What did you do to get stellar letters?

[peacedeer]

Hi guys,

To those who got into top 10s, would you plz share what you did to get your stellar recommendation letters?

 

What is it that makes your letter writer say, dear MIT, attention plz, this is the next Mr.Krugman/Miss.Duflo, take him/her asap?

 

Your sharing will help a lot of us out there in the next application cycle!

 

More details are always welcome (How you initially approached your supervisor, in what kind of context? classes/projects/thesis/r.a and what makes you stand out?)

 

Thanks!

Peace and blessings for Japan.

[econonerdics]

First I think I benefited coming from LAC, as there is A LOT of opportunities to interact with profs becuz they have nobody else to be busy with!

 

I have been with one prof (as RA & TA) since my freshman year, plus a couple of independent studies/research papers/honors thesis. So, that letter from this prof is spectacular!

 

Next, I do well in classes, and talk to prof very often, in terms of my goal to go to grad sch, and also potential careers. Profs in my school love to talk to students! (even more so especially when I say I wanna go grad sch)

 

However, one downside is that probably profs from LAC are less well known, thus harder to push you into top 10. Also, I think a slightly more unknown LAC will suffer more becuz of branding effect.

 

If you are in a top 20 PhD granting universities, I think the best way to get stellar letters is to do well in his/her class, and go to office hours to ask not only questions about courses, but also guidance for grad sch. I think they will be willing to help!

[rthunder27]

Certainly write a thesis, if your undergraduate program gives you the option to do so to graduate with honors. Then the key becomes selecting a good thesis adviser, because this person will certainly become a natural choice to become a LOR writer. There are a lot of different criteria for selecting a good thesis adviser, such as research interests and how much time they're willing to spend on you, but you should also consider how well connected they are to the schools to which you want to apply for a PhD program.

[IMTB]

I will 'try' and answer, though there is no answer. I was accepted at a good amount of the top 10 and had 1 'famous' letter and 2 letters from known professors. I went to a top 5 UG so I am not sure if this varies, however, all I did was really take an interest first semester and got into research by my second semester. I also lived on campus in the summer and continued to help which showed dedication (even though I had internships 2 of my first 3 summers), offered to help the TAs and professor with grading and such until I was an official TA. Then when it was that time I asked for letters and all three actually showed me the letter and basically asked for my approval. I think once you gain respect they will write great letters for you and basically throw in a line that says 'X is as good as the Y students I have written a recommendation for, and have been accepted, into your program, such as A, B, C, D, and E.' It really helps if someone in A-E is attending. If a professor really likes you and believes in you they may say, 'This is one of the top 3 people I have ever written a recommendation for and 8 of my students have been accepted there.'

 

In short, hard work gets you those letters. Being willing to help conduct study sessions, grade, TA, and RA. Writing a thesis is a way to get to know people you may not know and showoff your determination, however, the former ways, IMO, work best.

[thewhiterabbit]

I think the answer depends a lot on context, in particular on the tier of school you are coming from, and how frequently your recommender writes letters.

 

When you're coming from a higher tier school, and/or your writers are writing letters frequently, they have to take into account their currency and reputation as a writer. I've heard profs described as the type to always oversell candidates, and after awhile this will filter through and admissions committees will somewhat discount their letters. Conversely, a young prof might take the approach of being very demanding or harsh when evaluating a mentee, because they are new to the game and want to preserve their reputation. It's important to remember that when you go on to grad school, and then become faculty, you will either reflect well or poorly on them. So they have a vested personal interest in writing you a LOR that reflects your true quality as best they can.

 

So I believe the LOR thing is actually more of a crapshoot than most people realize. Here are two possible scenarios:

-You work for years for someone as an RA, and do good work that gets positive feedback. But you only do relatively menial RA tasks that they give you, so it doesn't actually reveal much about your quality, except that you're a good worker. Not wanting to bug them, you never approach them to talk about research ideas or a thesis. When it comes time to write you a letter, the letter (despite years of good work on your part) might only be mediocre because your writer is unsure of your true quality (prowess with Stata not being the best metric of research ability).

-You're the top student of a well-known professor who happens to take on lots of RAs. He tends to oversell them, however (but you don't know this). He writes a glowing letter, but it doesn't have much of an effect, because he doesn't have as much currency as you would think given his stature. He tells you you'll crack the top 5, but then you don't.

 

I think the most reliable way to get good letters is to identify, if you can, the professors at your school that tend to take 1-2 undergrads per year under their wing, help them develop, and reliably help them place at top schools. Then work as hard as you can (IMTB provides great suggestions) to impress them. The problem is that there may not be professors in your area of interest at your school that do this. It is possibly worth going for profs outside your area of interest if they seem to be excellent mentors (this never occurred to me as a good approach as an undergrad).

