The Insidiousness of Racism: Education, Machine Learning, and Coffee
Yesterday, I was feeling ill but needed to go to work anyway. I thought that a nice cup of coffee from one of my favorite places would be a perfect way to start the day and make me feel better.
When I walked into the coffee shop, there were people at the counter but no one was in a line. I stood between the two registers, thinking that would be a likely place to start a line. In a few moments, three other people arrived, one standing to my left and two others behind me.
When the first register became available, the person at the register looked at me, the person to my left, and then called forward the person behind me. The next register became available and the person called forward the other person standing behind me.
I looked around wondering if I was standing in the wrong place and wondering how I would have known I was standing in the wrong place. Then another register became available and the person to my left was called forward.
None of the customers said anything to me or the people at the register to acknowledge that I had arrived before them. Neither did the people at the register inform me that I was breaking protocol in some way.
I looked around again and then left.
Here’s the thing: I don’t know that this is racism as I don’t know what they were thinking. I do know I was the only Black person among the customers and the people at the register. I know that I surveyed my appearance and my actions before I thought about those of the people in the coffee shop. I know that I thought about the growing inequities in San Francisco that are epitomized by these interactions between service workers and the insta-millionaires that hoard the Financial District where the coffee shop was located, as if to say, “I know you’re not mad at me. You’re mad at systemic racism and classism.”
And hence comes the insidiousness of racism: It might not have been racist, at least in intent. But I certainly thought it was, couldn’t help but think it was, just like other times when I couldn’t help but think the same. It’s these moments that our biases and prejudices are supported by a system that allows this to occur without punishment or consequence. They can continue to treat any potential customers this way because there aren’t enough people who care to cause any impact in their sales. It wasn’t just that the people at the register did this; all three of the customers had no problem walking around me to put their order in.
And now I’m never going to that coffee shop or any of its franchisees again meaning I’m out a bunch of reward points too.
Over the last month, we’ve been training the counselors at Bridge to College and case managers at Vielka Hoy Consulting in all things racist in education. One of the things I’ve been most proud of in my work is the amount of research that we use to build our strategies, and the trainings are not short on stats and data and research and theory. It’s also been fascinating to see some of my work coming to fruition — thinking about how entrepreneurship and machine learning can address these systemic inequalities in education.
In both organizations, we’ve only hired people who are attached to the mission in some way. They may have worked in education and higher education, and they have some understanding of the barriers that get in people’s way. Even with this group though, they’ve been stunned by a few data points, mostly to do with undermatching:
- We are undermatching students at higher rates, mostly underserved students. Undermatching is when a student is qualified to do one thing but does something that requires fewer qualifications. About 30% of qualified students attend colleges that are an undermatch, including community college.
- Almost all of these students are students of color.
- Of the undermatched students, about 2% will go on to earn a degree…ever. They take longer to earn the degree, spend more money to do so, and tend to just leave.
While free public universities and community colleges tend to sound great, we also know that this will lead to more undermatching.
In short, underserved students are going to community college for four or more years, transferring to a public university for another four or so years, and sometimes, maybe, perhaps, graduating. All this when they could have attended a college that was a match, been funded, and finished in four years.
Why are we doing that?
In my imagination, the scenario looks like the coffee shop where the colleges, counselors, teachers, and students are moving about a system that is designed to give everyone coffee but feels no consequences when it doesn’t. So they just don’t.
- Just like in my coffee shop scenario, we make up stories that look at the person in the system, rather than the system. These are excuses such as, “Not everyone is made for college,” or “if you want to go to college, just go.” I know those are not true but here is the thing we all know to be true: In our future economy, no one will be able to make a living without a college degree. So if you think you aren’t made for college or not inspired to go, it doesn’t matter. It’s irrelevant. You have to go and you have to earn a degree.
While I have my feelings on Cosby himself at the moment, the perfect example of this is the pilot episode when Theo learns to budget.
- There’s another reason and I call it Fast Food Reasoning. I reference the fast food industry because they have mastered rationalizing bad behavior in the name of profits. It’s not that they make food that is intentionally addicting, with the amount of sugar, fat, salt, and caffeine. It’s that you didn’t pick the apple slices on the menu. It’s not that they mostly have restaurants in food deserts. It’s that you didn’t care enough to find a grocery store a few cities over. You could have exercised more. You could have cooked at home more. You could have done something more. All true, but again, doesn’t change the fact that they made something with the intention of making money first and providing a nutritious meal, if it’s on the priority list, is certainly no where on the top.
The reason I think about this here is because, in addition to making up stories, we tend not to look at who is guiding our storytelling. We have community panels on nutrition where half the panel is fast food executives. They are conducting the food programs in elementary schools. They own the organic food companies too. It’s no wonder the messaging seems so consistent, but it doesn’t mean it’s correct.
In terms of college, there are A LOT of entities that are guiding our storytelling. We have college access companies that are funded by banks and credit cards with large college loan programs. We have college programs and schools for underserved that are run by White men who never earned a college degree and are independently wealthy. We have college programs that undermatch students run by company owners who promise to hire graduates of these programs but they never do. Why? Because they didn’t go to a college that was prestigious enough.
Yet, they continue to weigh in on these issues. Or worse, we ask them their opinions.
It blows my mind sometimes.
Here is what becomes especially mind-blowing: These are also the people who build the technology to serve students. So they are able to do these illogical things…at scale.
I attend a fair number of race and technology conferences, especially those in academia. People are interrogating the shade of yellow on the emojis and what shows up for whom in the google searches. All important, I suppose, but we still return to this idea that a number is just a number. It doesn’t have a bias. There isn’t an African American version of the number three, goes the logic. But that’s disingenuous, and the people who fall into that argument know that…hopefully.
The number does exist within an algorithm and then a system that is designed by a person who holds bias. When I build algorithms, even machine learning ones, I still make a decision about what data sets will be included, which ones I want you to pay the most attention to, which ones I will pay the most attention to.
For this, I’m reminded of a scene in the movie Focus.
For both of my companies, we don’t bother with for-profit colleges and technical schools, and we don’t look at colleges with graduation rates below 70%. The data we use are items we know are critical for underserved students. The surveys are done in such a way to not only get good data, but so that the user knows why it is good data.
Beyond technology, we are intentional about other things that the technology allows us to do. We talk as though selective colleges are just colleges. Whenever I meet with students, I’m wearing college shirts, I’m using college pens, I’m surrounded by college penants. I talk about racism easily and get really nerdy often. It goes beyond basic college knowledge because I know that we are working on them getting a college list that will have some big names on it, and I don’t want them to be intimidated.
I also want them to get to a place where they see what’s happening around them and they see it as ridiculous too. Ridiculous but malleable. They need to be empowered in that way because they we will all need them to be.
Because maybe they will one day be at the register, in line, or own the coffee shop company. And then I can get a cup of coffee.
Originally on the blog