Recent comments

Search News & Views

Why teacher ratings don't tell much

The latest serving of data-driven mania from the city Education Department will likely produce screaming headlines about the city's "worst teachers." This virtual wall of shame (and fame) will live online for years to come. But does it actually help parents to find the best schools and teachers? Not really. Here's why.

The ratings are based on a complicated formula that compares how much 4th through 8th-grade students have improved on standardized tests compared with how well they were predicted to do. The system tries to take into consideration factors like race, poverty and disabilities. Teachers are then graded on a curve. It's known as "value-added," because it tries show how much value an individual teacher has added to a student's test scores.

Here are our top five reasons they won't help and why you won't be seeing them on Insideschools. Please add your own, or tell us why you think they will be useful.

1. The ratings are based on exams that state officials have said are invalid.

The reports being released are from 2007, 2008 and 2009, before state officials altered exams that they said were not a reliable indication of whether or not students were learning. The exams were only testing a small part of what students were supposed to know, and it was easy to predict what would be on the exam each year. So students who were drilled in a narrow set of skills might do well and their teacher might be rated highly. Teachers who were teaching the whole curriculum and not focused on test prep could be rated lower, even if their children had in fact learned more.

2. Test scores alone don't tell you how effective a teacher is.

The exams don't measure a student's critical thinking skills, creativity or if she works well with others, three things that many of the best teachers emphasize. Although the DOE tried to account for differences in student populations (such as poverty, disability, English language learner), it is not at all clear that they were able to measure all of the differences between classrooms that could affect scores.

3. The margin of error on the ratings is huge.

The DOE admits that a teacher whose rating is 50% (or about average) could actually have a rating as low as 25% or as high as 75%. Even though a teacher is assigned a score, the report includes a range of possible scores because the DOE acknowledges that the reports are imprecise.

4. Teachers of children performing well on the exams could be rated poorly.

The ratings are based partially on student "progress" on the exams. So if students score 3.8 out of 4 one year and then drop to 3.6 the next year, a teacher could be rated ineffective. This is true even though a 3.6 is considered above grade level, and the difference in the score could be a matter of getting one or two questions wrong.

5. The ratings are not stable from year to year.

If the measure was accurate, you would think that a good teacher wouldn't get a 20% one year and an 80% the next. But an Annenberg study found that a third of the English teachers who got the top rating in 2007 sank to the bottom of the pile in 2008. The same was true for 23% of math teachers.

Last modified on Friday, 24 February 2012 13:58

Please post comments

  • Give specific examples. Tell us why “this school rocks” (or doesn’t)
  • No profanity. No racial or ethnic slurs. No personal attacks
  • Criticism is fine but don’t be nasty.
  • Flag inappropriate comments. (Hover your cursor over comments to see flag)