 

If the professors in your orbit are unknown quantities in terms of their quality as mentors, or you're not sure where you stand with them, another thing you can do is cast a wide net and develop relationships with a number of professors. When it comes time to ask for letters, really try to get a frank assessment from them of the types of places they think you should apply, and whether the letter they will write you would give you a shot at those places. This is one thing I really wish I had done better (although I tried!). I found out after decisions started arriving that one of my letters was ridiculously stellar, and yet another of my letters was explicitly not top 5 material. I think this partly explains my weird results.

 

Finally, in keeping with the idea that writers have an incentive to reflect your true quality in their letters, let them know early that you want to go to grad school, get their advice on what they think you should do to get there, and keep them updated with the results of your actions (like, say, when you get that A in Analysis :)). Building up your credentials on the traditional metrics (like math courses) will reduce your expected variance in their eyes, and make it easier for them to write a great letter. I believe at least one of my letters would have been better had the results of my Analysis class been available when I was requesting LORs.

 

Finally, disclaimer: All the above is based off my own experience coming from a top tier school; I know it's very different if you're coming from a lower ranked school, and will refrain from giving advice on it since it's outside my realm of experience. :)

Edited March 16, 2011 by thewhiterabbit
typo

[fexical]

long story short, be a better slave.

[thewhiterabbit]

long story short, be a better slave.

 

I know you're being cheeky, but I actually think this is usually what *not* to do.

[Zeno]

I have two less-than-satisfying answers.

1) Show up, be enthusiastic, do what you can to let the profs get to know you.

2) Actually be one of the best students they've seen in however long. People on this forum frequently talk about how to get good letters. However, there's no prescribed process after which professors are going to automatically call you the next Roger Myerson. I feel we may want to think of it as how to put our best respective feet forward.

[WorkingHard]

These are all very good comments that I enjoyed reading as well. But I have wondered about one thing in regards to doing a thesis: since they are typically not completed until the final year (I don't plan on doing it until my last semester because I have been busy with other research, classes, etc), how does one emphasize it on graduate school applications? My assumption is that one must start the preliminary thesis work, at the latest, the semester they are applying and on their CV will be an abstract for the thesis (like job market papers appear on PhD candidates CVs). Is this true?

[vidamilonga]

My main piece of advice, which is I think not enough applicants do, is to go to seminars. Not only are they awesome and fun (and sometimes you'll even get a free lunch!), but you'll learn a lot really fast and it's an excellent way to get to know faculty. At least at my university, very few undergrads go to seminars (I think I'm the only regular), and professors will want to get to know you after you become a regular. This is a great opportunity to signal you're serious about econ research, and you might even get an RA gig out of the deal. It also gives you a chance to meet professors who don't teach or otherwise interact with undergrads.

 

In fact, if you hang out in the graduate program in your department, it's way easier to meet people than through traditional undergrad channels.

 

Also, random observations:

-Connections matter A LOT. I got accepted to both departments where my advisor used to be on the faculty; I doubt this is coincidental.

 

-If you're at a big school (like mine) realize that you have to take the initiative in getting to know your professors. Even if you get the top score in their class, that's not enough; they're too busy to seek you out to know more about you.

 

-Talk about your potential research ideas with your professors. Going to seminars can really help in developing your intuition of what research should look like.

 

-If you're looking to be an RA for someone, exploit your network of fellow students. I've recommended several of my friends to RA for my advisor, and he hired them based on my recommendation.

 

(I'm from a top 15 department, so this probably mainly applies for students in departments with good PhD programs and well-known faculty)

[Team3]

When you're coming from a higher tier school, and/or your writers are writing letters frequently, they have to take into account their currency and reputation as a writer. I've heard profs described as the type to always oversell candidates, and after awhile this will filter through and admissions committees will somewhat discount their letters. Conversely, a young prof might take the approach of being very demanding or harsh when evaluating a mentee, because they are new to the game and want to preserve their reputation. It's important to remember that when you go on to grad school, and then become faculty, you will either reflect well or poorly on them. So they have a vested personal interest in writing you a LOR that reflects your true quality as best they can.

This is well put, and I'll second it!

[ypdm]

These are all very good comments that I enjoyed reading as well. But I have wondered about one thing in regards to doing a thesis: since they are typically not completed until the final year (I don't plan on doing it until my last semester because I have been busy with other research, classes, etc), how does one emphasize it on graduate school applications? My assumption is that one must start the preliminary thesis work, at the latest, the semester they are applying and on their CV will be an abstract for the thesis (like job market papers appear on PhD candidates CVs). Is this true?

