Given the current DEFCON 1 level of anxiety about recommendation letters, and my belief that more transparency is usually a good thing, I thought it would be informative to assemble a word cloud of about 10 years of recommendation letters written for undergraduates applying to graduate schools or fellowships (with names and schools omitted).
As you can see, the top three themes are “research”, “project”, and “work” indicating that the letters are highly weighted towards what students have actually accomplished, rather than just pure intellectual firepower. The cloud for graduate students applying to postdoctoral positions would be more biased towards words like “thesis” and “papers”, but would again emphasize what one has done, not how intelligent one is.
So, if you are a student thinking about graduate school, you should make sure you get involved in research, and, when you do, make sure you get something done!
January 20th, 2009 at 5:21 pm
I wonder to what extent this is just because you use a wider range of words to describe promise. I see a lot of “excellent,” “remarkable,” and “extremely” all over the word cloud; maybe the weights would’ve been different if, like (e.g.) me in introductory Latin, you knew exactly one adverb… or, conversely, if you alternated “endeavor,” “undertaking,” and “venture” for project.
January 20th, 2009 at 5:44 pm
Somewhat, but those are mostly modifiers to research-related metrics. There are not all that many synonyms for smart, and “bright”, “smart”, “intelligent”, etc, make up a rather modest fraction of the cloud’s area.
January 20th, 2009 at 5:46 pm
The post linked to “get something done!” was excellent. In that context I would also like to point out that people like Newton and Einstein worked extremely hard! They would both regularly work non-stop through the night often forgetting to sleep or eat food. Yes, they were geniuses…but they “got something done” because they were extremely hard workers. I think this aspect of Newton or Einstein is often overlooked due to their inherent genius.
January 20th, 2009 at 5:52 pm
Since we put what we’ve done on our CVs (list of publications, talks at conferences, etc.), why do we need recommendation letters?
January 20th, 2009 at 5:56 pm
CI — I think you’re largely right at the more senior level, but undergrads and grads don’t really have enough of an independent track record to judge based solely on the CV. Recs are probably also needed when the majority of papers are on large collaborative projects, where it’s hard to know who did what — in that case the letters help frame the candidate’s contributions.
January 20th, 2009 at 6:35 pm
Getting into graduate school = postponing your eventual departure from academia (meanwhile, investing 4+ prime years of your life to improve your professor’s career)
January 20th, 2009 at 6:47 pm
Any interesting/significant differences between the word clouds of recommendation letters of successful and unsuccessful applicants?
January 20th, 2009 at 7:47 pm
Graduate: If you’re going to grad school, aren’t you pretty likely to end up in academia forever? My feeling is that most people who go to grad school aren’t exactly looking to depart from academia.
January 20th, 2009 at 7:51 pm
Julianne: I’m sure you’re right, esp. since “promise,” “potential,” and “talent” aren’t conspicuous either. On the other hand, both “research” and “project” have a paucity of synonyms, as I found out when applying to grad school. Incidentally, are all the “radios” part of the phrase “radio telescope”? A curiously large number of physics majors/grad students seem to have something to do with radio stations, but I don’t suppose that’s what’s going on here?
January 20th, 2009 at 9:07 pm
Where are words like “bonehead,” “sniveling,” and “cretinous”? What about “tawdry,” “scandalous,” and “lawsuit”? The letters I write are littered with such vocabulary. I think it spices things up.
January 20th, 2009 at 9:13 pm
Sean — I’m too busy swilling my chai latte and mocking the weak applicants to reply.
Sarong — the UW undergrads single-handedly organized the financing, construction, and operation of a radio telescope on the roof of a campus building, so I mentioned students’ involvement in that project, when appropriate.
January 20th, 2009 at 9:30 pm
You guys just made me drop my monocle right into my claret!
January 20th, 2009 at 9:47 pm
LOL! I’m impressed that you guys were able to hold back on the arrogant sarcasm for so long! I was curious to see how long you could continue this “caring” discussion with the weak applicants.
