PODCAST TRANSCRIPT

Announcer: 00:00 Steve Barkley Ponders Out Loud is sponsored by the AAIE Institute for International School Leadership. Preparing educators for the unique challenge of international school leadership through online courses led by international school leaders. Learn more at aaieinstitute.org.

Steve [Intro]: 00:18 Hello and welcome to the Steve Barkley Ponders Out Loud podcast. For over three decades, I’ve had the opportunity to learn with educators at all levels, both nationally and internationally. I invite you to listen as I explore my thoughts and learning on a variety of topics connected to teaching, learning and leading with some of the best and brightest educators from around the globe. Thanks for listening in.

Steve: 00:46 Repeat student-teacher matches. I had the opportunity to read an article that caught my attention by its title. And the title was “The Teacher Who Knows Me: The Academic Benefits of Repeat Student -Teacher matches.” The article was published by Andrew Hill and David Jones and we are fortunate enough today to have Andrew Hill, a professor from Montana State University, on the podcast with us. So welcome Andrew.

Andrew: 01:23 Thanks very much. Thanks for having me.

Steve: 01:25 So Andrew, before we jump into the study, when I had a chance to visit your website, I saw that a special interest of yours is education economics. And I’m going to guess to a lot of listeners, that’s not a phrase we as educators often come across. So you want to take just a moment to fill us in on the thinking behind that?

Andrew: 01:51 Sure. So basically, I’ll say economics is often more broadly defined than what people can expect it to be. It’s any kind of situation where you have constraints and incentives and people kind of operating in a world of a world of scarcity and responding to your incentive. So when we think about teachers and you know, what goes on in the classroom one way, one lens, which we can kind of look at that is kind of recognizing that they’ve got incentives, they want students to do well, they want to kind of impact students. If there’s, you know, different regimes where they kind of get paid based on performance or something like that, then that’s another example where they just kind of have all these incentives and I think economists are kind of just interested in any world in which you have these economic agents in a world of scarcity, kind of responding to incentives. Increasingly, what’s happened is economists have stepped into the education landscape. I guess, one reason is probably because now there’s – in the last few decades, there’s just a lot more data available, a lot more states collect administrative data where you see the students that are matched to teachers and it really gives a huge opportunity to really get inside what’s actually happening in terms of, you know, factors that affect student achievement and that kind of thing.

Steve: 03:08 Thank you. Thank you. So, kind of give us a thumb-sized sketch here of the breath of the study that you and your colleague did.

Andrew: 03:19 Right. So, my colleague and I have been working with data from the state of North Carolina for quite a few projects. So, what basically – they were kind of a very forward looking state in that they made their data available to researchers. Obviously it’s anonymized, so there’s no – you don’t know who the students and teachers are, but basically we have data from 1997 until 2013 of all the third, fourth and fifth graders in public schools in North Carolina. And we also able to see the teacher that administrated their kind of end of grade reading and math tests. So it’s a really, really rich data set. I think we’ve got about 2 million student observations in there.

Steve: 04:09 And so the bottom line statement that I pulled out was your finding that the students who matched up with the same teacher for a second year experienced an additional growth from other students. Have I worded that okay?

Andrew: 04:36 Right. I think that the tricky thing to always think about that we kind of spend a lot of time thinking about is recognizing that you can’t just compare students in the – you know, students who we see in these repeat matches to other students because there might be a whole lot of different reasons why they’re in a repeat match. And some of those could be correlated with their achievements. So essentially, the way I like to kind of describe our results is relative to the first year that a student is – that we see a student with a teacher in that second year that a student is with the same teacher, their school grows. And what we’re kind of doing is controlling for, say, the specific teacher they have. And, yeah. So I’ll say that’s kind of the phrase I would use. It’s in some sense, just the extra gain in student achievement the second year that a student is matched to a teacher in relation to the first year of that match.

Steve: 05:42 Is there a way that you can say that that growth the student makes the second year is greater than you would have predicted had the student just generically had a different teacher?

