Allow Me to Ramble: A SacerBro Paper

I finished up college coaching, and thought I’d post on some awesome research by Professor Sacerdote.

First off, where to start! Professor Sacerdote (SacerBro)’s website has the craziest collection of cool papers on an assortment of topics. I’m going to talk about his paper on Blitz, in true Dartmouth fashion. He basically measured the correlations between Dartmouth students’ friendships and the volume of email they exchanged via Blitz.

How Do Friendships Form?

I love this paper because a) it’s a dataset of Dartmouth students, and it relies heavily on the email-crazy culture of this campus, b) email volume is a kind of data that is rare and extremely interesting to look into, and c) the conclusions that are drawn about the factors that influence friendship-making are quite interesting.

The authors obtained a dataset of the number of emails exchanged between any two people in the dataset, in addition to other demographic variables. (I’m imagining what a nightmare the data window must have looked like in STATA, to have at least as many columns as observations.) They found that geographic location on campus made students much more likely to become friends, as one would expect, even for inter-racial friendships. Several interesting facts were drawn from the analysis – having a person in your dorm of a different race than you reduced your email interaction by a factor of almost five. Being in the same dorm as any student implies you will exchange four times more emails with them. These sorts of conclusions are stunningly precise and incredibly interesting.

Email volume as a measure of friendship – a weird but justifiable way of turning a qualitative variable into a quantitative one. It’s interesting to think how you would measure friendship if you didn’t have this sort of proxy variable – you’d have to use some sort of survey, which is fraught with errors, because people may not be able to answer questions about friendship honestly or consistently, and you would have to run a logit regression on this yes-or-no answer. An interesting conundrum, with an innovative solution.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: