How Tutoring Works (Cooking With The Grandkids)

My wife, Nancy, and I have three grandkids: Adaya (4 ½), Leo (3 ½), and Ava (8 months). They all live in Baltimore, so we see quite a lot of them, which is wonderful.

As with most grandparents and grandkids, one of our favorite activities with Adaya and Leo is cooking. We have two folding stepladders in the kitchen, which the kids work from. They help make pancakes, scrambled eggs, spaghetti, and other family classics. We start off giving the kids easy and safe tasks, like measuring and pouring ingredients into bowls and mixing, and as they become proficient, we let them pour ingredients into hot pans, scramble eggs on the stove, and so on. They love every bit of this, and are so proud of their accomplishments.

So here is my question. What are we making when we cook with the grandkids? If you say pancakes and eggs, that’s not wrong, but perhaps these are the least important things we are doing.

What we are really doing is building the thrill of mastery in a loving and supportive context. All children are born into a confusing world. They want to understand their world and to learn to operate effectively in it. They want to do what the big people do. They also want to be loved and valued.

Now consider children who need tutoring because they are behind in reading. These kids are in very big trouble, and they know it. All of them understand what the purpose of school is. It is to learn to read. Yet they know they are not succeeding.

The solution, I believe, is a lot like cooking with people who love you. In other words, it’s tutoring, in small groups or one-to-one.

The effectiveness of tutoring is very well established in rigorous research, as I’ve noted more than once in this series of blogs. No surprise there. But what is surprising is that well-trained, caring tutors without teaching certificates using well-structured materials get outcomes just as good as those obtained by certified teachers. How can this be? If tutoring works primarily because it enables teachers to adapt instruction to meet the learning needs of individual students, then you’d expect that students who receive tutoring from certified, experienced teachers would get much better outcomes than those tutored by teaching assistants. But they don’t, on average. Further, a U.K. study of one-to-one tutoring over the internet found an effect size of zero. These and other unexpected findings support a conclusion that while the ability to individualize instruction is important in tutoring, it is not enough. The additional factor that explains much of the powerful impacts of tutoring, I believe, is love. Most tutors, with or without teaching certificates, love the children they tutor in a way that a teacher with 25 or 30 students usually cannot. A tutor with one or just a few children at a time is certain to get to know those children, and to care about them deeply. From the perspective of struggling children, their tutor is not just a teacher. She or he is a lifeline, a new chance to achieve the mastery they crave. Someone who knows and cares about then and will stick with them until they can read.

This is why individual or small-group tutoring is a bit like cooking with your grandparents. In both settings, children receive the two things they need and value the most: love and mastery.

My point here is not sentimental or idealistic. It is deadly practical. We already know a lot about how to use tutoring effectively and cost-effectively. Yet there is a great deal more we need to learn to maximize the benefits and minimize the costs of effective tutoring. We need to find out how to extend positive effects to larger numbers of students, to learn how to maintain and build on initial successes in the early grades, how to successfully tutor upper-elementary and secondary students, and how to reach students who still do not succeed despite small-group tutoring. We need to experiment with adaptations of tutoring for English learners.

We know that tutoring is powerful, but we need to make it more cost-effective without reducing its impact, so that many more children can experience the thrill of mastery. To do that, we have a lot of work to do. Let’s get cooking!

This blog was developed with support from the Laura and John Arnold Foundation. The views expressed here do not necessarily reflect those of the Foundation.

Evidence Affects School Change and Teacher-by-Teacher Change Differently

Nell Duke, now a distinguished professor at the University of Michigan, likes to tell a story about using cooperative learning as a young teacher. She had read a lot about cooperative learning and was excited to try it in her elementary class. However, not long after she started, her principal came to her class and asked her to step into the hall. “Miss Duke,” he said, “what in blazes are you doing in there?”

Nell told her principal all about cooperative learning, and how strongly the research supported it, and how her students were so excited to work in groups and help each other learn.

“Cooperative learning?” said her principal. “Well, I suppose that’s all right. But from now on could you do it quietly?”

Nell Duke’s story exemplifies one of the most important problems in research-based reform in education. Should research-based reform focus on teachers or on schools? Nell was following the evidence, and her students were enjoying the new method and seemed to be learning better because of it. Yet in her school, she was the only teacher using cooperative learning. As a result, she did not have the support or understanding of her principal, or even of her fellow teachers. Her principal had rules about keeping noise levels down, and he was not about to make an exception for one teacher.

