Will your kid’s teacher soon be replaced by a robot?
As a researcher who specializes in education, Michael Hansen was watching closely as President Donald Trump took office.
Trump had made a campaign promise to steer billions of dollars into expanding the nation’s school choice options, and he had triggered controversy by nominating school voucher proponent Betsy Devos as his secretary of education.
But Hansen, a senior fellow at the Brookings Institution, said the noise being generated by the Trump administration on education in January 2017 quickly died down. And the biggest surprise a year later, he said, was how little the Department of Education had done to put Trump’s imprint on the nation’s education system.
“That’s not to say they’ve done nothing,” Hansen said during a recent visit to UNLV. “But they certainly haven’t moved in the directions that everyone was anticipating.”
During an interview with the Sun, Hansen recapped Trump’s moves on education and offered insight on other topics related to schools — a shortage of minority teachers, the best way to evaluate teachers’ performance, vouchers and more.
He also addressed a 21st century question: Are teachers on their way to being replaced by robots?
Excerpts of the interview follow:
In what ways have you seen Trump start to shape education?
There was certainly a lot of brouhaha upfront. There was a lot of concern about Betsy Devos making unprecedented harm and changes to public schools, such that the government would start divesting from them and putting them into school choices — private schools, etc.
So school choice felt like it was a leading issue and was going to be a prime area of focus. For example, on the campaign trail, Trump had a $20 billion school choice and charter school proposal. After he became President Trump, he made sort of a similar promise.
But that hasn’t come about. By and large, there has only been talk. There has not been any action.
Why do you think that’s the case?
One of the overarching philosophies governing the Department of Education is to reduce the role of the federal government in education policy. So it feels like that’s been more of an emphasis and, therefore, they’re actually doing very few things to exert influence into this school choice sphere. So there’s talk, but there’s no corresponding action.
Also, one of the most important actions Secretary Devos has undertaken has been more along the lines of deregulation.
The Every Student Succeeds Act was passed in December 2015, and it was very early in the rule-making process around that law when Trump came into office. There were two opportunities for the public to weigh in and for the Department of Education to write its own rules under the old administration. It did that, but then those rules were more or less reversed shortly after Devos came into power.
Also, on Title IX sexual assault guidance — all those “Dear colleagues” letters, transgender bathrooms issues — those have been reeled in.
Would school choice be good for the education system?
I’m a researcher, and I always try to be careful with my language, so I’m going to say it’s unclear. (Laughs).
But I will offer a few more nuanced points to think about.
We can think about school choice in several ways. We could have school choice in terms of charter schools in an environment where there’s already a thick distribution of traditional public schools.
In general, research has found that charter schools in that kind of setting — particularly in settings where traditional public schools have been underperforming — are making a significant impact.
This is a reform that has transformed urban communities.
What makes schools in those environments so effective?
More autonomy. They’re working without the same level of oversight as a public school. Then there’s the teacher workforce. In charter schools, those teachers are qualitatively different in several ways. They tend to be less unionized. They tend to skew a lot younger, and therefore they cost less. That cost efficiency isn’t necessarily a defining factor, but it doesn’t hurt.
Another factor is that charter schools have to give parents what they want. And often, particularly in these disadvantaged urban settings where there’s sort of a marketplace for schools, parents are selecting to get their kids in the highest-performing schools. So they tend to have more direct-market pressure than public schools, which are going to get their funding without that same level of risk.
OK, let’s get back to the broader point about how Trump’s proposal on school choice would affect education if it would come to fruition.
In expanding school choice in other contexts — in suburban and rural areas — the evidence there is not nearly as clear that this is necessarily good for students.
Overall, the expectation for how those schools perform compared with traditional public schools in those areas is more or less null. And in rural settings, they even tend to be slightly negative, so they’re underperforming compared with traditional public schools.
There are a few explanations for this that hold merit. One of my colleagues has done research on what parents want when they’re choosing schools, and what he highlights is that a parent seeking choice in an urban school setting chooses schools in a different way than a parent in a rural or suburban setting.
In the urban setting, it appears most of them are selecting for higher performance. But in a suburban setting, maybe a parent is driven by a focus on the whole child. He or she may be interested in the Montessori philosophy, or in more of a focus on the arts, those kinds of things. And understandably, those things come with a little more of a trade-off (in academic performance).
So even if they’re neutral on a performance standpoint, the parents don’t necessarily think that’s a bad thing.
We could also talk about school choice even within traditional public schools — some districts have STEM schools or magnet schools that kids can opt into. That’s a level of school choice that is generally popular with parents.
