Draft Statement on Letters of Recommendation

Thank you to all the members whose thoughtful feedback helped the Executive Council and me revise these guidelines. The final statement is now available on mla.org. I hope you will share these guidelines with job seekers and other colleagues involved in the hiring process.

I welcome your frank opinions on a new MLA document that would, if approved by the Executive Council, outline some best practices for those who write, request, or read letters of recommendation for graduate students seeking employment or postdoctoral fellowships in MLA fields.


For some time now, I have had conversations with graduate students and faculty members about the difficulties in requesting and writing recommendation letters. In the current stressful and often enigmatic competition for academic jobs and fellowships, students wonder how many letters are really required (perhaps over and above the number stated in a job advertisement), about what an effective letter looks like, and about who will write them the most effective letter. Faculty members also wonder what makes a letter effective. Even the term effective is enigmatic in the present network of academic social relations, as a comparison of current letters with those written in the past dramatically shows. The genre has a long and multicultural history; from what I have been able to reconstruct of it, the (almost always male) letter writer typically knew whether his letter was effective by whether its recipient did or did not give the recommended candidate the position or service respectfully requested on his or her behalf by the letter writer. Today, in contrast, institutions require multiple letters and typically delegate tasks of judgment to committees with several members who read dossiers that contain the candidate’s self-recommending letter along with letters from several knowledgeable (but not disinterested) faculty members, and there are very few ways for the writer to gauge the value of his or her letter—or even to know what systems of value are at stake.  It is no wonder that some faculty members find the work of writing letters time-consuming, perplexing, and sometimes unrewarding; they know that a single negative phrase can harm a student’s application for a job or fellowship, but they have little evidence that entirely positive letters help students achieve their goals.

Moreover, when departments request (or simply accept) six or more letters for each applicant, faculty letter writers (who are of course also often readers of job dossiers) cannot help being aware that the labor and time of writing letters are out of sync with the labor and time devoted to reading them (especially at the early stages of a job search). Recognizing that the terrible academic job market is at the root of many problems that faculty members encounter as they attempt to help their advanced PhD students gain an academic position, I think that the MLA can make some useful suggestions for best practices in this small but important arena of academic work.

The following draft proposals distill some of the arguments presented in a longer and more historical form in an article I wrote for the special topic Work in the October 2012 issue of PMLA, coordinated by Vicky Unruh, and incorporate suggestions from an Executive Council discussion in May 2013. The recommendations are modeled in part on those in the document entitled “Professional Employment Practices for Non-Tenure-Track Faculty Members: Recommendations and Evaluative Questions.” Prepared by the MLA Committee on Contingent Labor in the Profession, that document has been sent to department chairs, university administrators, and MLA members who are graduate students and professors in both tenure-track and non-tenure-track positions. Anecdotal evidence indicates that the committee’s suggestions are bearing fruit in some places.

Going Forward

I hope to bring a revised statement to the council in October 2013, incorporating your suggestions. I would be grateful for your comments on the recommendations below as well as for your suggestions for other possible recommendations. During the council’s discussion in May, there were differing views on limiting the number of recommendations (outlined in point 7), so I also invite you to comment specifically on that issue. Documents recommending changes in professional behavior are most likely to be read if they have been debated and revised in the process of being produced. I hope you will participate in that process now!

