiring
- dei-statements
- search-committee

