Making Effective Use of Paraprofessionals


The Education Endowment Foundation (EEF) in England has just released its first six reports of studies evaluating various interventions. In each case, rigorous, randomized evaluations were done by third parties. As is typical in such studies, most found that treatments did not have significant positive outcomes, but two of them did. Both evaluated different uses of paraprofessionals. In England, as in the U.S., paraprofessionals usually assist teachers in classrooms, helping individual students with problems, helping the teacher with classroom management, and “other duties as assigned.” As in the U.S., teachers, parents, and politicians like paraprofessionals, because they are usually nice, helpful people from the community who free teachers from mundane tasks so the teachers can do what they do best. Unfortunately, research in both countries finds that paraprofessionals make no difference in student learning. The famous Tennessee Class Size study, for example, compared larger and smaller classes, but also had a large-class-with-paraprofessional condition, in which student achievement was precisely the same as it was in the large classes without paraprofessionals.

In one of the recent EEF-funded evaluations, teaching assistants taught struggling secondary readers one-to-one 20 minutes a day for 10 weeks. The study involved 308 middle schoolers randomly assigned to tutoring or ordinary teaching in 19 schools. The tutored students gained significantly more in reading than did controls. Similarly, a studyin which 324 elementary students in 54 schools were randomly assigned to one-to-one tutoring in math or to regular teaching found that the tutored students gained significantly more.

The EEF reports add to a considerable body of research in the U.S. showing that well-trained paraprofessionals can obtain substantial gains with struggling readers in one-to-one and small-group tutoring.

What these findings tell us is crystal clear. Already in our schools we have a powerful but underutilized resource, paraprofessionals who, with training and assistance, could be making a substantial difference in the lives of struggling students. This resource is costing us a lot. Most of the $15 billion we spend on Title I every year is spent on paraprofessionals, as is a lot of state and local funding. From personal experience, paraprofessionals are caring and capable people who want to make a difference. Why not use the evidence to help them do just that?


School Improvement Grants Embrace Evidence of Effectiveness


Despite all of the exciting gains made by evidence-based reform in recent years, all of the progress so far has been limited to development, evaluation, and scale-up of proven programs outside of mainline education policy or funding. Title I, Title II, Race to the Top, School Improvement Grants, and other large, influential funding sources for reform have hardly been touched by the growth of proven, replicable programs sponsored primarily by the Institute of Education Sciences and Investing in Innovation (i3). Until the evidence movement crosses over from R&D to the real world of policy and practice, it will remain the domain of academics and policy wonks, not a real force for change.

In the recently passed Omnibus budget, however, appears a first modest step over the R&D/policy divide. This is a new provision in congressional authorization of School Improvement Grants (SIG). Up until now, SIG schools (ones that have suffered from very low achievement levels for many years) had to choose among four models, all of which require major changes in staffing. Each SIG school is expected to develop its own model of reform, usually with the help of consultants. The problem has been that each of the hundreds of schools receiving (substantial) SIG funding has to create its own never-before-tested path to reform, and then try to implement it with quality in a school that has just experienced a substantial turnover of its leadership and staff.

The “Fifth Option” recently introduced by Congress adds a new alternative. SIG schools can choose to adopt a “proven whole-school reform model” that meets at least a moderate level of evidence support, which includes having been tested against a control group in at least one rigorous experiment. The fifth option will let schools keep their leaders and staffs, but adopt a schoolwide approach that has been used in many similar schools and found to be effective.

The Omnibus bill was passed too late in the year to apply this fifth option to the 2014-2015 school year, and the U. S. Department of Education, as well as individual states, have a lot of work to do to prepare new regulations and supports for schools applying for SIG funds under this new option in 2015-2016.

However, the fifth option makes an important statement that has not been made previously. In a major school improvement (not R&D) funding program, the fifth option says “use what works.” Wisely, it does not mandate the use of any specific programs, but by highlighting evidence-proven approaches, it puts the government behind the idea that federal funding should whenever possible be used to help educators use programs with strong evidence of effectiveness. This could be the start of something beautiful.

Preschools and Evidence: A Child Will Lead Us


These are exciting times for people who care about preschool, for people who care about evidence, and especially for people who care about both. President Obama advocated for expanding high-quality preschool opportunities, Bill de Blasio, the new Mayor of New York City, is proposing new taxes on the wealthy for this purpose, and many states are moving toward universal preschool, or at least considering it. The recently passed Omnibus Budget had $250 million in it for states to add to or improve their preschool programs.

What is refreshing is that after thirty years of agreement among researchers that it’s only high-quality preschools that have long-term positive effects, the phrase “high quality” has become part of the political dialogue. At a minimum, “high quality” means “not just underpaid, poorly educated preschool teachers.” But beyond this, “high quality” is easy to agree on, difficult to define.

This is where evidence comes in. We have good evidence about long-term effects of very high-quality preschool programs compared to no preschool, but identifying exceptionally effective, replicable programs (in comparison to run-of-the-mill preschools) has been harder.

The importance of identifying preschool programs that actually work is being recognized not only in academia, but in the general press as well. In the January 29 New York Times, Daniel Willingham and David Grissmer advocated local and national randomized experiments to find out what works in preschool. On January 30, Nicholas Kristof wrote about rigorous research supporting long-term effects of preschool. Two articles on randomized experiments in education would be a good week for Education Week, much less the New York Times.

With President Obama, John Boehner, and the great majority of Americans favoring expansion of high-quality preschools, this might be an extraordinarily good time for the U.S. Department of Education to sponsor development and evaluation of promising preschool models. At the current rate it will take a long time to get to universal pre-K, so in the meantime let’s learn what works.

The U. S. Department of Education did such a study several years ago called Preschool Curriculum Evaluation Research (PCER), in which various models were compared to ordinary preschool approaches. PCER found that only a few models did better than their control groups, but there was a clear pattern to the ones that did. These were models that provided teachers with extensive professional development and materials with a definite structure designed to build vocabulary, phonemic awareness, early math concepts, and school skills. They were not just early introduction of kindergarten, but focused on play, themes, rhymes, songs, stories, and counting games with specific purposes well understood by teachers.

In a new R & D effort, innovators might be asked to create new, practical models, perhaps based on the PCER findings, and evaluate them in rigorous studies. Within a few years, we’d have many proven approaches to preschool, ones that would justify the optimism being expressed by politicians of all stripes.

Historically, preschool is one of the few areas of educational practice or policy in which politicians and the public consider evidence to have much relevance. Perhaps if we get this one right, they will begin to wonder, if evidence is good for four year olds, why shouldn’t we consult it for the rest of education policy? If evidence is to become important for all of education, perhaps it has to begin with a small child leading us.