Those Who Can, Teach—Those Who Don’t Understand Teaching, Make Policy

There is an old adage that many in education have undoubtedly (and unfortunately) encountered at some point in their education careers: that “those who can, do—those who can’t, teach.” For decades now, researchers, including Dr. Thomas Good who is the author of an article published in 2014 in Teachers College Record, have evidenced that, contrary to this belief, teachers can do, and some teachers do better than others. Dr. Good points out in his historical analysis, What Do We Know About How Teachers Influence Student Performance on Standardized Tests: And Why Do We Know so Little About Other Student Outcomes? that teachers matter and that teachers vary in their effects on student achievement. He provides evidence from several decades of scholarly research on teacher effectiveness to show that teachers do make a difference in student achievement as measured by large-scale standardized achievement tests.

Dr. Good is also quick to acknowledge that, despite the reiterated notion that teachers matter and thus should possess (and continue to be trained in) effective teaching qualities (e.g., be well versed in their content knowledge, have strong classroom management skills, hold appropriate expectations, etc.), “fad-driven” education reform policies (e.g., teacher evaluation polices that are based in large part on student achievement growth or teachers’ “value-added”) have gone too far and have actually overvalued the effects of teachers. He explains that simplistic reform efforts, such as Race to the Top and VAM-based teacher evaluation systems, overvalue teacher effects in terms of the actual levels of impact teachers have on student achievement. It has been well documented that teacher effects can only explain, on average, between 10-20% of the variation in student achievement scores (this will be further explored in forthcoming posts). This means that 80-90% of student achievement scores are the result of other factors that are completely outside of the teachers’ control (e.g., poverty, parental support, etc.). Regardless, new teacher evaluation systems that rely so heavily on VAM estimates ignore this very important fact.

Dr. Good also emphasizes that teaching is a complex practice and that by attempting to isolate the variables that make up effective teaching and focusing on each one separately only oversimplifies the complexities of the teaching-achievement relationship. For one thing, VAMs and other popular evaluative practices, such as classroom observations, inhibit one’s ability to recognize the patterns of effective teaching and instead promote a simplistic view of just some of the individual variables that actually matter when thinking about effective teaching. He provides the following example:

Many observational systems call for the demonstration of high expectations. However…expectations can be too high or too low, and the issue is for teachers to demonstrate appropriate expectations. How then does a classroom observer know and code if expectations are appropriate both for individual students and for the class as a whole?

The bottom line is that, like teaching, understanding the impact of teachers on student learning is very complex. Teachers matter, and they should be trained, treated, and valued as such. However, they are one of many factors that impact student learning and achievement over time, and this very critical point cannot be (though currently is) ignored by policy. VAM-based policies, particularly, place an exorbitantly over-estimated value on the impact of teachers on student achievement scores.

Post contributed by Jessica Holloway-Libell

Arizona’s Teacher Evaluation System, Not Strict Enough?

Last week, Arizona State Superintendent of Public Instruction, John Huppenthal, received the news that Arizona’s No Child Left Behind (NCLB) waiver extension request had been provisionally granted with a “high-risk” label (i.e., in danger of being revoked). Superintendent Huppenthal was given 60 days to make two revisions: (1) adjust the graduation rate to account for 20% of a school’s A-F letter grade instead of the proposed 15% and, as most pertinent here, (2) finalize the guidelines for the teacher and principal evaluations to comply with Elementary and Secondary Education Act (ESEA) Flexibility (i.e., the NCLB waiver guidelines).

Within 60 days, Superintendent Huppenthal and the Arizona Department of Education (ADE) must: (1) finalize its teacher and principal evaluation guidelines; (2) give sufficient weighting to student growth so as to differentiate between teachers/principals who have contributed to more/less growth in student learning and achievement; (3) ensure that shared attribution of growth does not mask high or low performing teachers as measured by growth; and (4) guarantee that all of this is done in time for schools to be prepared to implement for the 2014-2015 school year.

These demands, particularly #2 and #3 above, reflect some of the serious and unavoidable flaws with the new teacher evaluations that are based on student growth (e.g., and all other VAMs).

As per #2, the most blatant problem is with the limited number of teachers (typically around 30%, although reported as only 17% in the recent post about DC’s teacher evaluation system) who are eligible for classroom-level student growth data (i.e., value-added). Thus, one of the key expectations—to ensure sufficient weight to student growth scores so as to differentiate between teachers’/principals’ impact on student learning and achievement—is impossible for probably around seven out of every ten of Arizona’s and other states’ teachers. While most states, including Arizona, have chosen to remedy this problem by attributing a school-level (or grade-level) value-added score to classroom-level ineligible teachers (sometimes counting as much as 50% of the teacher’s overall evaluation), this solution does not (and likely never will) suffice as per #2 written above. It seems the feds do not quite understand that what they are mandating in practice leaves well over half of teachers’ evaluations based on both students and/or content that these teachers didn’t teach.

As per #3, Arizona (and all waiver-earning states) is also to demonstrate how the state will ensure that shared attribution of growth does not mask high or low performing teachers as measured by growth. Yet, again, when these systems are implemented in practice, 70+% of teachers are assigned a school-level student growth score, meaning that all teachers in any given school who fall into this group will all receive the same score. In what way is it feasible to “ensure” that no high or low performing teacher is “masked” by such a method of attributing student growth to teachers in this way? Yet this is another example of the type of illogical circumstances by which schools must abide in order to meet the arbitrary (and often impossible) demands of ESEA Flexibility (and Race to the Top).

If Arizona fails to comply with the USDOE requests within 60 days, they will lose their ESEA waiver and face the consequences of NCLB. In a statement to Education Week, however, AZ Superintendent Huppenthal stood by his position on providing school districts with as much flexibility as possible within the constraints of the waiver stipulations. He said he will not protest the “high risk” label and will instead attempt to “get around this and still keep local control for those school districts.” The revised application is due at the end of January.

Post contributed by Jessica Holloway-Libell