Teacher evaluations, part 1: To keep teachers in low-performing schools, DCPS needs to change how it evaluates them

DCPS has said it wants to invest in its 40 lowest-performing schools. If that’s the case, the school district needs to change its teacher evaluation system, which penalizes and discourages the teachers who work in them.

DCPS and the Washington Teachers Union will soon begin negotiating a new contract. As a teacher at a high-poverty DCPS school in Ward 8, I hope that one issue on the table is revising the IMPACT evaluation system. The current system creates unfair distinctions, stifles creativity, and has led to a high turnover rate among teachers at schools with the greatest needs.

DCPS introduced the IMPACT system in 2009, in an effort to provide educators with the tools necessary to become more “effective.” DCPS has made changes to IMPACT since its inception, and we’re now operating under version 2.0. But the system is still deeply flawed. Let’s hope the coming contract negotiations result in version 3.0.

The system divides employees into 20 different groups. Mathematics and English Language Arts (ELA) teachers are designated as Group 1. General education teachers, such as science and electives teachers, belong to Group 2. Special education teachers are designated as Group 3, and so on. Since DC’s standardized test, the DC CAS, only tests math and reading, Group 1 teachers are the only ones held accountable for student test scores.

As a result of their evaluations, all teachers receive one of 5 ratings, ranging from “Highly Effective” to “Ineffective.” Teachers with low ratings can be fired, and teachers with high ratings are eligible for additional compensation.

The Value Added Model

ELA and math teachers in certain grades are evaluated partly on the basis of a Value Added Model (VAM). According to DCPS, this model measures the impact of an individual teacher on students’ learning. The model calculates how a class is likely to do on the DC CAS, based on the average class score the previous year. It then compares that “likely” score with the actual score and determines how much “value” the teacher has added or subtracted.

One problem with the VAM model is that it can only apply to teachers whose students have taken the DC CAS the previous year. In a recent interview, Chancellor Kaya Henderson said that only 14% of DCPS teachers fall into this category.

But there are other problems as well, as this video explains. Although the model purports to take account of factors like students’ learning disabilities or the resources they have at home, in fact teachers’ VAM scores often vary depending on things they can’t control.

In addition, DCPS uses its VAM only to capture student growth as measured by grade-level proficiency scores on the DC CAS. Students in low-performing schools are far more likely to be below grade level than students in other schools. And if a student who is multiple grade levels behind randomly selects answers because he doesn’t understand the test questions, is that an accurate representation of a teacher’s ability to instruct?

There are fairer ways to measure a teacher’s impact. Schools already use assessments that measure actual growth, such as Lexile scores for reading and MAP for math, to diagnose students’ strengths and weaknesses. DCPS could base its VAM on these tests to evaluate teachers instead of on the DC CAS.

For example, let’s say a 6th-grade student enters middle school reading on a 2nd-grade level. If a teacher manages to bring the student up to a 4th-grade level, that should be celebrated rather than ignored. Sure, that student isn’t on grade level yet, but two years of growth is a realistic goal. Asking a teacher to bring a student up 4 grade levels in a single year is highly unrealistic.

Including test scores can make an evaluation plummet

A teacher’s rating can actually go from “minimally effective” to “highly effective” in one year, depending on whether or not DC CAS scores are included in his evaluation. I know, because it happened to me.

In 2011-12, when I was a Group 1 teacher, I scored just under highly effective on observations of my teaching and effective for my contributions to the school community (CSC). Yet, the test scores, which at the time counted for 50% of a Group 1 teacher’s evaluation, caused my overall rating to plummet to “minimally effective.” (DCPS has since changed the weight of test scores for Group 1 teachers to 35%.)

The following year I was a Group 2 teacher. I scored highly effective on observations and CSC. As a Group 2 teacher, 15% of my score was based on assessments other than the DC CAS, and I scored effective in that category. Needless to say, I finished the year with a “highly effective” rating.

So comparing Group 1 teachers in the lowest-performing schools to their counterparts in high-performing schools on the basis of “value added” is unfair. But the system also has negative consequences even within low-performing schools.

Group 1 teachers are under particular pressure because they’re the only teachers held accountable for their students’ performance on tests under the VAM model. But reading comprehension must be taught in all subjects. Why not hold all teachers accountable for a student’s Lexile score, rather than ELA teachers alone?

That approach would also encourage collaboration between Group 1 teachers and those in other disciplines. If our aim is to educate every child, we need to implement the same accountability standards across the board.

In part two of this post, I’ll look at how another component of teacher evaluations, classroom observations, can work against teachers in high-poverty schools.

A version of this post appeared on Edcentrist.com.