However, the problem of evidence-based reform for teachers as opposed to schools goes far beyond the problems of one noisy classroom. The problem is that it is difficult to do reform one teacher at a time. In fact, it is very difficult to even do high-quality program evaluations at the teacher level, and as a result, most programs listed as effective in the What Works Clearinghouse or Evidence for ESSA are designed for use at least in whole grade levels, and often in whole schools. One reason for this is that it is more cost-effective to provide coaching to whole schools or grade levels. Most successful programs provide initial professional development to many teachers and then follow up with coaching visits to teachers using new methods, to give them feedback and encouragement. It is too expensive for most schools to provide extensive coaching to just one or a small number of teachers. Further, multiple teachers working together can support each other, ask each other questions, and visit each other’s classes. Principals and other administrative staff can support the whole school in using proven programs, but a principal responsible for many teachers is not likely to spend a lot of time learning about a method used by just one or two teachers.

When we were disseminating cooperative learning programs in the 1980s, we started off providing large workshops for anyone who wanted to attend. These were very popular and teachers loved them, but when we checked in a year later, many teachers were not using the methods they’d learned. Why? The answer was most often that teachers had difficulty sustaining a new program without much support from their leadership or colleagues. We’d found that on-site coaching was essential for quality implementation, but we could not provide coaching to widely dispersed schools. Instead, we began to focus on school-wide implementations of cooperative learning. This soon led to our development and successful evaluations of Success for All, as we learned that working with whole schools made it possible not only to ensure high-quality implementations of cooperative learning, but also to add in grouping strategies, tutoring for struggling readers, parent involvement approaches, and other elements that would have been impossible to do in a teacher-by teacher approach to change.

In comparison with our experience with cooperative learning focused on individual teachers, Success for All has both been more effective and longer-lasting. The median Success for All school has used the program for 11 years, for example.

Of course, it is still important to have research-based strategies that teachers can use on their own. Cooperative learning itself can be used this way, as can proven strategies for classroom management, instruction, assessment, feedback, and much more. Yet it is often the case that practices suggested to individual teachers were in fact evaluated in whole school or grade levels. It is probably better for teachers to use programs proven effective in school-level research than to use unevaluated approaches, but teachers using such programs on their own should be aware that teachers in school-level evaluations probably received a lot of professional development and in-class coaching. To get the same results, individual teachers might visit others using the programs successfully, or at a minimum participate in social media conversations with other teachers using the same approaches.

Individual teachers interested in using proven programs and practices might do best to make common cause with colleagues and approach the principal about trying the new method in their grade level or in the school as a whole. This way, it is possible to obtain the benefits of school-wide implementation while playing an active role in the process of innovation.

There are never guarantees in any form of innovation, but teachers who are eager to improve their teaching and their students’ learning can work with receptive principals to systematically try out and informally evaluate promising approaches. Perhaps nothing would have changed the mind of Nell Duke’s principal, but most principals value initiative on the part of their teachers to try out likely solutions to improve students’ learning.

The numbers of children who need proven programs to reach their full potential is vast. Whenever possible, shouldn’t we try to reach larger numbers of students with well-conceived and well-supported implementations of proven teaching methods?

 This blog was developed with support from the Laura and John Arnold Foundation. The views expressed here do not necessarily reflect those of the Foundation.

Queasy about Quasi-Experiments? How Rigorous Quasi-Experiments Can Minimize Bias

I once had a statistics professor who loved to start discussions of experimental design with the following:

“First, pick your favorite random number.”

Obviously, if you pick a favorite random number, it isn’t random. I was recalling this bit of absurdity recently when discussing with colleagues the relative value of randomized experiments (RCTs) and matched studies, or quasi-experimental designs (QED). In randomized experiments, students, teachers, classes, or schools are assigned at random to experimental or control conditions. In quasi-experiments, a group of students, teachers, classes, or schools is identified as the experimental group, and then other schools are located (usually in the same districts) and then matched on key variables, such as prior test scores, percent free lunch, ethnicity, and perhaps other factors. The ESSA evidence standards, the What Works Clearinghouse, Evidence for ESSA, and most methodologists favor randomized experiments over QEDs, but there are situations in which RCTs are not feasible. In a recent “Straight Talk on Evidence,” Jon Baron discussed how QEDs can approach the usefulness of RCTs. In this blog, I build on Baron’s article and go further into strategies for getting the best, most unbiased results possible from QEDs.

Randomized and quasi-experimental studies are very similar in most ways. Both almost always compare experimental and control schools that were very similar on key performance and demographic factors. Both use the same statistics, and require the same number of students or clusters for adequate power. Both apply the same logic, that the control group mean represents a good approximation of what the experimental group would have achieved, on average, if the experiment had never taken place.

However, there is one big difference between randomized and quasi-experiments. In a well-designed randomized experiment, the experimental and control groups can be assumed to be equal not only on observed variables, such as pretests and socio-economic status, but also on unobserved variables. The unobserved variables we worry most about have to do with selection bias. How did it happen (in a quasi-experiment) that the experimental group chose to use the experimental treatment, or was assigned to the experimental treatment? If a set of schools decided to use the experimental treatment on their own, then these schools might be composed of teachers or principals who are more oriented toward innovation, for example. Or if the experimental treatment is difficult, the teachers who would choose it might be more hard-working. If it is expensive, then perhaps the experimental schools have more money. Any of these factors could bias the study toward finding positive effects, because schools that have teachers who are motivated or hard-working, in schools with more resources, might perform better than control schools with or without the experimental treatment.