Andrew: 05:55 Yes. So are essentially – part of our comparison is against – is with the students who went on and went onto the next grade and had a different teacher. So, I guess it’s in some – it’s kind of a tricky comparison because part of the comparison is versus how they themselves did the first year of the match. And part of the comparison is kind of, the students who went on to have a different teacher. And it’s kind of like, there is a statistical way of disentangling that.

Steve: 06:29 So a generalization I pulled out that I want to check – and that is, increased student teacher familiarity improves academic achievement.

Andrew: 06:44 Right.

Steve: 06:44 And you’re just looking at students from grades three, four to five?

Andrew: 06:48 Yeah. So we, basically interpreting this effect that we’re seeing – essentially, we’re kind of, we’re just looking at raw data. We don’t have survey data on how this, you know, the student tells us how they actually experienced the class or survey data from the teacher where they kind of talk about how they experienced their interaction with the student. But the one thing we know for sure is that the student and the teacher know each other more than otherwise because they had this extra year together. And we just kind of, essentially describing that, as the student teacher familiarity. And we’re saying that kind of causes this increase in student achievement.

Steve: 07:31 So, when I first contacted you, I was using the word looping. So, that’s embedded – schools that practice looping, meaning the teacher moved on with the students to the next grade level, that would be included, but there may be other models where say students were placed with a teacher for a two year time frame. I started my teaching career teaching fifth and sixth grade. Half of my class were fifth graders, half were sixth. So I kept the same – at the end of the year, my sixth grade students moved on, my fifth grade students stayed with me for another year and half of my class came in different. So what would a setting like that have fallen into your study?

Andrew: 08:23 Right. So, I guess – two kinds of things I guess you’re talking about. The one is this policy of looping, which I guess might be, it might just be a single grade class with the whole fourth grade class goes to fifth grade with the same teacher – this multigrade teaching where you just, you know, the teachers always teaching fourth and fifth grade every year. So first they see the fourth grade students and I mean, your first year with the students when they’re in fourth grade and then you’ll also see them again when they’re in first grade. So yes, both those settings would be incorporated into our [unintelligible]. So what we actually find, you know, we have these 2 million students observations. It’s relatively rare in our data to actually see a student see the same teacher in fourth or fifth grade.

Andrew: 09:12 So we only see that happened about 60,000 times in our data set of a sample of 2 million and only about 5,000 of those are actually looping classes where the whole class has gone from fourth to fifth grade. So they’re kind of part of the story for sure and we think our analysis kind of speaks to the debate around looping, but we’re not ready. We haven’t conducted an explicit analysis on looping or an explicit analysis on multi grade classrooms. Because, I guess, there are all kinds of subsets of this kind of story, but they might have their own kind of specific things that happened inside those environments.

Steve: 09:53 Gotcha. So one of the pieces I found intriguing was what you called the the spillover effect of students who were new. They weren’t necessarily in the second year but had a positive effect on them.

Andrew: 10:15 Yeah. So, what we did there is we basically said, okay, we see this primary effect where students who are with a teacher for the second time do better. And then we could kind of imagine a different stories about how that might happen. It might just be that that teacher kind of specifically [unintelligible] and they’re able to kind of adapt their teaching, perhaps, in small ways to kind of help that particular student or they know that student’s home environment and there might be some mechanisms like that which kind of just impact that student. But then we also thought, well, another thing that could be happening is, you know, if you have a better relationship, the student – that maybe, that kind of improves their behavior in class and their focus and their discipline and that could actually then go on and affect the classmates who might be experiencing this teacher for the first time.

Andrew: 11:08 So what we did is we just looked at classes and we looked at the share of students in the class that we’re experiencing the teacher for the second time. And we noticed that as that share increased, you actually started seeing test score gains for the students who were experiencing that teacher for the first time. So not only did the students in the repeat match do better, but once you started getting classes where maybe about 40 or 50% of the students were in a repeat match with the teacher, the other 60% of students in the class also started experiencing test score gains relative to what they would have done if they were in a class where none of the students were in a repeat match with the teacher.