 

Yeah, I had exactly the same concern when I first met my senior thesis advisor at end of September last year. I felt there wouldn't be enough time for him to know me well before I ask him for recommendation letter. However, things worked out very nicely. First of all, I had a pretty good proposal that he found quite interesting. I tried to schedule meetings with him and discuss my progress at least once every two weeks. So things were moving pretty fast for the first semester of my senior year. Also I told him that I am applying to graduate school in our first meeting. Though I only met him several times, we talked a lot about thesis and Phd studies in general. So by mid-November, he offered to write a letter for me!

 

I think this really depends on the professors. I am really lucky to have found someone whom I feel so comfortable working with. Now when we discuss about the applications, my advisor tells me candidly that when he was writing the letter for me last year, he had to make some predictions about my research skills because I didn't have a completed work back then. So I guess a lot of the nice things that he wrote about me came from his feelings and observations of our weekly conversations.

[peacedeer]

Thanks guys. Awesome advice~

[peacedeer]

So I believe the LOR thing is actually more of a crapshoot than most people realize. Here are two possible scenarios:

-You work for years for someone as an RA, and do good work that gets positive feedback. But you only do relatively menial RA tasks that they give you, so it doesn't actually reveal much about your quality, except that you're a good worker. Not wanting to bug them, you never approach them to talk about research ideas or a thesis. When it comes time to write you a letter, the letter (despite years of good work on your part) might only be mediocre because your writer is unsure of your true quality (prowess with Stata not being the best metric of research ability).

Thanks for pointing this out. It was my exact concern~ R.A work tends to be time-consuming and employer-oriented and can hardly be a channel to show originality. I guess it really says more of your work ethics than ability~ so I take it most of you DID NOT get letters from people you just R.Aed for but more from thesis advisors/grad course professors/summer project supervisors?

[FromTheHip]

I'm going to chime in here -- I'm not sure it's appropriate since I have not had the results of TWR/ypdm/others who got into top-5 programs, but I'm guessing my letters were pretty good and I have a few points to make.

 

On RAing:

I currently work as a RA, and I got a letter from my boss. I do think RAing can be a good way to get meaningful letters -- but it really depends on who your boss is and how long you work for. My boss has always been very nice about seeing to the academic development of people who have worked for him as RAs, and also is very generous about letting RAs who he deems competent take on larger chunks of work with less oversight. No, I never came up with any research ideas, etc., but it did let me demonstrate that I understood the research well enough to work with a good degree of independence, etc (and of course, show that I worked hard). It's nevertheless true that not all RA work offers the potential for distinction, but if you do it long enough and you have a good boss, you will get your chances. And the time my boss put into my academic development definitely helped me gain understanding of the work I was doing/past and current research in the (broader) field -- which I was then able to demonstrate to him. I suspect all of it made for good material for a letter, if not quite a "his work could already be published in AER" kind of letter. So while I certainly believe that there are RA jobs which are dead-ends, if you're careful (or very lucky, as I was!) in who you work for, it can work out well. From what I've seen, professors who tend to have more experience having RAs are generally better at understanding how to help their RAs grow and show that they are ready for top programs. Additionally, if a professor has had RAs recently, you can probably contact them for the inside scoop.

 

On courses:

I think courses besides a thesis are a good (and thus far unmentioned) way of getting letters. If you're lucky enough to have access to a top econ department (and are able to enroll), graduate classes are a great way of building letters. Not only will the grades look good on your transcript, but if you can manage a top performance, there's probably a recommendation in it, too (obviously, bad grades are a bad signal that you probably want to avoid). I'd also guess that more economists are willing to pay attention to students who can advertise that they are already doing well in graduate coursework.

 

Finally, I really think Zeno has had the best point this thread: what we're really discussing is good ways to signal that you're good enough -- you still have to put in the work.

[thewhiterabbit]

I do think RAing can be a good way to get meaningful letters -- but it really depends on who your boss is and how long you work for. My boss has always been very nice about seeing to the academic development of people who have worked for him as RAs, and also is very generous about letting RAs who he deems competent take on larger chunks of work with less oversight.

 

I think FromTheHip gets it exactly right about RA positions, and this line is key. I think some advisors don't know how to do this, leading to the scenario I mentioned before, and on the other hand, some will will reward menial RA work with a stellar letter (even if they didn't actually get to know that much about your ability). The best kind of advisor is the kind FromTheHip seems to have had, who can find opportunities to give you autonomy and let you shine, and then you're able to do so.