Thank you, and all best wishes,

Margaret Ferguson

First Vice President

Proposed Recommendations

  1. Faculty letter writers should consider whether the length and content of their recommendation are appropriate in relation to the student’s letter of application. Some faculty members who write letters of more than three pages may be offering descriptions of the student’s intellectual project that compete with, rather than support, the student’s account of his or her project.
  2. Graduate students should ask faculty members for letters of recommendation well in advance of going on the job market so that letter writers can visit graduate students’ classes to gain firsthand knowledge of their teaching practices. The corollary of this recommendation is that faculty letter writers, especially but not only dissertation directors, should be able to comment knowledgeably on the students’ syllabi, student evaluations, and classroom practices. Many postdoctoral fellowships require some teaching as well as a strong intellectual project, so faculty letter writers, especially dissertation directors, should be prepared to write an informed letter that speaks to both pedagogical and scholarly issues.
  3. Directors of graduate studies should consider reviewing all letters of recommendation written for students going on the academic job market. Alternatively, graduate directors or department chairs could ask a trusted faculty member to perform this service, which could include proofreading as well as fact-checking (e.g., how many courses the graduate student has taught) or discrepancies in projections about important matters such as the dissertation completion date. Reviewing of students’ recommendations was a standard practice in many departments and other academic units before electronic uploading of individual letters became the norm. Some departments may have found ways to continue the reviewing practice in the digital age. (If your department has done so, please describe its process in your comments.)
  4. Hiring departments (and deans who oversee departmental hiring protocols) should consider requiring letters of recommendation only for finalists or semifinalists in a job search rather than for all applicants. With electronic communication resources, there is no longer a need to have all parts of a dossier submitted at the same time. Many faculty readers of dossiers say that they don’t read letters of recommendation carefully until the applicant is at the “long short list” stage of the selection process. (Note: It’s my sense that faculty members could tailor their letters more effectively if asked to comment at a later stage of the hiring process—but I don’t want to make the possibility of such tailoring, which involves extra work and which some of us already do, a linchpin for this general recommendation.)
  5. Deans and provosts should consult with department chairs in language and literature departments before making decisions (or delegating them to human resources administrators) about adopting a platform that has an automatic limit on words or on characters for uploaded letters of recommendation. Setting automatic length limits requires extra work from recommendation writers, who submit many letters for many students applying to different institutions. Academic administrators should also consult with their counterparts in other institutions about how the rapid and unstandardized changes in the reception of job search materials are affecting the labor time of faculty members in fields where detailed narrative evaluations are the norm.
  6. Administrators in agencies and foundations that award postdoctoral fellowships should consider changing automatic word or character limits for recommendations if such limits have been set for the convenience of the staff or of outside reviewers. Stating clearly in the directions to letter writers that letters should not exceed two or three pages is reasonable; cutting off letters automatically is not. To find that your letter must stop in midsentence (or midword!) creates bad will and a collective loss of faculty time. Moreover, if faculty members think carefully about the above proposal, concerning the appropriate length for most letters, the automatic cutoffs introduced by electronic uploading procedures may become unnecessary. That is, fellowship administrators can join a field of negotiation and debate rather than allow technology to decide on an issue that has not yet been adequately discussed in the profession.
  7. Hiring departments should consider limiting the number of required letters of recommendation to three (four, perhaps, when the position requires competence in more than one area). At present, there is considerable confusion on the part of graduate students about how many letters are “really” needed; some students are asking for—and receiving—as many as nine or ten letters while other students (especially those going on the market for the first time) feel doomed to failure because they have letters coming only from their dissertation committee members and perhaps from someone else who will write a “separate” teaching letter. If letter writers adopt a holistic approach to their letters and if students request letters in advance, allowing letter writers time not only to observe classes but also to comment on job search materials and writing samples, departments could at least have a robust discussion of how many letters are “really” needed for candidates for a certain job. Once such a decision is made, departments could give it heft by stating clearly in their MLA Job Information List advertisement that the search committee will read only the first (three? four?) letters in any dossier.
  8. Faculty members who write letters of recommendation for graduate students (or undergraduates or colleagues) should carefully consider the legal opinion that “recommendations should be written on the assumption that the subject will read the letter.” This opinion, formulated by lawyers employed by the University of Alabama, Huntsville, occurs in a document that examines, among other things relevant to faculty recommendation writers, a student’s right, granted by the Family Educational Rights and Privacy Act (FERPA), to read statements about him or her. While students may waive this right on the letter of recommendation form, faculty letter writers should understand that such waivers may not hold up in a legal proceeding. (Note: Confidentiality is also a vexed issue in external review letters; some states prohibit public universities from keeping letters of evaluation confidential. That issue might be taken up in these proposals.)


Kristin E. Pitt

When I was applying for jobs in the early 2000s, many job applicants, particularly those still in school, submitted letters of recommendation via a dossier service provided either by their departments or by university career centers. This saved faculty time: they wrote one letter that was submitted to the dossier service and then copied and sent out as requested by students. However, for those of applying to jobs across multiple disciplines and interdisciplinary programs (for example, I was applying for jobs in CompLit, English, Spanish, Women’s Studies, Latin American Studies, American Studies, and Humanities programs), it meant that our dossier contained letters that may or may not be of particular interest to all search committees, and at least the dossier service at my university wouldn’t tailor the contents of the dossier to individual jobs. I don’t know if such dossier services are as common now in an age of electronic submissions, but if they are, limiting the number of letters might be a real problem for students in similar situations.

Tilottama Rajan

I’m in agreement with much of what is said here but have three concerns. One concerns limiting the request for letters to semi-finalists. While it may be that my/our positive letters are largely useless, I do think that the student’s application letter, record, writing sample and recommendation letters make up a package. To reduce this package to just the application letter and cv is to reduce the applicant’s chance to make his or her case, while on the other hand it isn’t a lot of extra work for me to send the standard letter I’ve already written for a student in whom I realy believe to a few more places (or to a dossier service). I understand that committees may disregard letters because “all letters are good.” Still there’s good and excellent, and I fear that if letters are removed, the filter will be (and maybe already is) the place of the degree: Ivy League students become semi-finalists, others do not; amongst Canadian universities (where I am), only Toronto counts. Etc. The hiring committee is likely to use the application letter as an initial filter, and perhaps not move beyond that in many cases, but the materials should all be there.

2. Your recommendation that all letters of recommendation be reviewed by the graduate chair or someone else. I believe my dept. followed this practice briefly several years ago and gave it up. I’m in the position of sometimes writing letters for several students applying for the only Romantics position in this Victorian country to have materialised in three years, or for the many positions in the US that do turn up but for which my students doubtless have no chance. They’re all “good” letters that none of these students would take offence at if the Family Act (which I don;t think we have here) allowed the students to see them. But reading between the lines, I trust it’s discernible that I’m speaking to the different strengths of A and B and putting C a bit lower while making the best case for him or her. I would resent, frankly, being forced to beef up my letter for someone I don’t rate as highly. What is the point of the letter, if they all have to be equally good?

