Let me preface by saying that I don't think reference letters should be done away with entirely. Someone should be able to speak to an applicant's qualities, skills, productivity, and character beyond the applicant him/herself. Rather, I believe that asking for three (or more) letters from references who have acted in a supervisory role to an applicant (i.e. professional references) is excessive. My reasoning is as follows:
How many graduate students, post docs, junior faculty, etc, have worked with three or more supervisors/advisors/professors closely enough to write a detailed letter in a relatively short period of time? One might have had three such references over the course of their career, but hiring committees are generally not interested in references from 10+ years ago. They want references who have worked with an applicant more recently (and rightly so). Additionally, hiring bodies don't tend to look favorably on "peer" recommendations. Where are grad students, post docs, new researchers supposed to get letters?
Thus, I think one or two detailed, professional letters are both sufficient and more realistic. However, I'm curious what others in the academic community think about the matter.
I think a more fundamental problem of recommendation letter is that it is an almost unbreakable step somewhat like Catch-22：
to get into area B, you need one or two or three recommendations from area B;
to get recommendations from area B, you need to get into area B.
to get a good job, you need recommendations from a good employer;
to get recommendations from a good employer, you need a good job;
to get a good job, you need good skills;
to get good skills, you need to pay to learn skills and apply the skills;
to pay to learn skills and apply the skills, you need a good job (or perhaps a fortune, or perhaps to become Federal Reserve Bank).
Xinyuan gu Yes, there are some inherent difficulties when trying to gain the skills, references, and experience one needs. Though, I'm not sure the situation is quite as rigid as you propose. With most careers, positions can be (roughly) characterized by levels (i.e. entry, intermediate, experienced, expert), each requiring the requisite skill and experience for the position. Thus, the majority of people have to start at the bottom and work their way up - gaining what they need as they progress.
My concern with this situation is that it is a long-term endeavor. For many, it can take several years working in different positions. However, references are expected to be current (or at least relatively recent), and I don't know many applicants who can amass three or more professional references in a short period of time - and if they can, it is a red flag. It might suggest that the applicant spends a very brief time at each position before moving on to the next (i.e. not likely to stick around long-term).
To Anna Soska
's point, I fear a lot of talented applicants fall to the wayside due to this stringent practice, which could inadvertently contribute to discouragement and burnout. In addition, bear in mind that the goalpost keeps migrating further back for applicants. Many years ago, my undergraduate advisor and professor of mine said to me, "Thirty years ago, it was typical that a student just helped some research faculty member with their research, and it didn't really matter in what topic. I would never have been accepted into a graduate program with the requirements today".
Blaine Tomkins Thank you for the comment. I certainly agree with you on the point that everyone starts basic. My point is a bit different and perhaps I did not extend it as much as I could.
The fundamental problem of recommendation letter, when placed in a buyer's market, is likely to result in class solidification, or in other similar forms of solidification. In a buyer's market, resume readers are flooded by resumes, and disadvantaged background leads to permanent disadvantage. This means a looser recommendation rule in a buyer's market would further flood resume readers, while a tight recommendation rule might fail in the long run.
Xinyuan gu Interesting points. Yes, requiring 1-2 letters rather than the traditional 3 will likely increase the applicant pool - thereby requiring more time and resources on part of hiring committees. However, there are more effective ways to reduce the applicant pool than using an arbitrary count of references. (What if the supply continues to grow while the demand remains stable, do we transition to 4 letters? 5?). To me, it makes more sense to place more stringent criteria on experience, productivity, skills, and applicant fit to the specific position. That is, things applicants have more control over and can improve.
I may get some flak for this, but I'm not very moved by the "it will take too much time and resources" argument. If you want to fill a highly competitive, long-term position and you want the best candidate, it should! Make it a priority, not a subsidiary task. Furthermore, cutting down the number of letters should have the effect of highlighting quality over quantity. Two strong professional letters is, in my opinion, much more desirable than (1) One strong professional letter, (2) One weak professional letter, and (3) One character letter. Which is also a waste of hirer's time.
Just for illustration purposes. A position posted to HigherEdJobs just today:
This raises another issue on the topic of reference letters. Where publications, awards, skills, and other accolades speak for themselves, a reference letter boils down to the subjective opinion of the writer. I've heard many stories similar to yours about pernicious or unfavorable letters.
I read a shocking story in the Chronicle of Higher Ed last year from a Ph.D. recounting their experience in graduate school, where their advisor threatened to withhold a reference unless the grad student added some of the advisor's colleagues (who were in no way involved) as co-authors to a publication. I'd like to believe this kind of unprofessional and underhanded behavior is rare, but I've been around academia long enough to know better.