 

This is more likely to happen when the RA position is full-time than in a term time RA position, also. FWIW, I got my really really good LOR the same way, so I definitely don't think RA positions aren't a good way to get letters, it's just you have to be aware of the nuances of it.

[kt253]

One thing that you might do, if you're at a lower-ranked or less well-known school, is seek out professors at well-ranked programs and see if they need any RA help. I've had a friend at school do this (with a Harvard professor, no less), and have also been given this advice by a few faculty members at various economics departments that I've interacted with over the past couple of years.

[treddi]

I think thewhiterabbit is going to be a great prof/advisor

[thewhiterabbit]

Thank you! I hope you're right!

 

Having supervised undergrad RAs in the past, it's a lot harder than it looks I think.

[JRav]

I'll add my two cents, just from my own personal experience. In my case, I was fairly certain I was into macroeconomics so I tried to find classes taught by people who do macro, and who also went to top schools and knew people. My department is very small and while plenty of students major in econ, only "one every two years or so" ever actually apply for grad school. I made it a point to get to know my professors, one in particular. I introduced myself after the first class and made an appointment to discuss graduate school and research opportunities. Interestingly enough, I was told that RA positions for undergrads were very rare because all of the positions go to grad students. While I couldn't be an RA, I asked if he could supervise my senior thesis an he agreed. RA positions are one of the best ways to get to know a professor, but in no means the only way. I guess this goes without saying, but if you're going to discuss research, don't sound like an idiot. Have some idea of what the professor does, who they work with, some of their papers... Also, show that you can work independently. If you need your hand held every step of the way, it's not going to help you.

 

So to sum it up, basically try to get to know your professors, especaially outside the classroom. Chances are, especially if you're in a large program, the professors are not going to seek you out. Inevitably, you'll find that some professors can be unwilling to help and close to useless for advising purposes (my "official" adviser, for example) so you'll have to look around. Someone suggested going to the seminars. This is a very good suggestion (I think Mankiw suggested this on his blog). You'll be amazed how many professors get to know you simply because you're the only undergrad who goes. Also, take graduate courses-they're typically smaller and it gives you a chance to stand out.

 

One more interesting note about my experience: don't count on connections. you'll definitely want to seek out professors who are well-known and have lots of well-known friends. In my case, my best letter (I would hope!) came from a well-known economist who was classmates with and remains close friends and a co-author with several of the top macroeconomists in the world (graduates of the MIT classes from the early-mid 80's). Despite this, I was flat-out rejected from the schools where his contacts are, despite referencing these connections (with permission, of course). That's not to say that connections will hurt you, but I was surprised that they didn't help me at all.

[Dimitrvi]

Thank you! I hope you're right!

 

Having supervised undergrad RAs in the past, it's a lot harder than it looks I think.

 

And adding a tangent, how did you get the Profs to send the letters? did you ask for a paper letter o did they fill the online application? I had real trouble getting them to fill just 8 applications each, I wonder how did whiterabbit with over 20 apps??? Paper I guess??

[JRav]

And adding a tangent, how did you get the Profs to send the letters?

Often times, it depends on the school. Some schools (Harvard comes to mind) specifically say they will not accept paper letters. Others (MIT, Columbia, Berkeley) are a lot more lenient and will allow either. When you have a choice, always make it easier for your professors. If the professor prefers to send paper, make sure you print all the forms, fill out what you have to, and provide them with addressed envelopes and the like (for stamps, usually the department can mail the letters but be sure to check). If they prefer electronically, make sure the professors get the e-mail notifications and check with them that everything is okay.

 

For me, two of my professors had no preferences and uploaded them electronically. Another professor preferred paper so I made sure to give him everything (but in the end, he said he would get with the 21st century and sent everything electronically).

[thewhiterabbit]

And adding a tangent, how did you get the Profs to send the letters? did you ask for a paper letter o did they fill the online application? I had real trouble getting them to fill just 8 applications each, I wonder how did whiterabbit with over 20 apps??? Paper I guess??

 

All my professors have assistants and so it was easier. I am extremely lucky in this way.

[Popolo2]

I'm a little worried after reading this, because I don't have access to any professors who came from top 10 institutions who publish in anything outside of ARE. I'm hoping for at least one top 10 admit.

 

Moreover, I'm doing well (A in analysis and all that), but I'm hardly the only student doing well. A number of other students are doing as well as I am.

 

Is it hopeless?

[Zeno]

Based just on that, not necessarily hopeless. I went to a school with no faculty from a top 10, and in the last couple years two undergrads got between them (at least) three top 10 admits.