3. This is more a question than a concern. You indicate that the dissertation advisor should speak to the student’s teaching. Our students, for collective agreement reasons, are not allowed to teach their own courses and are only TAs in larger courses. Sometimes they later acquire part-time experience at some other university and I can;t sit in on their classes. In addition, I haven’t taught undergraduate classes for 20 years, and can’t speak effectively to pedagogy, though I can extrapolate certain kinds of teaching strengths from the nature of the dissertation, conference presentations etc. I therefore do a bit of the latter but concentrate on the research, and my students generally then also have a “teaching letter” from the person for whom they’ve taught or someone else who has observed their tutorials. IS this a bad way to proceed? I’m not sure I can do otherwise …

— Tilottama Rajan, Canada Research Chair, Univ. of Western Ontario

Jenni Halpin

Reducing the number of letters to five or fewer and delaying the letters until the first cut has been made strike me as enormously sensible, as does curbing the overly-long letters. But one thing I notice is that there’s an emphasis in this set of recommendations on reducing volume, but we also have a problem in the opposite direction: letters that say too little.

I have read a surprising number of letters of recommendation doing little more than saying that the writer has known the candidate in some capacity for some amount of time and then adding a sentence or three highlighting details already presented in the CV. (I’ve seen a lot more of these too-brief letters than of the too-long ones.)

Such a letter is as much a waste of time for the reader as for the writer. Particularly in searching for a new colleague, I want detail. A CV may just “tell,” but a letter should “show” the reader something of who the candidate is, preferably something readers would be unlikely to learn from the CV and the candidate’s own letter, or with some degree of personalized insight that would help those CV lines come forth as examples rather than bullet points.

Perhaps the first recommendation could be expanded to offer warnings away from the overly long letter and the unduly brief one?

Kristin E. Pitt

Another thought about the recommendation that letters not be requested in the first round of applications: while I agree that lots of search committees don’t read a lot of letters until they’ve established a “long short list,” those same committees may nevertheless wish to request letters from all applicants if they anticipate that they will need to be move from a “long short list” to a “short short list” quite rapidly. There have been a lot of compressed hiring timelines in recent years arising from budget uncertainties: many institutions have approved searches much later than usual, and many programs have tried to complete searches much more rapidly than usual in the hopes of avoiding the consequences of anticipated budget freezes. If you only have a couple of weeks–or less–to winnow your long short list down to the 10 candidates you’d like to interview at the convention, or perhaps even down to the 2 or 3 candidates you’d like to bring to campus, then you’ll probably want to have all the materials necessary to make that decision available to you from the start.

Martha E. Cook

I read the PMLA article on letters of recommendation and was frankly shocked at the bias against short letters–and I did not feel that anything like a thorough effort had been made to establish any objective criteria for making such statements. I have primarily written letters for B.A. and M.A. students to recommend them for graduate school; but I have also written letters for former students with Ph.D.s, including for deanships and college presidencies. I doubt that I have ever written a letter longer than two pages, with single-spaced paragraphs with double spaces in between. Often I have written one-page letters. My goal is to say what I have to say about my experience with a student, not to repeat what he/she has said in an application letter, dissertation abstract, statement of goals, etc. I have read a lot of files of applicants for teaching positions at my institution and find it hard to imagine the effectiveness of half a dozen letters running three or four pages each. A letter of application plus writing sample, teaching philosophy, etc., added to twenty-plus pages of references? Remember that many professors on search committees are teaching four or even five classes a semester–all of your students are not applying to research institutions. Let’s exercise some common sense.

George L. Justice

Just as a thought experiment: what if hiring departments and universities no longer required letters of recommendation at all in the job search process? What would we really lose?

Applicants could list references on their CVs, so hiring committees, chairs, and deans could contact them for further information about the candidate at the point it would be useful. Comments could be by phone or in writing. This would reduce labor, tailor the comments to the questions the hiring entity has about the candidate, and perhaps make the process more fair.

There must be better ways than the typical (bloviated) letter of recommendation to discern a job applicant’s scholarship and teaching.

Gaurav G. Desai

A further thought experiment– what if search committees were asked to come up with their finalists in a “blind submission” process — having access to the content of all material but not to the names or institutional affiliations of either the candidates or their recommenders? While there is much of merit to many of the proposals made above, I think we are missing the real problem. Most committees, I suspect, put more weight to institutional affiliations and celebratory status of the recommenders than to the actual content of the letters. In the inevitable instance of then being faced with a number of superlative letters from the same few individuals the tendency is to mine the letters for any “red flags” that they may raise. I confess that over the years I have put less and less weight on such letters especially when they come from within the U.S. (European and particularly British letters are an entirely different matter).

Marguerite Waller

I am very grateful that the subject of recommendations has been brought up and is under discussion. All of the questions I see raised here are intrinsically fascinating, and some of the most pressing issues could probably be expeditiously addressed. It would be enormously helpful, for example, if universities and colleges could agree upon a simple, user-friendly letter submission interface. (Recommendation 5) Perhaps we could also curb the desire to “upgrade” for a few years. I was resentful of Interfolio at first, but came to appreciate its convenience for both students and faculty and worked diligently to write very comprehensive letters.