Because of this problem of selection bias, studies that use quasi-experimental designs generally have larger effect sizes than do randomized experiments. Cheung & Slavin (2016) studied the effects of methodological features of studies on effect sizes. They obtained effect sizes from 645 studies of elementary and secondary reading, mathematics, and science, as well as early childhood programs. These studies had already passed a screening in which they would have been excluded if they had serious design flaws. The results were as follows:

	No. of studies	Mean effect size
Quasi-experiments	449	+0.23
Randomized experiments	196	+0.16

Clearly, mean effect sizes were larger in the quasi-experiments, suggesting the possibility that there was bias. Compared to factors such as sample size and use of developer- or researcher-made measures, the amount of effect size inflation in quasi-experiments was modest, and some meta-analyses comparing randomized and quasi-experimental studies have found no difference at all.

Relative Advantages of Randomized and Quasi-Experiments

Because of the problems of selection bias, randomized experiments are preferred to quasi-experiments, all other factors being equal. However, there are times when quasi-experiments may be necessary for practical reasons. For example, it can be easier to recruit and serve schools in a quasi-experiment, and it can be less expensive. A randomized experiment requires that schools be recruited with the promise that they will receive an exciting program. Yet half of them will instead be in a control group, and to keep them willing to sign up, they may be given a lot of money, or an opportunity to receive the program later on. In a quasi-experiment, the experimental schools all get the treatment they want, and control schools just have to agree to be tested.  A quasi-experiment allows schools in a given district to work together, instead of insisting that experimental and control schools both exist in each district. This better simulates the reality schools are likely to face when a program goes into dissemination. If the problems of selection bias can be minimized, quasi-experiments have many attractions.

An ideal design for quasi-experiments would obtain the same unbiased outcomes as a randomized evaluation of the same treatment might do. The purpose of this blog is to discuss ways to minimize bias in quasi-experiments.

In practice, there are several distinct forms of quasi-experiments. Some have considerable likelihood of bias. However, others have much less potential for bias. In general, quasi-experiments to avoid are forms of post-hoc, or after-the-fact designs, in which determination of experimental and control groups takes place after the experiment. Quasi-experiments with much less likelihood of bias are pre-specified designs, in which experimental and control schools, classrooms, or students are identified and registered in advance. In the following sections, I will discuss these very different types of quasi-experiments.

Post-Hoc Designs

Post-hoc designs generally identify schools, teachers, classes, or students who participated in a given treatment, and then find matches for each in routinely collected data, such as district or school standardized test scores, attendance, or retention rates. The routinely collected data (such as state test scores or attendance) are collected as pre-and posttests from school records, so it may be that neither experimental nor control schools’ staffs are even aware that the experiment happened.

Post-hoc designs sound valid; the experimental and control groups were well matched at pretest, so if the experimental group gained more than the control group, that indicates an effective treatment, right?

Not so fast. There is much potential for bias in this design. First, the experimental schools are almost invariably those that actually implemented the treatment. Any schools that dropped out or (even worse) any that were deemed not to have implemented the treatment enough have disappeared from the study. This means that the surviving schools were different in some important way from those that dropped out. For example, imagine that in a study of computer-assisted instruction, schools were dropped if fewer than 50% of students used the software as much as the developers thought they should. The schools that dropped out must have had characteristics that made them unable to implement the program sufficiently. For example, they might have been deficient in teachers’ motivation, organization, skill with technology, or leadership, all factors that might also impact achievement with or without the computers. The experimental group is only keeping the “best” schools, but the control schools will represent the full range, from best to worst. That’s bias. Similarly, if individual students are included in the experimental group only if they actually used the experimental treatment a certain amount, that introduces bias, because the students who did not use the treatment may be less motivated, have lower attendance, or have other deficits.

As another example, developers or researchers may select experimental schools that they know did exceptionally well with the treatment. Then they may find control schools that match on pretest. The problem is that there could be unmeasured characteristics of the experimental schools that could cause these schools to get good results even without the treatment. This introduces serious bias. This is a particular problem if researchers pick experimental or control schools from a large database. The schools will be matched at pretest, but since the researchers may have many potential control schools to choose among, they may use selection rules that, while they maintain initial equality, introduce bias. The readers of the study might never be able to find out if this happened.

