Wednesday, October 14, 2009

Three Letters: At Least 3, or Only 3?

This issue came up in a side conversation in a previous thread, but I thought it warranted its own discussion. Anon 12:55 asks,
Letters of recommendation: they ask for 3. Sending 4 instead: always a bad idea, always a good idea, something in between? Elaboration appreciated.
And I said,
I always thought that the rule for letters is "as many as possible." I did a little research, and I don't think I know anybody who sends just 3. I was told that getting more and better letters, particularly from people who are not directly invested in your success, was one of the best things a candidate can do to improve her chances of nailing a job.
But anon 3:28 says,
I also heard the opposite advice -- they ask for 2, they want the strongest 2, and anything further would be diluting. that said, I always send (what I assume to be) the strongest x (2 or 3 have been small numbers I encountered recently) plus a teaching letter which I note in my cover letter is sent with that intent.
But I've never heard of that. Plus, I don't have any idea how to rank my letters from strongest to weakest. I've never read my letters. Plus, I'm not the one who sends them out. Plus, I don't even think there is even any provision in my Ph.D.-granting program's placement procedures for this sort of thing. And as if that wasn't enough, it would make the department secretary whose job it is to put these files together that much more involved. So I say, as many letters as possible is the right strategy.

--Mr. Zero


Anonymous said...

Your advisor could try to rank them for you, but I think the rest of your points are legitimate, especially the departmental administrative staff (one person in my PhD department) would probably quit if any of us made that request. This person is overworked as it is.

Anonymous said...

This is something I've thought about in the past so let me share my experience.

I've held the same view as you in the past (send as many letters as possible), but last year I only used my top 3 refs who were on my dissertation committee (unless I felt one of my other refs had some special influence for a particular job), and I ended up getting a TT job (it was my third year on the market, so I had used more than 3 refs for 2 years prior; and my other refs were from other instutitions than my own).

I'm not sure if landing a job was due to my change in strategy (that could be due to anything), but my rationale was that it was better to keep my letters as strong as possible (i.e., not to dilute them). In my interviews in previous years, noone ever mentioned anything said by anyone beyond my top 3 letter writers, so that played into my decision. But I should qualify my story by saying that my different strategy in sending letters didn't lead to any change in the number of interviews I got in previous years (although the market was a bit worse last year).

Overall, I'd suggest sending a reference beyond the top 3 only if they know your work quite well, and you think they will write you a letter as strong as your top 3 (it might be worth asking them candidly how strong their letter will be). Otherwise, you might be just diluting your refs. As you say, it's impossible to know what they'll say, but I'd take familiarity with your work to be a good gauge.

This might be a question best posed to someone who's been on a search committee, but I'm guessing there will be disagreement on this issue. I know that people on SCs hold very different views on the relative value of letters vs. research achievements (at some places, a lot of weight is placed on letters, which seems a bit strange to me given that letter writes vary on how effectively they craft their letters).

Dr. Killjoy said...

Wow, Zero's potentially career-killing advice brought me out of comment retirement.

I know someone who had accumulated ten letters (six of which were outside her department), and so thought she was some hot shit. Then she had the placement director vet those bad boys, and guess what, two were comparatively mediocre and may have called the other stellar letters into question and one was nothing less than pure fucking poison.
She of course sent out the stellar seven instead of the torpedoing ten and lived happily ever after.

Letters are perhaps the single biggest aspect of your dossier, and just one less than awesome letter can torpedo your application. So, I don't care how many letters you may have, if you send any of that shit out un-vetted, then you are a fool.

Moreover, I guarantee that the folks at the top programs have placement directors that vet everyone's letters. You must, must, must, get all of your shit looked at before you send it out; you have no idea how often even a single awkward sentence can transform an otherwise stellar letter into a giant red flag.

Vet that shit!

Anonymous said...

(Thanks for noting the department secretary angle. I've not followed the advice from the previous post, sending my complete dossier for each job that I am applying for by mail, in part because that makes it easier on our single overworked department secretary.)

Mr. Zero said...

Hi Killjoy,

I thought it was obvious that you should only send out *helpful* letters. I thought the question was, if the ad asks for 3, and you've got 4 that it would be good for the committee to see, how many should you send? I did not think I would be interpreted as saying that more is better no matter what. I was trying to say that more is better if they're good.