Whatever is decided, I hope we can steer clear of efforts to standardize. I take the genre seriously, ever since as a graduate student I was able to get a nontraditional undergraduate student into Yale Law School, apparently by the sheer persuasiveness of my rhetoric. I try to write, and I like to read, letters of recommendation that have attitude, voice, style– or whatever one wants to call it, dramatizing in some way who the candidate is and what s/he is about.

I was intrigued by the suggestion that applications be read in an institution-blind way. I doubt this is practical, but it corresponds with the way I try to read files. If we want to keep our departments and institutions aerated, then we should be looking in all directions for scholarly edge, theoretical originality, and beautiful teaching.

After 39 years in the profession, I have no doubt that three or four letters are sufficient to give the hiring institution a multi-angular view of the candidate.

Thank you to First Vice President Margaret Ferguson for broaching this discussion. Opening a window on current practices will help us all read each other’s letters more fruitfully.

–Marguerite Waller, Chair of Women’s Studies and Professor of Comparative LIterature, U.C. Riverside

Amy Vidali

I appreciate many of the pragmatic suggestions here. My concern is around issues of diversity and recommendation letters. I’ve done some work on disclosing disability in letters of recommendation, and I would like to see a suggestion regarding speaking with candidates before disclosing disability, and possibly other identity traits. While letter writers may believe it “common sense” to avoid disclosing particular traits (such as marital status, sexuality, or race), disability is frequently disclosed in problematic ways, and there are more and more disabled candidates moving up in our ranks (appropriately, I think!).

Perhaps we can make suggestions not only on how we disclose (letter length, when required, etc.), but what we write, even if that opens a large can of worms. (My article: http://www.tandfonline.com/doi/abs/10.1080/07350190902740042#preview.)

Mary Nyquist

These recommendations are extremely helpful, and I agree with *7 on limiting the number of letters required. But it seems that the conversation has shifted as a result of T. Rajan’s comments regarding place of degree as, in reality, the primary “filter.” Since institutional affiliation and celebrity status of referees undoubtedly play a decisive role, the radical suggestion of “blind” candidacy seems worth considering. If this practice is recommended for publications, why not for jobs? The other radical proposal — that letters or interviews (via telephone or Skype) be required only at a late stage of the search process —. also deserves consideration. The problem of compressed time-lines, mentioned above, would have to be addressed, but I don’t think it’s insurmountable..

John Allen Stevenson

First, many thanks to Prof. Ferguson for taking on this issue: doing something to improve how we write and read letters of recommendation would have a huge impact, not just on professorial quality of life (no mean feat) but it would also begin to address larger problems in the way we conduct searches in general.

Letters of recommendation have become an otiose arms race, ever longer, ever more detailed, ever more panegyric, ever more useless. I like George Justice’s suggestion: let’s try quitting the whole wretched business. In the last several years, I have chaired two searches, one for a senior administrator, one for an assistant professor. Both had good outcomes but in neither case did letters of recommendation have any bearing on that ultimate success. The difference is that, in the senior administrator search, we did not ask for letters. Rather (and this is of course standard procedure) we asked for names of references, whom, at a late moment and after reducing the field to finalists, we contacted for a conversation. What we heard was more candid and more useful than any letter I have ever read as part of a search. Let us not fiddle around at the margins. Let’s attack a recommendation system that wastes enormous time and energy and which contributes substantially to the way the academic job market has become the melancholy picture we contemplate today.

David C. Lloyd

Like others commenting here, I am grateful that this issue has been raised and discussed so thoroughly. Having worked with many graduate students at several institutions, served both as a search committee member on many occasions and as a placement director recently, and written all too many letters both of recommendation and external review, I am like everyone who has commented all too aware of the enormous quantity of academic labor that goes into these tasks and of the importance, for graduate students and promotion candidates as for letter writers, of finding ways to make the process fairer, more efficient, and less laborious. I am largely in accord with the proposals made above. I would hope that the MLA would at least make a strong recommendation in relation to letters of recommendation for hiring purposes that no more than two letters should be required until the short list has been determined by the search committee. More seems to me at that stage unnecessary and, as several people have commented, when facing 300 or more candidates’ files, any search committee member will skim most letters (another part of academic labor inflation is this inevitable result of a poor job market where candidate numbers are increasing and many are applying to jobs at which they have a very long shot if any chance at all). Similarly, the number of external review letters for tenure and promotion has become excessive. Surely no institution really needs more than six or maybe eight letters of evaluation for candidates (and I am aware of at least one institution whose upper administration sought in some cases twenty or more, which made the process a virtual fishing expedition). In my experience, writing external reviews is enormously time-consuming if done responsibly (not to mention coinciding with the job letter season), but more importantly the simple physical limits to what one can do does adversely affect candidates whose departments request letters later in the year. All too often declining to write for a candidate, even if one is simply unable to take on any more reviews, is taken as a negative comment on the significance of their work, which places the potential review writer in an unenviable ethical position. The fewer of either kind of letter, recommendation or review, that are asked of us the more responsibly we will be able to perform the task.