Pre-Specified Designs

The best way to minimize bias in quasi-experiments is to identify experimental and control schools in advance (as contrasted with post hoc), before the treatment is applied. After experimental and control schools, classes, or students are identified and matched on pretest scores and other factors, the names of schools, teachers, and possibly students on each list should be registered on the Registry of Efficacy and Effectiveness Studies. This way, all schools (and all students) involved in the study are counted in intent-to-treat (ITT) analyses, just as is expected in randomized studies. The total effect of the treatment is based on this list, even if some schools or students dropped out along the way. An ITT analysis reflects the reality of program effects, because it is rare that all schools or students actually use educational treatments. Such studies also usually report effects of treatment on the treated (TOT), focusing on schools and students who did implement for treatment, but such analyses are of only minor interest, as they are known to reflect bias in favor of the treatment group.

Because most government funders in effect require use of random assignment, the number of quasi-experiments is rapidly diminishing. All things being equal, randomized studies should be preferred. However, quasi-experiments may better fit the practical realities of a given treatment or population, and as such, I hope there can be a place for rigorous quasi-experiments. We need not be so queasy about quasi-experiments if they are designed to minimize bias.

References

Baron, J. (2019, December 12). Why most non-RCT program evaluation findings are unreliable (and a way to improve them). Washington, DC: Arnold Ventures.

Cheung, A., & Slavin, R. (2016). How methodological features affect effect sizes in education. Educational Researcher, 45 (5), 283-292.

 This blog was developed with support from the Laura and John Arnold Foundation. The views expressed here do not necessarily reflect those of the Foundation.

Why Can’t Education Progress Like Medicine Does?

I recently saw an end-of-year article in The Washington Post called “19 Good Things That Happened in 2019.” Four of them were medical or public health breakthroughs. Scientists announced a new therapy for cystic fibrosis likely to benefit 90% of people with this terrible disease, incurable for most patients before now. The World Health Organization announced a new vaccine to prevent Ebola. The Bill and Melinda Gates Foundation announced that deaths of children before their fifth birthday have now dropped from 82 per thousand births in 1990 to 37 in 2019. The Centers for Disease Control reported a decline of 5.1 percent in deaths from drug overdoses in just one year, from 2017 to 2018.

Needless to say, breakthroughs in education did not make the list. In fact, I’ll bet there has never been an education breakthrough mentioned on such lists.

I get a lot of criticism from all sides for comparing education to medicine and public health. Most commonly, I’m told that it’s ever so much easier to give someone a pill than to change complex systems of education. That’s true enough, but not one of the 2019 medical or public health breakthroughs was anything like “taking a pill.” The cystic fibrosis cure involves a series of three treatments personalized to the genetic background of patients. It took decades to find and test this treatment. A vaccine for Ebola may be simple in concept, but it also took decades to develop. Also, Ebola occurs in very poor countries, where ensuring universal coverage with a vaccine is very complex. Reducing deaths of infants and toddlers took massive coordinated efforts of national governments, international organizations, and ongoing research and development. There is still much to do, of course, but the progress made so far is astonishing. Similarly, the drop in deaths due to overdoses required, and still requires, huge investments, cooperation between government agencies of all sorts, and constant research, development, and dissemination. In fact, I would argue that reducing infant deaths and overdose deaths strongly resemble what education would have to do to, for example, eliminate reading failure or enable all students to succeed at middle school mathematics. No one distinct intervention, no one miracle pill has by itself improved infant mortality or overdose mortality, and solutions for reading and math failure will similarly involve many elements and coordinated efforts among many government agencies, private foundations, and educators, as well as researchers and developers.

The difference between evidence-based reform in medicine/public health and education is, I believe, a difference in societal commitment to solving the problems. The general public, especially political leaders, tend to be rather complacent about educational failures. One of our past presidents said he wanted to help, but said, “We have more will than wallet” to solve educational problems. Another focused his education plans on recruiting volunteers to help with reading. These policies hardly communicate seriousness. In contrast, if medicine or public health can significantly reduce death or disease, it’s hard to be complacent.

Perhaps part of the motivational difference is due to the situations of powerful people. Anyone can get a disease, so powerful individuals are likely to have children or other relatives or friends who suffer from a given disease. In contrast, they may assume that children failing in school have inadequate parents or parents who need improved job opportunities or economic security or decent housing, which will take decades, and massive investments to solve. As a result, governments allocate little money for research, development, or dissemination of proven programs.

There is no doubt in my mind that we could, for example, eliminate early reading failure, using the same techniques used to eliminate diseases: research, development, practical experiments, and planful, rapid scale-up. It’s all a question of resources, political leadership, collaboration among many critical agencies and individuals, and a total commitment to getting the job done. The year reading failure drops to near zero nationwide, perhaps education will make the Washington Post list of “50 Good Things That Happened in 2050.”

This blog was developed with support from the Laura and John Arnold Foundation. The views expressed here do not necessarily reflect those of the Foundation.