Why Data Science Should be Applied to College Applications
Back in the 90s, the Wu-Tang Clan taught us through a handy 5-letter acronym that cash rules everything around us (that’s C.R.E.A.M. for all you late-stage millennials). Now that we’re deep in the 2010s and cash has been replaced with Data, it’s time for D.R.E.A.M.
Terrible word play aside, data is everything these days. From recommended Netflix shows and Spotify playlists to Tinder matches and curated Facebook newsfeed content, all of our data and meta-data are constantly being analyzed to guide our next decision.
True, a lot of these decisions aren’t of much import. Whether Tide or Gain detergent is recommended to you on Amazon likely won’t substantially affect your household finances or your social capital. But that’s not the point. The point is that that information is now available to you to capitalize on. Even if you have no digital footprint, I bet your purchasing decisions are still made based on recommendations from people who’ve made certain assumptions about you based on data they’ve gathered from interactions with you. We’re constantly sourcing recommendations, whether from automated analytics about our personal preferences, predicted behavior based on our demographic information, or from people who know our likes and dislikes.
The beauty of data science and data analytics is that it’s based on data compiled about us that we supply. It’s not just a survey about how we think we behave; it’s how we actually behave. We can then elicit trends across sectors and demographics. Recommendations tailored to us individually are thus better informed by large-scale, empirical, quantifiable data.
Okay, now, what if you could apply data analytics to the black box of college admissions?
We can. And we have.
College Admissions Reimagined
If you’re choosing your detergent based on automated recommendations, just wait ’til you see what you can do with actual admissions data, in the aggregate, sourced from people just like you.
At AdmitSee, we believe in the power of transparency and in universal access to higher education. It starts with information transparency.
Our database features thousands of real application files uploaded by verified college students themselves into LinkedIn-style profiles. These profiles include not just traditional quantitative info, such as grades and test scores, but also the good stuff — the previously unmeasured stuff — including legacy status, extracurriculars, essays (personal statements as well as supplements), and student advice about college interviews and campus life. All of these translate into millions of data points.
Okay, so what does this mean?
It means that applying to college can now be data-driven in the same way we make purchasing decisions.
Not only can we examine trends (e.g., the essay topics male vs. female applicants tend to gravitate toward or essay language and tone preferred by different top schools), we can now make very targeted recommendations to find you the best-fit school or to improve your chances at your dream school.
For example, our college students share in their profiles 5 adjectives that describe a typical student at their school. With thousands of aggregated responses, we can then get a overarching descriptive sense of a student body culture. Now enter 5 adjectives you as the applicant would use to describe yourself. And we’ve got a match!
Now, imagine copy/pasting a draft of your personal statement and having a database retrieve accepted essays ranked by percent similarity based not just on essay topic but also essay structure and sentiment.
We can do that.
Imagine being able to extract the differentiating factor that got an applicant admitted to a school despite lower than usual grades and test scores.
We can do that.
Imagine plugging in your stats and extracurricular activities as a junior and having a system recommend that you shoot for the varsity captain position next year because it’ll double your odds of getting accepted to your dream school.
We can do that.
What’s the Big (Data) Deal?
We started out with the mission to level the college admissions playing field. At an average price of $4,035 in the United States, private college consultants are prohibitively expensive for many students. The draw of hiring private help is to get personal attention and tap into that consultant’s wealth of knowledge and past case studies. The good ones are really good; they bring years of experience having sat at the admissions table and boast tons of connections.
However, if you don’t have the financial means, your options are limited. Your school counselor is overworked and under-resourced. (The average number of students to one school counselor in the state of California is 1,016. The national average is 478.) Those who have parents or siblings who’ve gone through the application process at least have first-hand knowledge, even if it’s outdated. But what about the first generation college applicants? What about international students? What about veterans?
We decided to create an affordable, accessible alternative that, like Yelp, is crowdsourced and centralized. We launched AdmitSee as the first searchable database of successful application files. Along the way, we discovered we were onto something bigger. We now had information that even those who hire top private consultants have never had access to.
The perspective of a near-peer turns out to be invaluable — it’s more applicable, more relevant, more current. Thanks to social media sharing and the now common practice of upvoting/downvoting, millennials turn to their peers to confirm decisions more than ever before. Sure, an expert consultant opinion is constructive, but who better to tell you whether a school has the vibe you’re looking for than someone on the ground, in real time?
Say you want to climb Half Dome in Yosemite National Park. Tough feat. You can seek out an expert climber to ask for his advice on how to train and what to prepare. He’ll undoubtedly provide valuable suggestions, but keep in mind that his caliber and industry experience can lead him to give one-size-fits-all answers or graze over key facts that he takes for granted but an amateur might not yet know. Alternatively, you can join a discussion board or Meetup for climbing aficionados. Share some information about yourself: your skill level, fitness stats, past favorite climbs. Seek out a multitude of opinions. Then pick and choose the advice most applicable to you offered by people most similar to you. Hell, do both. Ask Alex Honnold to share his pro tips and crowdsource from like-minded people.
But the takeaway is this: we want to be your Meetup group.
Now, Let’s Level Up.
Data science applied to college admissions isn’t about gaming the system. It’s about bringing an old-world institution up to a 21st century tech standard and helping all players make better decisions with data.
The information available isn’t meant to crack the code on admissions.
Hot tip: There’s no code to getting in. There’s no holy grail formula.
At the end of the day, the holistic admissions process comes down to human connection — did the application reader connect with you and your story? Does s/he feel you’ll be a value-add at the school? Have you effectively told your story? We can’t turn you into a candidate you’re not, but we can advise you on how to maximize your candidacy.
With data, we can get better at answering questions and figuring out what questions to ask in the first place. Taken beyond the applicant experience, this data can be also be used by universities to examine unconscious biases and revamp admissions practices. What kinds of application essays do schools tend to prefer? What kind of student character traits do they admit?
Now, take the 5 adjectives example I gave above. Wouldn’t it be good for a university administration to know how its students describe the school? Wouldn’t that be helpful in how the school markets and represents itself to potential applicants? Couldn’t this help with yield? Let’s even start to identify early warning signs for attrition and pair incoming freshmen with similar mentors — this can be based on personal background, academic interests, or career aspirations. The impact on retention, graduation rate, and even student career planning can be immense.
College admissions needs to be brought into the data world. Sooner or later, it’s going to make the full shift over — it’s already happening. We’re hoping to lead the charge because why wouldn’t we choose be more informed in every decision we make? And what choice is more important in the life of a young adult than where s/he will spend the next four years forging connections, discovering passions, and chasing dreams?
This is the future of college admissions. This is 2017. D.R.E.A.M.
Stay tuned for AdmitSee’s version of OkTrends. Here are some fun interactive charts to play with in the meantime: AdmitSee data insights.
For more of the data science behind AdmitSee’s analysis, read Mining the Common App: Part 1 (article discusses predictive models) and Mining the Common App: Part 2 (article delves into essay insights).