January 20th, 2009 at 10:01 pm
The pity is that I actually do care. But when someone accuses me of drinking chai (EW!), that’s over the line.
January 20th, 2009 at 10:31 pm
HEY!?!, what’s w/the ripping of chai lattes all of a sudden?
Some of us need a bit of variation in our daily espresso intake.
January 20th, 2009 at 10:51 pm
I am deeply disappointed grad student. This field has no room for chai sympathizers like yourself.
January 20th, 2009 at 11:26 pm
Glad to see some of that charming arrogant humor back on this blog. The whole reason I went into theoretical physics is to become smart enough to justify my arrogance which comes so naturally
to me. Isn’t that the whole point of learning all this stuff about our universe? Now when I sip chai with snooty people, I can bullshit with authority.
January 21st, 2009 at 3:46 pm
Alexander: Of course most people go to graduate school aspiring towards a distinguished academic career, but you’re more likely to complete your training for an academic career and then find yourself at a dead end (unless your definition of an academic career includes adjunct community college faculty). Or, maybe you’ll face the dead end after one or two or three postdoc positions. To paraphrase the advice once given to me by a professor at a major research university: you are unlikely to become a professor at a major research university unless you are “the best student that your professor ever encounters in his or her career.” Do you think you might be that best student? If so, go for it, and good luck to you.
January 21st, 2009 at 4:31 pm
If you are smart enough to seriously consider a career in academic science, it shouldn’t take that long to figure out what Graduate cited above is true, on average. Assuming the number of positions stay constant, over his or her lifetime a faculty member needs only one successful advisee to replace that position when he or she retires. Anything more will produce an overabundance of “postdocs,” and many professors advise more than one grad student at once. Therefore it is a bad sign when there is an astonishingly high number of people hired at the postdoc level in your field of choice.
January 21st, 2009 at 6:28 pm
I don’t think it’s very useful to see wordclouds of all recommendations. I mean, far fewer than half of all applicants are accepted into graduate positions for the places where you have access to the letters, right? That means unsuccessful letters are weighted more heavily than successful letters in your model.
Let’s see what successful letters have in them! Or, better yet, let’s see what successful letters have in them once you trim out the top and bottom 5% of words or something. Who cares about seeing a word like “research” or “work” in the cloud? Wouldn’t that show up frequently in almost all applications, successful or not?
Let’s see what adjectives show up, too (of course, getting rid of numbers (which are adjectives) and other things like that)!
January 21st, 2009 at 9:35 pm
Kyle G — All of the students I wrote for were successful in applying to various programs/fellowships (i.e. the letters were letters I have written, not received). I gracefully bow out of writing letters for students who I can’t back fully. Not every student got into every program they applied for, of course, but all found places they were happy with. So, the word cloud is indeed for successful applicants. And also, I don’t think that “research” and “work” would actually show up in all applications, since a significant fraction of letters are from people who only supervised a student in coursework. Those letters tend to be short, since there’s usually not much to talk about (grade in the class, performance on problem sets vs labs vs tests, interactivity in class, and that’s about it). However, the strongest letters will always contain involved discussions of research, since in that case the student performed a significant enough bit of work that there’s something to talk about!
January 22nd, 2009 at 12:19 pm
The best description I ever read in a reference letter was “gormless troglodyte”.
January 22nd, 2009 at 12:42 pm
Thus proving that British letters are indeed different than those in the US.
January 23rd, 2009 at 3:36 pm
Or, maybe you’ll face the dead end after one or two or three postdoc positions.
In this case, what are graceful exit strategies?
January 24th, 2009 at 12:21 am
Good question, devicerandom. I wish I had an answer for you (I can think of many exit strategies, none of them “graceful”). Maybe those on this blog who are encouraging grad school applicants should try to answer your question.
January 24th, 2009 at 4:53 am
Well, start from the non-graceful ones. Since I am a fresh molecular biology/biophysics postdoc, I’d like to know them.