Anonymous said...

I agree with Dr. Killjoy on this one. Get, then vet. All a placement director needs to do is indicate which ones are in and which are out and indicate rank.

Anonymous said...

Here is a related question. I have three people listed on my CV under "references" but they are not my letter writers. Is that wise to do at all? ( Two of them never read my dissertation, but I took some grad seminars from them and did well etc.) Also, how common is it for the hiring committees to contact those CV "references" who are not letter writers?

zombie said...

Gee, if only the placement director at my school did anything remotely like help jobseekers and vet letters. If only I could get my entire committee to write their damn letters. (Actually it's just the one of them, and I had the same problem with him last year, and missed a bunch of deadlines waiting for his letter -- not again!) Therefore, if only I had more than 3 letters. Heavy sigh.

Anonymous said...

I was told your references should be familiar with your work. Period.

Anonymous said...

I am the placement director at a top 30 dept with a good placement record. Here is my opinion: more letters is not necessarily better. 3 research letters + 1 teaching letter is the norm (and less than 3 research letters looks very bad since most thesis committees have at least 3 people, so 2 research letters implies that one committee member didn't think well enough of the work to write a letter). 4 research letters + 1 teaching letter is ideal. 5 research letters is OK but only if they're all really good and one is from someone outside your dept. More than 6 letters total is overkill. Having someone (sensible) vet the letters is crucial. You should never list a reference on the CV who hasn't read your work and written you a letter unless you're a senior candidate.

Anonymous said...

I'm the anon 5:05 who posted a question about references. Many thanks. Valuable advise.

Anonymous said...

Anon 8:08 is spot on. I've served on search committees and there is such a thing as too many letters.

Let me add that getting hotshot philosophers to write you letters is not an unqualified good. If the hotshots don't really know you well, their letters won't be as strong. A really great, enthusiastic letter from a less well-known advisor that knows you well is much more impressive than even an reasonable strong letter from someone who you clearly asked primarily because they had a recognizable name.

At least that was my reaction.

Anonymous said...

I am a placement director at a top 500 dept. with a horrible placement record. Here is my opinion: you need a note from your mother and a physical by a board-certified chiropractor.

Anonymous said...

Is a "teaching letter" required? What if two of your research letters also address aspects of your teaching?

Anonymous said...

"Is a "teaching letter" required? What if two of your research letters also address aspects of your teaching?"

"Required" is too strong, but a good teaching letter can be really helpful, especially if one is applying for a job at a so-called "teaching institution." Moreover, a strong "teaching letter" will often be taken more seriously or given more weight if it comes from someone who is not at your Ph.D. granting institution. This is because your home institution stands to gain something from your finding employment, but a place where you have done independent teaching has nothing to gain (in fact, if you are hired they will actually lose your good services).

Anonymous said...

All of this is very instructive, but of course not everyone is applying straight out of grad school. How do search committees' expectations change (if they do) when one is applying after a year or two in a VAP or TT position?

Anonymous said...

Re. "You should never list a reference on the CV who hasn't read your work and written you a letter unless you're a senior candidate."

I have had a couple of VAP positions and would like to list the department chairs as references for teaching and collegiality. I don't think I want to add more letters to the 4 research and 1 teaching letter I already have. What's wrong with listing department chairs on the C.V. (with their permission, of course) even though I am not asking them for letters? On my C.V. I note which of my references have written letters and which should be contacted for further information about my performance as a VAP.

Anonymous said...

In response to 8:18.

Sorry. I didn't mean that teaching letters shouldn't be listed on the CV (so shouldn't specifically have said that listed references had to have read your work.) But I think that you shouldn't list a reference on a CV who doesn't already have a letter in your file. If you don't want to ask the chair to submit a letter, then just indicate in the cover letter that s/he is willing to be contacted.

I emphasize this only b/c when people see the reference list on the CV, they expect to see a corresponding letter in the file. Just as they expect to see a letter from everyone on your thesis com. In my opinion, that is what the reference list on the CV is *for*, namely to signal who exactly will be sending letters of recommendation.

Anonymous said...

Did it ever go well for you when you answered the 6th of 5 essay questions when told to answer only 5?
As a search chair I see it the same way.