I do agree that letters should not be standardized, in length or in response to bureaucratically generated questionnaires. Common sense suggests that they should be detailed, evidencing real understanding of the candidate’s work and, where possible teaching skills, and normally not more than two to three single-spaced pages. However, I have seen all too many letters (often from senior and eminent colleagues) that seem to assume their very general afflatus will carry weight just on account of who they are. Hopefully that proves not to be the case, though, as a hedge against such hangovers of the old-boy network, I am attracted by the idea of blind job submissions. The difficulty with that, however, is that a candidate has to reveal their institution in describing their academic career, if only in their cv, and in my experience the tropism towards the ivy league precedes the reading of recommendation letters and is often confirmed by deans and provosts, especially at non-ivy schools with higher aspirations for themselves. However, as graduate programs shrink, it is worth all job search committees recalling that an inevitable consequence of that–as of the diminishing job market–is that talent is becoming much more widely spread: grad students who formerly might have been admitted to the University of A+ are now going instead to University A- or B+, but are no less intellectually interesting as a result. Indeed, across the board more interesting, riskier and more challenging work may be being done outside the highly ranked programs and the more we can do encourage an ethos that is more open to a variety of institutions the better. Multiplying the number of reference letters (which usually means that candidates are being encouraged to seek out letters from eminent professors with whom they have not worked) will not address that issue.

An MLA guideline that limits the number of letters and/or the timing of their being requested would help graduate students who do often feel the more external support they have the better but are then made unnecessarily anxious if they do not have it at their stage in their career. We should also reiterate an old MLA guideline that sample writing should not be requested until a candidate has made it through the initial selection process.

Michel Chaouli

Many thanks for these thoughtful and sensible suggestions. I love George Justice’s proposal: do what other professions do and ask applicants to provide names of referees, who, if needed, will be contacted by the hiring committee. The proposal has the merit of tackling the hyperinflation of letter writing at its source.

John Marx

A practical comment and a broader (thus less helpful) thought:

Practical comment: At UC Davis the placement advisers vet recommendation letters electronically and students on the market have found this to be helpful. When students are told by the dossier service that all their letters have been posted, the placement advisers log into the service (which should be able to grant such admin authority if asked to do so) and look at the letters. I never found this to be an onerous task when I served on the placement committee. In my experience, this kind of vetting tends to be limited to error and typo correction (does the letter have the right date on it, or does it look like it was written last year, etc.) and to general assessment of the dossier as a whole (does it appear that all the letter writers are describing the same person, etc.). On one occasion, I had a follow up conversation with a letter writer whose recommendation struck me as a little light on substance (the problem Jenni Halpin raises). The goal of that conversation was standardization of the present, perhaps undesirable, letter writing mode.

Broad and unhelpful comment: As both Gaurav Desai and David Lloyd suggest, there is a broader frame here so obvious that it might appear not to bear noting. We still write our letters amidst hierarchies both of the institution and the scholar, even if the old boy networks you invoke no longer dominate. Barring something like blind submission, I don’t know that standardization will always help students doing intellectually interesting work in what Lloyd describes as “University A- or B+.” Maybe leveling the playing field is not the goal, and maybe it shouldn’t be, but it seems to me that standardizing the letters may have the inadvertent effect of elevating the value of letterhead from “University of A+.” I have certainly advised students in the past who have sought letters from scholars outside their institution to demonstrate that they are building a network or included a separate teaching letter to emphasize that important part of their training. Do we want to take those options off the table?

Gina Bloom

As much as I would love, like others above, to scratch the whole obligation to write letters of recommendation, I doubt most members would be willing to go that far and would be thrilled to see implemented at least the more moderate of the suggestions above. I like David Lloyd’s suggestion of limiting the number of letters at the early stage of the process to two. I fear, though, that this will not be enough to slow down what John Allen Stevenson aptly calls the current “arm’s race.” If we do not limit the total number of letters considered, even for short-list candidates, then we will not be able to grapple effectively with the labor issues at stake here. I very much support the proposal to limit letters to no more than three, notably the same number required for most fellowships. I would also like to see MLA at least make a statement about optimal length. Certainly, faculty should be allowed to write more if they wish, but if MLA were to endorse the shorter letter as acceptable, then many of us would not feel compelled to write long letters simply out of fear that failure to to do so will do a disservice to our students. For instance, I have found that colleagues at many non-American institutions often write short letters (sometimes just over a page). When members of hiring committees see these letters, they do not necessarily conclude anything negative about the candidate; they read the letter in terms of the cultural conventions of the letter-writer. If MLA were to articulate some best practices in terms of letter length, faculty who would like to write shorter letters could do so without harming their students’ job prospects.