Steve: 11:54 Yeah. Let me click through a couple of these others because they kind of all added up to an interesting conclusion for me. Then you found out about that it had a greater impact on minority students. Speak to that a little bit.

Andrew: 12:12 Okay. So again here, just to emphasize, we are kind of just, you know, we just have this administrative data where we can identify the race of the student and the race of the teacher as well, which isn’t such a big part of this particular paper. But we kind of, you know, I mean there’s thousands of papers trying to understand these achievement gaps in education between different race groups in the states. So we kind of thought, well, how could these repeat student matches kind of speak to those debates. So then we kind of just essentially took our data and we saw that for black students, the test score gain that they experienced from a repeat match was actually greater than it was for white students.

Andrew: 12:59 And then after that, we can’t really definitively pin down the mechanism, but our story was kind of going along the lines that one thing that might explain this, which, you know, you would probably need survey data to confirm was that, you know, most teachers in North Carolina are white. So they might have like, a greater social distance from, you know, to their black students. And maybe that gets narrowed more when you get these, repeat matches with black students. But that, just to emphasize, that’s kind of our interpretation of of the results. There might be some other story that explains that finding. But we’re just kind of saying that that kind of could be consistent with what we see.

Steve: 13:50 It could also – it would strike me, could also be the student to student understanding and relationship with more time. So the fact that the students got to know each other better, could change that whole learning environment for the minority students in the class.

Andrew: 14:09 Okay. That’s interesting. I’m trying to think – that that definitely could be a story. The way we’ve done our data – the way we’ve analyzed our data up to this point, I probably wouldn’t be able to empirically speak to that point. But that’s certainly something that I guess we could look at. Try and disentangle, I guess, you know, actually look at maybe the share of my minority students in the class as an additional fact and see if that kind of affects the pattern that we’re observing. But yeah, that definitely could be a plausible mechanism for the effect.

Steve: 14:47 The other point that I found interesting was the effect size was greater when the teacher was less effective. So if I read that right, the less effective teacher got an even bigger growth – what, would that be percentage wise, the second year?

Andrew: 15:08 Yes. So basically – and this is just, basically what we did is, we divided up their teachers by their value added scores. And obviously completely recognizing that that’s one small measure of teacher effectiveness. There’s a whole lot about the stuff that goes on inside the classroom that we don’t see in our data that’s, you know, obviously key to the education process. But when we divided up teachers into four quartiles based on their value added scores, we saw that the overall effect size was actually larger for the teachers who had the lowest value added scores. It’s difficult. I mean, so that’s a finding, which I guess, again, it’s kind of one of the signals – we have this finding, whether that, you know, the exact interpretation of that’s kind of a, you know, a whole lot of different story. So we kind of just are reporting that finding. That’s what we find and just even the size, it’s not – a prime result from a statistical perspective is very, very strong.

Andrew: 16:13 Statistically, I would say it’s more suggestive than clear. But it’s what we see in the data and we’re interested in that because we kind of think, well, if this is true, we’re not really sure how it happens, but it might be, you know, improving the student teacher relationship. That actually might be a mechanism which could help teachers, you kind of, their value added scores are like on the low end, this could actually be a mechanism to make them more effective in terms of test scores.

Steve: 16:48 I had to do some thinking about that piece as I read it. I mulled over it a few times. My wife’s an educator, so she and I were pondering it together. But then it clicked to me that your higher end effectiveness of teacher would be a person who would build those relationships and build the learning community faster. So the idea that somebody less effective took longer to do that in order to get the payoff for that learning environment, kinda clicked to make sense then.

