A new study in one school showed that children with bigger feet did substantially better on tests. Educators were amazed to find out that simply looking at feet could predict how students would rank on a test. But, the statistics don't lie.
That is, until they realized the study was comparing 3rd graders and 8th graders. If you look hard enough, you'll find someone who has a study to support your position - and a lot of studies don't survive the science and statistical test. Be sure to look really hard before you leap!
Which brings me to teacher evaluations based on test scores. To the common man on the street (and apparently our new State Superintendent for Education), comparing test scores that students achieve is a great way to evaluate/compensate their teacher. The LA newspaper did an evaluation of teachers based on this and published the results. However, a follow-up by some statisticians in Colorado, using the same data, came up with some different results.
Is one analysis better than the other? Maybe, maybe not. The real issue is both are trying to differentiate within the common cause variation of the data - ranking within common variation actually creates more variation and doesn't produce desired results. A better system is to give appropriate recognition to the top and bottom 1% using multiple data inputs.
And this is not easy to do and in many cases, teachers will get evaluated based on system problems they have no ability to change. For example, if you evaluate teachers based on parent input, does the district provide the teacher a classroom phone with voice mail to be used to communicate with parents? Postage paid cards and envelopes to be mailed to parents? Free transportation for parents to open houses, parent/teacher meetings? Baby sitting services and meals for class meetings with parents? Some school districts do this. Do you think their teacher evaluations will be higher than the districts who do not?
Then there is the variation of students - and what a variable that is! You should all know about the impact one or two students can make to the classroom environment.
Then there is the issue of comparisons within/between subjects. How do you evaluate a PE teacher with a Calculus teacher?
Don't get me wrong. The teacher evaluation and pay system is broken. We need to have the discussion and make improvements. Change that - WE don't need to have it - it needs to be done by the teachers.
Good teachers already use feedback from students and parents to make themselves better teachers. Lets be sure something doesn't get put in place that makes things worse.
That is, until they realized the study was comparing 3rd graders and 8th graders. If you look hard enough, you'll find someone who has a study to support your position - and a lot of studies don't survive the science and statistical test. Be sure to look really hard before you leap!
Which brings me to teacher evaluations based on test scores. To the common man on the street (and apparently our new State Superintendent for Education), comparing test scores that students achieve is a great way to evaluate/compensate their teacher. The LA newspaper did an evaluation of teachers based on this and published the results. However, a follow-up by some statisticians in Colorado, using the same data, came up with some different results.
Is one analysis better than the other? Maybe, maybe not. The real issue is both are trying to differentiate within the common cause variation of the data - ranking within common variation actually creates more variation and doesn't produce desired results. A better system is to give appropriate recognition to the top and bottom 1% using multiple data inputs.
And this is not easy to do and in many cases, teachers will get evaluated based on system problems they have no ability to change. For example, if you evaluate teachers based on parent input, does the district provide the teacher a classroom phone with voice mail to be used to communicate with parents? Postage paid cards and envelopes to be mailed to parents? Free transportation for parents to open houses, parent/teacher meetings? Baby sitting services and meals for class meetings with parents? Some school districts do this. Do you think their teacher evaluations will be higher than the districts who do not?
Then there is the variation of students - and what a variable that is! You should all know about the impact one or two students can make to the classroom environment.
Then there is the issue of comparisons within/between subjects. How do you evaluate a PE teacher with a Calculus teacher?
Don't get me wrong. The teacher evaluation and pay system is broken. We need to have the discussion and make improvements. Change that - WE don't need to have it - it needs to be done by the teachers.
Good teachers already use feedback from students and parents to make themselves better teachers. Lets be sure something doesn't get put in place that makes things worse.
Posts
No comments:
Post a Comment