Jean Elizabeth Howard

Thanks to all for the helpful comments on this issue. I very much like the idea of requiring only 2 letters at the initial stage of an application and then no more than 2 more at a later stage. This would be feasible if the primary letters writers really did take responsibility for writing about pedagogy as well as research. I think a concerted focus on reducing the number of letters required for each candidate would go a long way to reduce faculty work load in this area and still provide all the information a search committee could possibly want. I also think a strong suggestion that no letter exceed 3 pages is a very good one with the norm being 2. We all spend too much time re-describing what the candidate should have described, i.e., the argument of the dissertation. If candidates use Interfolio, it is still possible to vet the completed dossier by having it sent to the placement officer or dissertation advisor before sending it to schools. I have not myself had arbitrary space limits put on my letters by the software with which I was “interfacing,” but would be very annoyed if that happened. But for me the big issue here is the overkill in the present system. We need to urge shorter letters and fewer of them.

John Allen Stevenson

Let me make an addendum, despite the fact that the suggestion that we do away with written letters altogether does not seem to have generated a groundswell of support.

There is an anxiety, running through this thread, that messing around with letters as currently consituted will disadvantage students from the non-elite institutions, that “reference checks” as opposed to carefully crafted documents will send us back to the dark ages of private calls amongst the old boys. I simply think that fear is exagerrated. My experience is that a responsible search committee can, when the time is right, make the necessary calls in a professional manner–this is not friend to friend but search committee to reference, a very different animal in my experience.

I think what we are reluctant to face, perhaps, is a job market where we focus completely on the writing and (for those who make it that far) the interview (s). I would like to see an institution-blind application where reviewers did not even know where the student was doing his or her work, but I am probably well into science fiction by now.

Elizabeth Freeman

I very much appreciate Margie’s willingness to take this subject up. I think recommendation letters are most valuable when a committee is hiring in a field not represented on the committee, as sometimes happens in emerging fields. That’s the only reason I can see for not getting rid of them entirely (and is particularly important in promotion and tenure cases, where senior faculty can be unaware of newer field formations, journal rankings, etc). Therefore, for job candidates I’d vote for a director’s letter and a teaching letter only until the finalist stage — the director ought to visit classes and speak to teaching, but I think other faculty may be situated to write a more comprehensive teaching letter. Directors ought to be charged with explaining the significance of the work in one or more fields, as graduate students sometimes just don’t have the large frames to put around their work — but not with reiterating the argument. The finalist stage ought to occur no less than one month before the selection of MLA interviewees, to allow other letters to be prepared, and the limit on those ought to be two more. The nice part about delaying them is that a committee might request more on research, more on teaching, more on something else, or a combination.

For what it’s worth, as someone at a non-elite R1, I read files backwards: writing sample (which usually doesn’t have the name of the institution or faculty mentors on it) first. If it doesn’t impress me, the rest gets a pretty cursory look; no recommendation can offset a writing sample I think is weak. That’s the argument for blind review — but I think a purely blind review process risks committees tilting toward work that fits into their pre-existing frames of analysis. So my compromise position is two letters with an option for two more.

Michael G. Cronin

On recommendation 4, it might be worth noting that in the Irish and British systems it is standard practice to only request names and contact details of referees at the initial application stage. Letters are only requested from those nominated by the applicants shortlisted for interview.
As a (relatively) recent graduate who has been applying for posts on both sides of the Atlantic over the last few years I have noticed this difference, and have indeed felt rather awkward at times asking people to write letters in support of some applications, to US institutions, where I know my chances of being shortlisted are slim. In short, as somebody applying to posts rather than writing such letters, this recommendation seems very sensible.

Zachary Zimmer

I support George Justice’s suggestion of cutting the Gordian knot of letters altogether when it comes to the hiring process. For those candidates who advance to a semi-finalist or finalist pool, the search committee can speak with the candidate’s references (as listed on the CV) directly. As others have outlined, a great LOR may have little to no impact on a candidate’s dossier, while a throw-away line or sloppy proofreading might eliminate a candidate from contention for arbitrary or unjustified reasons. That is to say: a LOR may be more likely to be ignored or actively hurt a candidate than to advance his/her case.

In the meantime, Matt Kirschenbaum and George Williams have an excellent set of guidelines for requesting LORs. I distribute these to all my advanced undergraduate and graduate advisees each fall. Perhaps an MLA best practices document could include similar recommendations.

Letters do seem to be more valuable in promotion and tenure cases, for reasons mentioned by Elizabeth Freeman. LORs also serve an important purpose in grant/fellowship applications, where they can document institutional support for a project, demonstrate the need to access to a particular archive/research site, or attest to a project’s potential contribution to a developing field of study. That said, LORs do not serve any of these purposes in the first-round of a job search.

One final observation: I think the MLA should recommend against using private dossier services such as Interfolio. These “services” charge exorbitant fees, considering that many recent PhDs are applying to dozens of positions (I applied to nearly 60 postings the year I defended my dissertation, including t/t, visiting, fellowship, and postdocs). Furthermore, these companies privatize and monetize what used to be shared responsibilities between an Advisor, a Director of Graduate Studies, and a Graduate School office. As most universities move to online applications, the main appeal of a digital dossier service have become superfluous, as candidates already submit digital files of every single part of the dossier save official transcripts. (Interfolio currently charges $6 US to upload your digital materials to an online application portal; that is, $6 for EACH job posting.)

–Zac Zimmer

Karen Newman

Limiting the number of letters, both for job applications and for tenure reviews, another area where the number is escalating, is important and worthwhile, but I think the number should be a range, not a rigid figure since, as a comparatist myself, I recognize the need for letters that address different areas or fields. I also don’t think two at first, and two later, is feasible. Students prepare a dossier made up of letters from their committee and other appropriate faculty members, and cherrypicking just doesn’t make sense–not having seen the letters, how can a student decide which to send initially, and which to send later? I think it does make sense to request only the names of recommenders, and then to ask for letters only for those the committee has decided to pursue, but I don’t feel strongly about it. Guidelines about the number of letters, and their length–I also agree that we spend too much time/space re-describing the candidate’s work–would be helpful. 3-4 letters, 2-3 pgs. We have experimented with various ways to insure faculty have sufficient time to write letters–usually by requiring that the dissertation director review and approve the CV, dissertation abstract, teaching philosophy and sample application letter in advance–by August 15–if the student expects support from the department in terms of placement.

I often write letters that don’t address a student’s teaching, because I have no knowledge of his or her teaching, and it would be difficult to obtain it. Our students rarely teach independent classes of their own, but instead are TAs, and I would find it awkward to visit another colleague’s section–others are often in a better position to address teaching than I, but I understand this might vary depending on the teaching circumstances at different universities. Though I appreciate the idea behind “blind” applications, it would prevent efforts to contact colleagues elsewhere when a student has applied for a job to help him or her “get off paper,” and I just don’t think it is feasible–how avoid any identifying tags?

Mary Beth Rose

I find everyone’s comments very helpful. In truth I am not sure the number of letters should be limited below three, just to give each candidate a maximized chance in case one of the letters should be odd. But I am completely on board with everyone’s suggestions about limiting the length of the letters, which should be no more than two pages. I have two rhetorical problems with the ways letters are written. First as many have said already, we often find ourselves repeating – – and sometimes even do a better because more experienced job — of describing the candidates’ dissertation projects. In some cases as a reader I almost feel as though the writers (especially the advisors) are competing with the candidates in the length and efficiency of their descriptions. Second is the well-known problem of hyperbole, which it’s difficult to curb. But both of these elements – – overlong descriptions of projects and exaggerated claims – – can do the candidate a subtle kind of harm. The economics of the job market are so bad that I know we all want to do anything we can do to add to a worthy candidate’s credibility. So straightforward and to the point seems to me a more helpful, because more persuasive, style of letter.

Margaret W. Ferguson

I’ve had two recent emails I want to share with you. Many thanks for taking the time to comment on my Letter of Recommendation proposals! One email, from Elizabeth Crachiolo, provides a link to a Chronicle of Higher Education article from Sept. 1, 2013, which contains several statements about letters of recommendation from the job candidate’s perspective. Entitled “Requests from a Pair of Job Seekers,” and written by Anne-Marie Womack and Rebecca L. Harris, the article is at http://chronicle.com/section/Home/5. If you can’t access it through the web and would like to see a copy, please email me at mwferguson@ucdavis.edu.

The second email is from Laura Stevens, Professor of English at the University of Tulsa. Here is her statement:

Thanks to Margaret Ferguson for tackling this important issue. It would be wonderful to have the MLA take some steps to halt the escalation of reference letters’ length, number, and rhetoric. It also can’t hurt, I think, to have more attention called to this form of invisible labor in our profession.
Like George Justice and John Stevenson, I too dream of a day without letters of reference, at least not the many and very long letters that are our current stock in trade. Still, I honestly have found letters to be helpful in hiring, and at any rate I think incremental changes are more realistic.
I have a few suggestions in response to the draft:
1. I wonder if the document might be divided into recommendations for students or recent Ph.D.’s seeking letters, faculty writing letters, hiring committees, and university administrations.
2. For recommendations to administrators, I wouldn’t mind using this discussion as an occasion also to assert the importance of having as little administrative interference in faculty hiring as possible. For example, although I had uploaded an electronic letter, I once had one university insist that I fax them a signed copy of my letter after someone I was writing for became a finalist, due to various regulations. This seemed silly to me, it was a pain in the neck, and I know the faculty didn’t come up with this rule. A standard of minimal administrative interference would include giving hiring committees some leeway in deciding when and how to seek and evaluate letters of reference or make reference calls. It also would be great to have online systems for uploading letters that were set up for compatibility with dossier services like Interfolio. [The hassle of uploading all these letters by the way, may be an even bigger issue for grad school applications.]
In communicating with administrators, I think it would be helpful to join forces with the AHA, APA, and other liberal arts professional societies, encouraging them to develop similar recommendations.
3. For recommendations to students, I’d like to see more detailed guidelines, noting for example that they should be prepared to provide their advisors with the information that’s needed to write a solid letter. It would be helpful for the MLA to advise students (and empower them) regarding the delicate issue of when and how to remind their advisers to turn in their letters. Students should also understand that few things look worse than a letter with an old date, so from year to year they need to remind their advisers to update their letters.
For what it’s worth, I’ve placed a section on my (currently outdated) personal web page, for students seeking recommendation letters. I find that it’s cut down on my workload a bit, economizing on my communications with students seeking letters and making them take on some the responsibility for organizing information:
I copied some of this language from Toni Bowers’s web page at Upenn.
4. For recommendations to faculty, why not set a standard that at the very outside, a reference letter should not be longer than a student’s letter of application? This has been stretching a bit recently, but still I think 2 pages is something of a standard for entry-level positions. Also recommend that faculty be open to having their letters read – and proofread — by colleagues before they go to hiring committees.
I don’t know if we should recommend that all faculty observe their students’ teaching, as I often read dossiers with one letter focused on teaching while others focus on scholarship. This seems to work well. Teaching observation from all letter writers is desirable but perhaps not practical. Rather, suggest that dissertation chairs or job officers make sure at least one of the students’ recommenders comments in detail on teaching, while at least one focuses on scholarship.
5. For hiring committees, a recommended limit of 2-3 letters for assistant professor positions sounds good. Also exhort hiring committees to consider their process carefully, and determine whether they really need to see all letters right away. I do like the idea of moving the letters to a later stage in the process. There could also be some flexibility here for some hiring committees, so that after an initial screening one letter is sought, presumably from the dissertation chair, with a full dossier reserved for finalists.
6. I agree with those in this discussion who also have called attention to outside evaluations for tenure cases. There seems to be a great deal of redundant labor surrounding the process, with the time and energy of our most prominent scholars unduly taxed. Perhaps the MLA could make some recommendations on this also.
7. Another possible compromise for a new standard: Job applicants could routinely seek 1-2 letters of reference, but supplement with a short list of people who could be called for additional comments.