Andrew: 17:30 Right. Yeah. I mean, that’s definitely a plausible story for what we see. Coming from like a data angle I guess, you know, unfortunately, I can’t get, you know – it would be great if I had survey data every month, you know, how has the relationship with their students because that’s exactly the kind of thing could empirically investigate. But it sounds to me like that’s definitely a thing that could be happening.

Steve: 17:56 And then lastly, the point that hit my interests were the comments that you made regarding schools that are looking at departmentalizing their teachers in fourth and fifth grade so students are working with more teachers thus decreasing that time for the relationship building.

Andrew: 18:23 Right. Yeah, so basically, there was some interesting work coming out of the Houston Public School districts done by Roland Fryer who kind of conducted this experiment in the school district, where kind of, schools were randomly assigned to either do this teacher specialization in these early grades or just maintain having the same teacher for fourth and fifth grade. And what he actually found is that these schools where they introduced a system of specialization, the early grades, the students actually did a little worse. So though you might think you could choose who are perhaps [unintelligible] math and let them specialize in math in fourth grade and the teacher who teaches them is the best at doing reading, let them specialize in reading. They actually found that any gains that might have come from that specialization, it seems that they were counted by the fact that, you know, teachers perhaps didn’t build the same relationships with their students. So we just kind of say, well, we saw those findings, what we find in a very different context seems to be consistent with what was what they found in this.

Steve: 19:35 Well Andrew, it was a pleasure for me to read through this and to end the talk with you because it is strongly reinforcing for me because I place a great value on teachers building that student relationship — teacher to student, as well as building that community of learners. So I frequently have looked at it across various grade levels in lots of different designs and your study gave me one more piece of reinforcement. Is there any point that you think I may have missed here in asking you about?

Andrew: 20:22 I guess one thing just to say is that, you know, we definitely have a clear finding here. I mean, with a lot of studies in education, it’s not a massive effect. It’s not like going to take students from C’s to A’s type of thing, but I, you know, I was kind of struck by the fact that we only – I saw such a small share of looping classes in our data and I kind of think just, you know, as one policy that could be relatively low cost, this could be, I mean, it would be exciting for me to see this kind of thing explored more in a lot more schools.

Steve: 21:03 It’s actually pretty much no cost.

Andrew: 21:05 Right, right. I mean, I guess the one thing I guess teachers — I mean, I’m sure that first year when you are third grade teacher and you’ve got to prepare a whole fourth grade syllabus for the first time, that that might be be quite a big thing. But once you’ve done that and that fixed cost as being kind of spent, then, you know, I agree and it seems like would it be a relatively low cost kind of thing.

Steve: 21:29 And let me add a piece to that. I’m big on trying to get teachers to work vertically because to do an efficient job of teaching the fourth grade curriculum, it’s actually critical that you really understand the third grade curriculum. So there’s another whole bonus besides knowing the kids in knowing the content and what it is the student’s experienced. And if that teacher at the end of that second year loops back to pick up another three, four cycle, then I find them to be even a stronger teacher in third grade the second time they’re teaching third grade because they now have an understanding of the fourth grade material that the student needs to go on to.

Andrew: 22:18 Right. Yeah, that’s what — I think it would definitely be exciting to see more research on that for sure. I mean, I know there’s a lot out there, but yeah, I definitely think it would be an interesting and potentially a rewarding kind of change to see more of that.

Steve: 22:39 Well Andrew, thank you. And we will put a link to your article into the lead in to the podcast so listeners can go back and check it out.

Andrew: 22:53 Great.

Steve: 22:54 Thanks again. Have a great day.

Andrew: 22:56 Thank you very much, Steve. Thank you.

Steve: 22:57 Bye, bye.

Andrew: 22:58 Bye.

Steve [Outro]: 22:59 Thanks again for listening. You can subscribe to Steve Barkley Ponders Out Loud on iTunes and Podbean and please remember to rate and review us on iTunes. I also want to hear what you’re pondering. You can find me on twitter @stevebarkley or send me your questions and find my videos and blogs atbarkleypd.com.