Then, if we look at private school choice, where students get access to public vouchers that pay tuition to private schools, there are 13 or so states that have voucher programs in place.
Up to about four or five years ago, the vouchers were targeted to specific segments of the student population — students who were in poverty and didn’t have good public schools around them, or those with special needs.
In general, the evidence on those programs has by and large been pretty positive. Students are benefiting, high school graduation rates are better, etc.
But what happens when you broaden that out and include more income levels?
Several states have extended their voucher programs to broader bases of students, and as they’ve done so, it increases the number of students demanding private schools. And this is opening it up to students who could have been just fine in the public schools they were attending.
So in expanding, among users of the voucher, you’re getting ones who are less and less high-need. And on the other hand, there’s a competitive effect — now you’re competing with other voucher holders, so perhaps there’s a crowding-out in private schools.
The evidence on these kinds of programs has actually been negative in quite a number of circumstances.
Some argue that we still have some accountability for private schools, so by requiring them to do standardized testing and so forth, it’s disruptive to the private school. And therefore private schools that participate in that program are not elite.
We don’t know which theory to believe, but the evidence on those expanded programs is that in a couple of states, there’s no difference, or those who get vouchers are actually worse off.
What’s the effect on the other end of that — the public schools that would lose funding that they might have had otherwise?
The short answer is it’s disruptive. And the way it’s disruptive is that in building a large school district, you have a lot of fixed costs that don’t simply go away when students go away. So you have buildings you’re maintaining. And particularly in underserved areas, if they have fewer students in those buildings, basically what that means is the cost of those facilities per student increases. That begins to crowd out instructional costs.
That’s something we need to think more clearly about and be very careful about in expanding school choice.
Your research emphasizes the challenges districts face in the diversity of the teacher workforce. Can you describe some of the key factors behind these challenges?
Nationwide, our student population is roughly 50 percent minority, whereas our teaching workforce is less than 20 percent minority.
The diversity of our teacher workforce has significantly increased over the past 10 to 20 years, but it’s not entirely optimistic moving forward. Among the youngest teachers, the millennials coming into the workforce, they tend to be more white.
So where are we losing minorities so they’re not showing up to be teachers? Think of this as a pipeline moving into the profession — it has many leaks. The most significant ones are basically getting minority youths to and through college. And that’s particularly the case among Latinos — (we’ve seen) particularly low college completion rates there.
So that would be priority No. 1. If we could get more minorities to and through college, that would plug a significant hole in the pipeline.
And once a student graduates with an education degree, the white graduates of those programs are more likely than minority students to end up teaching in the classroom. Part of that could be due to licensure policies. Some studies have shown significantly higher passage rates for whites than blacks on licensure tests, and there’s no strong evidence that this is due to the quality of candidates. So people have raised questions about whether these licensure tests are culturally biased.
You’ve written about how technology is poised to disrupt education.
Probably the positions in schools that are no longer going to be there, or maybe will still be there but in lower numbers, are administrative positions, accountants, receptionists.
As for teaching, a significant part of what a teacher does is providing child care, essentially. I’m not trying to reduce what they’re doing, but there’s a significant element of that. But it’s that element that is really hard to automate, and that will likely be the thing that saves teaching from being wiped out in a tidal wave of robots.
So I do speculate that there will be a lot more centralization and a lot of technological efficiencies on the lesson preparation side, developing and preparing curricula. Many places have become a lot more effective at it already.
But having an adult in the classroom supervising kids, that’s not something to easily replicate with a robot.
In Nevada, we’ve wrestled over how to best evaluate public school teachers. Have you come across any assessment tools that are particularly effective?
We have no obvious, really great way of assessing teacher performance.
First, it’s hard to define what good teaching is, and there’s a disagreement among experts as to which of these measures we should be using for assessments.
So we could measure teaching based on an outcome — student test scores. We can try to measure based on the practice we observe in the classroom. We can try to measure based on students’ experiences, how much they’ve learned based on self-reporting.
Each of these has its own set of problems.
So in value-added (assessment based on how a student performed on tests versus how the student had been expected to perform based on his or her previous results), how well does that capture an unbiased estimate of a teacher’s performance? We can construct statistical models where it looks like it’s really good, but can we say all teachers get an unbiased estimate? We can’t.
Likewise, even looking at an observation type of rating, one observer doing the rating might see the teacher’s performance as really strong, where another may disagree.
So it’s a mystery. What is a great evaluation system?
The common factor across all of these evaluation systems seems to be that they don’t place heavy reliance on any single factor, but rather are getting multiple signals from various types of performance. So they’ll have a value-added rating, they’ll also have observations, and they’ll have a teacher peer professional rating.