As detailed here, new teacher evaluation systems that measure teaching effectiveness based on students' achievement test scores are riddled with serious flaws.  One of the most obvious is the likelihood that those systems will simply rate teachers with the weakest/most challenged students as the most ineffective.  The hope of some, however, is that one might be able to demonstrate that districts are assigning teachers who are, in fact, the least effective to disadvantaged schools--a serious equity concern under both federal and state law. It is hard to say definitely which of these two possibilities a new study of Chicago schools reveals, but it tends toward the later.  Either way, it is bad for Chicago Schools.  The new study by the University of Chicago Consortium on School Research finds that:

teachers with the lowest scores on the REACH Students teacher evaluation system are overrepresented in schools serving the most disadvantaged students, while teachers with the highest observation scores are underrepresented in these schools.

The study uses data from the 2013-14 school year, which represents the first comprehensive snapshot of evaluation scores for Chicago Public School teachers under the new REACH Students teacher evaluation system. This includes value-added scores based on students’ gains on tests, as well as scores from observations of teaching practices in classrooms.

It finds 26 percent of teachers with the lowest value-added scores are in schools with the highest concentrations of poverty, while 13 percent are in schools with the lowest concentrations of poverty. The differences in observation scores are more pronounced: 30 percent of the lowest-scoring teachers are found in the highest-poverty schools, while only 9 percent are in schools with the lowest poverty. In other words, observation scores have a stronger relationship with school characteristics, such as poverty, than value-added scores.

While more research needs to be done in Chicago to understand why these differences exist, other research suggests these differences could arise because it is more difficult to recruit and retain high-scoring teachers in high-poverty schools, or because it is more difficult to get a high observation score if teaching in a high-poverty school.

The report also finds teachers in schools with better organizational and learning climates tend to have higher value-added and observation scores, and these differences remain significant when comparing schools with similar student characteristics, including poverty level.

REACH and other teacher evaluation systems employ multiple measures to capture different aspects of teacher performance. Value-added scores are intended to capture student growth on test scores, and explicitly control for measures of student disadvantage, such as poverty and previous achievement. Observation ratings are intended to capture a teacher’s level of instructional practice, and do not control for any student or school characteristics, such as poverty.

The study also finds that, on average, African American, Latino, and other minority (i.e. Asian, Hawaiian/Pacific Islander, Native American, and multi-racial), teachers’ observation scores are lower than white teachers’ observation scores. However, for African American teachers, who are overrepresented in the highest-poverty schools, most of this difference seems to be due to the relationship between observation scores and school characteristics, such as school-level poverty. There were no significant differences by teacher race/ethnicity on either reading or math value-added scores.

Other key findings include: There are some differences in teachers' evaluation scores, depending on experience and credentials. Teachers with more experience have higher scores on value-added and observations than new teachers. Differences between teachers with National Board Certification or advanced degrees, compared to those without those credentials, were found only on observation scores, not value added.  Male teachers have lower observation and value-added scores than female teachers. On average, male teachers scored lower than female teachers on observations and slightly lower on value added than their female counterparts.