Margaret W. Ferguson

Dear Participants (past and future):

Colleen Flaherty has written an interesting article for Inside Higher Education about the LoR proposals and your comments on them through the third week of August. She interviewed me and several of you for her article in addition to quoting from (or paraphrasing) what’s written on this site. If you’re interested, here’s the link: http://www.insidehighered.com/news/2013/09/04/mla-explores-ideas-streamline-letters-recommendation-academic-jobs

Jennifer Wicke

As someone whose life has often been made a misery in the fall semester by an onerous number of letter requests, I should fall into the George Justice camp, but much as I admire that clean proposal, I am not at all persuaded that the absence of letters is a good thing for all but the Ivy League letter-headed. I have been told too many times to count that a letter of mine has been persuasive or important in a candidacy–at various stages, not just the final push. We also need to remember that it’s crucial for candidates to be empowered to get their 3-4 letters ready to go in a given job season, rather than having to play catch-up later. The fact that so very many jobs now come in later in the form of visiting assistant professorships, last-minute replacements, or the new lectureships now replacing some t-t hires, means that search committees often go back to their files, even after the MLA interviews and sometimes after the call-backs, in order to re-assess candidates–and the letters are often the best guide. I do agree that people need to be encouraged to write their best letter for someone, and I can’t imagine that those in a position to know a doctoral student well wouldn’t already be commenting on teaching and collegiality, rather than simply addressing the dissertation. Marguerite Waller makes a fine point when she mentions the role of rhetoric in our own letters; the perfunctory, standardized, and extremely brief letter is going to do a grave disservice. Letters or 3 pages or so are not necessarily “bloated,” especially when they contain the very attention to teaching and administrative aspects of a profile that has been noted as a key feature of the best letters. I think keeping things open to 4 letters is crucial, given the complexity of present fields, but it seems great to ask for 3-4 as a limit. Places can be encouraged to run searches where the first cut is made only with names of references, before those letters are actually downloaded, but let’s face facts that since there’s no way to know in advance whether a candidate will make a short list or not, the letters are going to have to be ready to go to interfolio nonetheless. What might make a real difference is having placement directors remind faculty at the beginning of job market season what goes into a strong letter, and also having the writing of these careful documents be recognized as a critical form of academic labor that is currently unacknowledged. With all the stringencies of the job market I wouldn’t want to err on the side of failing to give job seekers any possible additional chances to become visible. I find the cavil that people only look for “names” of the letter writers really cynical and somewhat anti-intellectual; if people in a field aren’t in some sense reproducing important aspects of that field in the next generation of scholars, why are we doing this? That approach, to me, gives ammunition to those who are eager to see the humanities disciplines as less rigorous and less worthy in their research. Finally, to me these letters are something of a sacred obligation. That isn’t always what I am muttering under my breath as I complete one, but it nonetheless is the case. I’m all for applying standards that are supple and sensible, more stream-lined and reasonable, and surely no one anywhere should have to collect more than 4 letters at most, but at this juncture turning the letter writing process into a one-page exercise in boilerplate strikes me as less than generous.

Paula M. Krebs

My own institution does not request letters of recommendation at all. The search committee phones references for the top candidate only, before deciding whether to make an offer. How do recommenders feel about being phoned about candidates? Is that easier? More difficult? An added burden? Would you recommend such a system as a replacement for letters?

Comments are closed.