Could Statistics Help You Find the Right College?

By Catherine Groux
Posted November 27, 2012 03:00 PM

Could statistics help students find the right college for them?
Could statistics help students find the right college for them?
Today, people can use facts, figures and statistics to calculate almost anything, from what movies you are likely to enjoy on Netflix to what ads you will probably click on while browsing Facebook. Therefore, it only makes sense that researchers could use similar equations to determine the right colleges for you based on your likes, dislikes, geographic preferences and interest in other schools.

One website that is trying to put this theory into effect is, which uses more than 765,600 previous college applications to help students see what their odds are of getting into certain schools and recommend other institutions that might suit them.

Matthew Pittinsky,'s chief executive officer and co-founder of education software developer Blackboard, recently spoke with U.S. News University Directory about and what he thinks a statistics-based system could do for students.

Q: What exactly is statistics- or social-based college selection?

A: The best analogy is from movie rentals, and Netflix is always presented as an example. We look for movies and instead of looking for the perfect movie by searching for "science fiction," a certain length or other attributes of the movie, we recognize that people like us tend to like similar movies. If we can share the data about the movies in general, what we will discover is there is a whole set of movies that are potentially good matches for us.

We are seeing the same kind of social-driven tools enabling students to find a collection of colleges that may be a good match. If I am a student who's interested in Caltech, MIT and Carnegie Mellon, but I am also interested in Reed College and Williams College, then I should know about Harvey Mudd because it is small. That small-school experience I have expressed some interest in is captured well, but also it is an engineering-focused institution, and clearly from the other universities I've listed, I have an interest in engineering.

Students today have a great opportunity to say, 'Here is who I am like. Where are they applying or expressing interest? Based on that, what are a collection of colleges that represent a good match for me?'

Q: How does use other students' data to help individuals find the right college?

A: is the student-ordering site for transcripts from 30% of high schools in the country. We have a very large database of students, where they have applied, where they have been accepted, where they have been rejected and where they ultimately chose to go in order to power those kinds of tools. We really think that is the next generation of finding that best match and recognizing that it is not a needle in a haystack to look for a school.

Students like small schools and big schools. They often want to be in an urban place and in a rural place. Students will express interest in the University of Michigan and NYU, and those are two very, very different places. So how do you recognize that that is a signature? The fact that you have that particular vector of colleges in mind is a signature for you, and it is a signature that can allow us to match you to other students with a similar signature and then from that, recognize a small cluster of colleges that are probably the best match for you. 

Q: Could using this data- and social-driven college selection method help the nation improve its graduation rate?

A: Absolutely. It starts with the right match and it starts with helping college students be much more informed about the college admissions process. It starts with trying to move away from this notion that there is a match, or that there is just one school for you, or even two or three schools for you. I had an amazing experience at American University. I think it changed my life. I met my wife there, I found a business partner there. But there were several institutions that would have been a great match for me. The question is how do we locate that cluster of colleges that, given your signature - your particular set of interests in terms of academic programs, geography, school size, class size, sports programs and all of those other dimensions - so that you start off with a higher probability of success? That enrollment match is absolutely critical. 

We recommend