should i sacrifice quality for a diversity of letters?

[Guest Gnome Chomsky]

I hope my title wasn't too vague. Anyway, here's my situation: I'm a linguistics undergrad. I've been in undergrad for over 4 years and have completed all my linguistics requirements. So the past semester I've been exclusively taking math and computer science classes. I wanted to 1) challenge myself, and 2) give myself opportunities in fields such as computational linguistics. To be exact, I'll be getting a minor in computer science and applying to computational linguistics master's programs for Fall 2014. 

 

Now, I've never had a problem establishing relationships with my professors. And I'm definitely not a "kiss ass." I have 4-5 professors who have promised me very good letters of recommendation a year in advance. I'm definitely not worried about getting 3 letters of rec from linguistics professors. The problem is since I'll be applying to a somewhat interdisciplinary program (half the faculty is typically from computer science and the other half from linguistics) I was thinking I should provide a diversity of letters of rec. In other words, if a school requests 3 letters, I'd like 2 from linguistics and 1 from comp sci (or vice versa).

 

The problem is I just started my first computer science class this summer (which is taught by a TA), and by the time I submit my application I would only have taken 2 computer science classes. I'm not going to get a letter from a TA. And I doubt that within a month or two of the Fall semester that I could establish such a great relationship with my professor (provided that he/she isn't another TA) that he/she would be willing to write me a high-quality letter. Once again, I've never had a problem establishing relationships with professors, but I've also never tried to force it in a fixed amount of time. 

 

I guess my question is should I sacrifice the quality of 1 of my 3 letters by getting the generic "This student received an A in my course" letter from a computer science professor? Or should I just go with 3 high-quality letters from linguistics professors? And if I do go with 3 linguistics professors, should I mention in my statement of purpose that I was aware of the fact that I didn't have letters from computer science professors but that I didn't want to sacrifice quality? I've always had a problem of being a little too real and honest. 

[fuzzylogician]

I guess my question is should I sacrifice the quality of 1 of my 3 letters by getting the generic "This student received an A in my course" letter from a computer science professor? Or should I just go with 3 high-quality letters from linguistics professors? And if I do go with 3 linguistics professors, should I mention in my statement of purpose that I was aware of the fact that I didn't have letters from computer science professors but that I didn't want to sacrifice quality? I've always had a problem of being a little too real and honest. 

 

1. No, don't sacrifice a good letter for a "diverse" DWIC letter. How is it diverse if the letter can't say anything particularly insightful about you? Get the best letters you can, it's fine if they're all from linguistics professors. If you got three letters from Physics professors (to pick a random field) that would be different, but we're talking about linguistics professors writing letters for (computational) linguistics grad programs -- those should be well accepted and appreciated by the adcoms.

 

2. No, don't mention the composition of your letters and not wanting to sacrifice quality. It's assumed that you chose your best options so anyone wondering why you didn't have a letter from a CS professor will guess that it's because it would not have been as strong as the letters you got from the ling professors.

[Guest Gnome Chomsky]

Thanks, Fuz. Yea, I think I'll be fine with 3 letters from linguistics profs. One idea I had to show my "diversity" (if I design an impressive program by the time I apply) is to submit a code as my writing sample. Even then, I think the quality of my traditional research papers (one which was for a grad class) would probably be more impressive than a program/code I wrote for an intro to computer science class. But we'll see. Thanks for the advice. 

[TakeruK]

Agree with fuzzy. Also want to say that I think it's perfectly fine to enter a diverse / interdisciplinary field (e.g. "A-B" where A might be astronomy and B might be chemistry, or in your case, computational science and linguistics) with a stronger preparation/experience in one or the other. Most of my cohorts in our Planetary Science (which is a combination of geology, geophysics, physics, chemistry, and astronomy) have undergrad backgrounds in just one (or two related ones) of these subjects.

 

One of us had an interdisciplinary degree in the wide field of planetary science, and that person says they feel it's easier to take something you know really well (e.g. physics, or chemistry) and then go and apply it to the interdisciplinary field and learn the other stuff as you need it, instead of having a wide background but no specialties. I'm not sure if it's necessarily true, since no one can see both sides (can't be both a generalist and specialist at the same time!) but it's something worth thinking about. I would personally also prefer to be really good at "A" and have working knowledge of "B" when going into an "A-B" field, rather than know a general amount of both "A" and "B".

[rising_star]

I changed fields for my master's (went from humanities to social sciences) and all three of my letter writers were professors from my undergraduate major. It was not a problem at all with admissions.

 

But really, if you read the wording of your topic title, the answer should be obvious.

[fuzzylogician]

One idea I had to show my "diversity" (if I design an impressive program by the time I apply) is to submit a code as my writing sample. Even then, I think the quality of my traditional research papers (one which was for a grad class) would probably be more impressive than a program/code I wrote for an intro to computer science class. But we'll see. Thanks for the advice. 

 

Holy mother of bad ideas, batman! 

 

The writing sample is supposed to showcase your ability to conduct meaningful research as well as your writing abilities, not to show that you can write basic code. You should submit your best paper, not some code you wrote for an intro class.

[Guest Gnome Chomsky]

Holy mother of bad ideas, batman! 

 

The writing sample is supposed to showcase your ability to conduct meaningful research as well as your writing abilities, not to show that you can write basic code. You should submit your best paper, not some code you wrote for an intro class.

Well, that's why I said if it's impressive enough. I don't know what type of stuff I'll be designing by the time I apply. Some of the schools say they encourage code for a writing sample. But I feel like I'm a much more polished (and likely will still be) writer of research papers than of computer code. I guess I'll just keep it linguistics and show them I can program when I get there. Thanks for the advice. 

[dat_nerd]

Is there a way to append code to your application without making it *the* writing sample? As a computer science major, there's no such thing as "impressive" code. There's also a lot of stylistic and "syntactic sugar" battles out there in the CS world that can make a reviewer hate your code based on your style alone. Oftentimes what one person considers "impressive" code is a jumbled mix of headache-y code to another, so be careful!

 

Even so, I don't know how much a code sample would actually matter... I mean, the CS applications don't even ask for a code sample. It's assumed you're either good at programming, or will learn quickly.