Blaine Tomkins, Anna Soska
Thanks for the very interesting comments. I certainly wish there could be a system better than the current one from which I could be benefited.Some social math might be very challenging, though. For instance, I have 10 posts, 400 applicants. If I somehow "objectively" rank all the applicants based on skills, experience, and recommendations, I will almost for sure to get more than 10, let's just say 15, that count as the best among 400. This is why, the most reasonable thing for me to do, is to set the common factors of the 15 as the baseline, and I will always have plenty of daredevil applicants that waste time to play the game due to the blindness to the baseline. What would be the common factors of the 15? Proficient in-need skills, a lot of experience from the past, and THREE perfect recommendations. Why would this be the case? Because the three common factors are interlocked.
This indicates that requiring only one or two recommendations would not change the ranking of the applicants, whether the number of the applicants is increased or not, unless the evaluation is different. This also indicates, if one or two daredevils survived the common-factor screening, it must be the case that the daredevils survived because of something irrelavent to the common factors or because of something significantly overriding the interlock of the common factors.
Reforming the evaluation would be very delicate, for it involoves maintaining the fairness of the hiring process and counter-engineering the accumulated social unfairness. So one might observe that some daredevils are in fact very rational, for if there exists any place on this planet, where the hiring process is a little bit smarter than the common one, perhaps smart enough to maintain fairness and to counter unfairness, the place could not be common.
And so I agree on the point that, how to evaluate is critical.
Making references open to applicants may be helpful, for applicants would better know their chance in surviving the common-factor screening.
Anna Soska
No...I mean the current academic recommendation system must have come from a time in which there were no global scale application pool. Unfortunately, I can't find the origin of the current academic recommendation system via google. Note that in the time there were no global scale application pool, the nation/city/area-based inequality did not matter.Are we on the same page?
Xinyuan gu I see what you're saying, but there are a few issues with your example.
First, I don't believe it's possible to "objectively" rank applicants with respect to experience, skills, and recommendations. For instance, experience can vary with respect to (1) amount of time, (2) the kind of experience, (3) how productive someone has been in their years of experience, (4) the role(s) someone played in their past experience, (5) how closely the work they've done in the past matches the work they're being hired to do....and there are surely many others. The same is true of skills. The point is there will always be some subjective element in such decisions.
Second, related to Point 5 above, you're missing the crucial aspect of "fit". Just because two (or 15) applicants are all qualified and with similar resumes, does not necessarily mean they would all be an equal fit for the job. I'll give you a real-life example: A few years ago I was being interviewed -along with a very small and select pool of applicants- for a college teaching position. I wasn't hired, and I asked the gentleman I interviewed with what I could improve to become a stronger candidate. He said, "There was nothing wrong with your application or interview. You were a strong applicant and definitely qualified. It came down to a matter of fit. This position requires teaching class sizes around 300, and we had an applicant who has such experience" (*I was teaching classes of 20-40 students). And that decision was perfectly reasonable.
Lastly, the harsh truth is that hiring practices are never going to be 100% fair. It's not pleasant, but it's reality. In some cases, there will be multiple, superstar applicants for a position, but the committee can choose only one. It's inevitable that qualified applicants are going to be rejected. I think we can amend practices to reduce unfairness, but it will never be eliminated entirely. I also think that, of all the formal factors involved in an application, using a strict, arbitrary cutoff for number of letters to narrow down an applicant pool is the most unfair option.
Blaine Tomkins Thank you for the very interesting example. To give an "objective" evaluation of experience is indeed very difficult even for humans.
I think the fit problem you mentioned is due to chance, and the due-to-chance problem can be compensated if there is a sufficiently large job pool, in the sense that a qualified accessing perfect information is going to get a fit job somewhere. A real-life problem is that many qualified due to chance can only find suboptimal jobs, let alone the endings for those less qualified.
Since required only one or two highly relevant recommendations could increase the chance for an applicant to enter a very small selected applicant pool, perhaps fewer recommendations do increase the chance for the very small pool to include the best fit. So maybe you are right!
I've heard/read in passing that recommendation practices are considerably different across countries. My experience is pretty much restricted to the U.S. It would be great to hear about approaches and/or practices used in other countries. Perhaps there is something we can learn.
There is a good reason for requiring three letters of reference. With only two letters there is the possibility of a tie between a strong recommendation and a weak recommendation.
Also, nowadays in addition to a letter there is typically an online check-the-box assessment form which requires the referee to rank the applicant (e.g. top n% of students) in various categories. With only one referee, one is less able to judge whether one is getting a very one-sided or idiosyncratic assessment.
Karl Pfeifer I appreciate your sharing your thoughts on this question. From the hiring side of the table, the rationale for requiring three letters is fairly clear. However, my goal in raising this question is to address the issue from the perspective of the applicant - primarily the inherent difficulty of most applicants to obtain three strong and thorough letters of recommendation from individuals serving in a supervisory role over the applicant in a relatively short stretch of time. This is simply not how the career path of (most) aspiring academics and/or researchers progresses. What advice do you have for applicants?
- Badges
- Science topic
- Similar topics
- Computer Science
- Research Topics