Stats PhD: I have 4 recommenders but need 3... Which to choose???

[open_ball]

I am applying to 16-18 stats/biostats PhD programs and I have 4 recommenders. Which do I choose? 

1. Tenured math professor I've known for 3 years. Easily one of my favorite professors. He taught me 30% (unfortunately diminishing now) of everything I know. He was also my REU mentor. 

2. Tenured professor who does biomedical engineering. I'm a research assistant in his lab but I feel like I'm not able to commit enough time to this project and I worry that I let him down. He's been understanding though — I remember he was impressed that I was a straight-A student 

3. Professor that I did another applied math REU with. Don't think he knows me that well or is that impressed by me lol. 

4. My real analysis professor. I've only taken 2 semesters of analysis with him, but I was the best (or at least among the top 3 students) in the course, and I came to office hours every week. He can attest to my mathematical ability and work ethic. 

[bayessays]

1 and 4 sound like easy choices.  There's not a ton to go off of for 2/3 - applied math is closer than biomedical engineering probably, but basically you just want somebody to write a nice letter that you're able to work independently and are smart, so whoever you think is more likely to do that.

[John Watson]

I agree that 1 and 4 are easy choices. Between 2 and 3, of course I recommend 2. Indeed, you're doing Statistics, and Biomedical Engineering is not that close to Statistics. However, one thing valuable about your experience with Professor Number Two is that you are a research assistant in his lab. That's research experience, and it will be valuable and important for your PhD applications, remember that doing a PhD is all about doing research. Your other three recommendation letters are all coursework letters, they can only say something like you did well in some specific courses, but only Letter Two is a research letter, which can testify your research abilities and potential.

 

Don't worry if you don't have any super big research achievements or publications, we all know that having publications in fields like Maths and Stats can be difficult. Also, don't worry if your research experience in Biomedical Engineering appears to be not closely connected with Stats, some research experience in another field can also be helpful. Many people study some non-Business subjects as undergrad, and they happen to do a PhD in Business, so previous research experiences in another field will also be helpful.

 

As for Letter Three, you already know that that professor doesn't know you well. A solid recommendation letter should certainly come from someone who is close to you and knows you well.

[John Watson]

Oh I was looking super quick. But, yeah, a recommendation letter should be positive and have something solid in it, I don't think that Professor Number Three can do as well as Number Two. A short paragraph of some broad, generic statements usually do not do good