Our Failed Investments in Teacher Preparation

Our Failed Investments in Teacher Preparation
X
Story Stream
recent articles

In this Monday, Dec. 14, 2015 photo, second-grade teacher Ann Johnson works with Sullivan Murphy on a math problem at Sageville Elementary School in Sageville, Iowa. (Jessica Reilly/Telegraph Herald via AP)

RCEd Commentary

Each year, new teachers collectively spend about $4.85 billion and 302 million hours on their preparation work. But there is no evidence that any of it really matters.

Most of this money and time is spent on state-required “inputs”: things like passing certification tests and taking a certain number of hours of certain courses. States put other restrictions on the programs that prepare teachers, like requiring a minimum GPA for teacher candidates.

In theory, these requirements on inputs ensure a baseline of quality for teachers and preparation programs. In reality, research shows they amount to little more than meaningless barriers to entry.

Take the evidence on GPA and SAT scores. Some research suggests that these screens can predict teacher effectiveness, but the differences are small, and there’s no clear tipping point guiding states on where to set their expectations. The evidence on coursework and certification requirements is even weaker.

In other words, these inputs can’t guarantee that programs will produce effective teachers.

A handful of states and the federal government are moving away from using inputs to define quality. They continue to regulate inputs, but they’re shifting their focus toward a teacher’s performance after she leaves the preparation program. These states measure certain outcomes of teacher performance – like impact on student learning, job placement, retention, and evaluation rating – and link those outcomes back to the preparation program.

The idea is appealing: States loosen input requirements, give providers more freedom to design their programs as they saw fit, and then make decisions about programs on the basis of the success of their teachers. States can then differentiate among poor, satisfactory, and excellent alternative and traditional certification programs, and “consumers” — prospective teachers and potential employers—can act on that information. Early research from Louisiana and Tennessee suggested this approach was possible.

Six months ago, we started writing a paper looking at these efforts. We were as excited about outcomes as anyone could be. We thought that, with some guidance from states already doing this work and the endorsement of the Department of Education, all states could affect the quality of educator preparation, and consequently of future educators, by holding programs accountable for their completers’ performance. But we were too optimistic.

Instead of overwhelming evidence, we found conflicting research. Recent studies from Missouri and Texas suggest that completer outcomes may not differentiate preparation programs as distinctly as hoped. In Missouri, researchers reviewed three years of classroom performance records for more than 1,300 teachers, all of whom had recently graduated from one of the state’s major preparation programs. In the Texas study, researchers looked at nearly 6,300 new math teachers and 5,000 reading teachers from 100 preparation programs of all types. In both studies, researchers reached the same conclusion: the differences between programs are very small and practically indistinguishable, and almost all of the variation in teacher preparation occurs within programs. Studies of North Carolina and Washington State came to similar conclusions.

Instead of state exemplars, we found implementation challenges. The issues range from the ideological, like selecting and defining quality measures, to the technical issues of determining the right sample size and deciding if a college or university as a whole, or its individual programs, should be responsible. Because of these issues, very few states are actually using outcomes for accountability purposes. Others intend to, but haven’t yet worked out how. For example, many states would like to look at teacher preparation programs holistically by looking at how well their graduates are rated on multi-measure evaluation systems. But many of those underlying evaluation systems rate all teachers the same, making it impossible to differentiate between preparation programs.        

The implication can’t be overstated: If states can’t identify meaningful differences in teacher effectiveness between programs, it’s as good as having no information at all.

We look forward to the time when this analysis is wrong, but at this point the sobering reality is that none of the current measures -- whether inputs or outcomes -- can guarantee a teacher will be ready on Day One. We’re holding out hope that research finds new ways to more accurately predict who will become a great teacher. At that point, we would gladly endorse it as worthy of policy. As a field, the burden should be on us to find ways to help teachers prepare for the challenges of the classroom. But we're not there yet, and we shouldn't be asking candidates to bear the burden for us in the meantime.

Comment
Show commentsHide Comments

Related Articles