A Meeting with Congress, Higher Ed, and a Bold New Way

I recently completed a trip to Stanford University where I participated in a roundtable discussion with the Subcommittee on Education and Workforce in Congress. Along with the congressional members were 7 other EdTech leaders in the room — many of which had raised over $100MM including Coursera, Udacity, and AltSchool. Undoubtedly I was representing the smallest and youngest start up in the room, yet I was peppered with questions.

So why the heck did they invite me? Who am I? And why should anyone care?

As a country, our higher education system has been the envy of the world for decades, but we are facing some very serious issues. Among them include student debt — a $1.4 Trillion dollar problem!

I was there to talk about an even bigger issue…

“The Biggest Crisis in Higher Ed Isn’t Student Debt, It’s Students Who Don’t Graduate “

— Dr. Michael Crow, President of Arizona State University

According to NCES data, at 4 year institutions today, 1 in 5 Freshman will drop out after the first year. At those same institutions, less than 60% will graduate from the school they started at within 6 years.

Not only does the student dropout problem lead directly to billions of dollars of wasted taxpayer money and drops in rankings and budgets of institutions, but it also means that 10s of millions of people are left with large amounts of debt without a clear path toward a higher paying job while leaving many of them with a lesser sense of self. This is a travesty.

The reason I was invited to speak with some of the highest elected officials in the country was because I founded a company that attacks the student completion debacle in a bold new way. We set the precedent that colleges have the ability to act — differentiating from the idea that colleges would like to act and are simply not able to or unsure how to.

During the discussion, as some political views were beginning to emerge, Virginia Foxx the Chairwoman of the Committee and representative for North Carolina’s 5th congressional district said, if I may paraphrase:

“We may disagree on some issues, but we all want the same thing — we want results!”

Results in Higher Ed are hard to come by. Particularly because the reasons why students leave are multidimensional, the one-size-fits-all approach greatly underserves the majority of students, and because institutions are slow to adapt and change. What has developed over the last few years are better 3rd party applications that allow Universities to better serve students. For example, course scheduling that optimizes a student’s path to graduation, classroom apps for discussions, etc…

The Analytics Problem in Higher Ed

These are all great tools, but in general, Universities and councilors are flying blind — they don’t actually know what leads students to drop out. What is missing is actual insight into students; that is, not simply what demographics are at-risk, but what actually causes the students to be at risk?

For instance: nearly every single 1st generation college student would be flagged as being at-risk by traditional approaches — yet many of these students will graduate. Likewise, nearly every traditional approach would mark a white middle-class male with a 3.3 GPA to be “safe” —yet a recent report indicates that 98% of institutions have more students drop out with a GPA above 2.0 than under.

The basis of all these approaches is to use basic linear or logistic regression models, more or less correlating profile and academic data, to identify who’s at risk and who’s not. At the end of the day, I don’t believe any of these “insights” actually help graduate more students. Sure, they help identify which populations of students tend to drop out, but are they really that successful in driving action?

Question from Congress: Are you able to determine which students should be in college, and which students shouldn’t be?

First, I do not believe that college is the right choice for everyone — that would be foolish. But, I do believe that if a student is willing to attend a university and pay $1000s of dollars to attend, they intend to graduate. It is the university’s duty to do everything within their power to enable that student’s success.

Many analytics providers or even some universities fall into a dangerous trap of using “profile variables” to determine success. They do this because, well, its what they have to work with. In general, this is what may have created the fallacy that the dropout rates of highly at-risk students may never be able to really be improved.

That being said, I refuse to believe that ethnicity, gender, socioeconomic status, or high school scores prohibit educational success. As a result, we do not use those variables to create at-risk scores. Anyone is capable under the right conditions to learn anything and succeed — the real question is, what exactly are those right conditions?

And that is the core of the problem: a chasm between the understanding of student behavior and actionable insights, and it is the reason I created Degree Analytics and the reason I was invited to speak with members of Congress.

A Smarter Path

We decided that there has to be a better way to evaluate and understand students. Traditional approaches are deeply flawed and universities need deeper and more real-time indicators they can act on.

That led us to a unique idea: using sensor data readily available around campus to gain an understanding of how students actually behave, often called “student engagement.” We then perform machine learning techniques that identify the different ways students succeed and provide real-time insights to schools that allow them to understand which students are at risk, why certain groups succeed over others, and the actual behaviors they can act on.

Basically it means that we are able to identify the 1st generation students that are both safe and at-risk. Similarly, we will identify that white middle-class male with a 3.3 GPA to be at risk, because as it turns out, he stopped going to class altogether over the last two weeks.

Question from Congress: How do you account for all of the different types of students? On vs Off Campus or 1st Generation for example?

It’s important to understand that we don’t presuppose anything about students — what behaviors lead to success or failure. We have a hunch that attending class is a good thing, but ultimately the machine learning figures out what’s important and then we validate those outcomes.

Personally, my favorite piece about this approach is that it is designed to understand and interpret the different paths of student success.

For example, on-campus and off-campus students should exhibit different behaviors. IE, a student living in an on-campus dorm should be on-campus more than a commuter student. Our solution actually figures out those different types behaviors or success paths of students.

Question from Congress: Is this scalable to other institutions? Do students everywhere have the same indicators?

Every single institution has a unique community and culture, and it is exactly why “out of the box” solutions often get vastly different results from school to school. Knowing this ahead of time, we built a system that generates insights according to each institution’s unique identity.

For every school, we identify what behaviors lead to student success. That means that each institution will have its own indicators that represent its student body.

Instead of a one-size-fits-all approach, it’s a methodology that identifies all’s sizes.

Closing Thought

One of the interesting ideas that many members of the University Innovation Alliance believe is that universities were designed for faculty and administrators, not for students and their success— changing that paradigm is going to take a lot of work. We believe that understanding how students engage, personalizing their experience, identifying reasons for failure and success, and creating informed personalized advising is critical towards achieving that goal.

Thanks again to the Subcommittee on Education and Workforce for allowing me to share this new approach towards student success. If you’d like to help us attack this massive problem facing our country and the world, please don’t hesitate to reach out.

Aaron Benz
CEO, Founder of Degree Analytics

Degree Analytics is a start up in located in Austin, TX currently in the Capital Factory Accelerator.

A Meeting with Congress, Higher Ed, and a Bold New Way was originally published in Austin Startups on Medium, where people are continuing the conversation by highlighting and responding to this story.

#URLinkedUp AustinStartup https://austinstartups.com/a-meeting-with-congress-higher-ed-and-a-bold-new-way-6455ec80a332?source=rss—-9504c035b990—4

#Austin Check out URLinkedUp > http://www.urlinkedup.com


Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s