Texsox Posted April 29, 2011 Share Posted April 29, 2011 The state of Texas just changed how teachers are evaluated. I don't mind accountability and being judged on how well I do or don't do my job. But, evaluating teachers, and comparing them to other teachers based on how their students do on standardized tests, is fraught with danger. Students are not randomly assigned to school districts, campuses, or classrooms. The majority of my students have parents who never graduated from high school. On my campus I have one Pre-Ap course, another teacher has none, and two others have two. Every student takes the same test. How do you compare any of our results against each other, then compare it to the really wealthy districts like where Kap lives. Using a local Illinois example, does anyone believe that there is not a difference in students between New Trier and Ford Heights? How do you compare those student's test scores and decide which teacher is more effective? Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Tex @ Apr 29, 2011 -> 01:01 PM) The state of Texas just changed how teachers are evaluated. I don't mind accountability and being judged on how well I do or don't do my job. But, evaluating teachers, and comparing them to other teachers based on how their students do on standardized tests, is fraught with danger. Students are not randomly assigned to school districts, campuses, or classrooms. The majority of my students have parents who never graduated from high school. On my campus I have one Pre-Ap course, another teacher has none, and two others have two. Every student takes the same test. How do you compare any of our results against each other, then compare it to the really wealthy districts like where Kap lives. Using a local Illinois example, does anyone believe that there is not a difference in students between New Trier and Ford Heights? How do you compare those student's test scores and decide which teacher is more effective? Improvement rates would be the biggest way. In general, I think that is a terrible idea to use a standardized test to judge how good of a teacher someone is. I support evaluations being done by in-building administration, just like any other job would be done. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 I believe the national standards are changing. I know Illinois is undergoing large changes right now. Link to comment Share on other sites More sharing options...
NorthSideSox72 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 01:03 PM) Improvement rates would be the biggest way. In general, I think that is a terrible idea to use a standardized test to judge how good of a teacher someone is. I support evaluations being done by in-building administration, just like any other job would be done. I agree with this. The school is a business or an office in many ways, and evaluation of staff should be handled the same way. Heck, do some 360's even. Test scores can be a part of that, but I also agree that improvement should be what is looked at. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 The problem still is linking accountability to the Federal dollars with a national standard of evaluation, right? Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (StrangeSox @ Apr 29, 2011 -> 01:45 PM) The problem still is linking accountability to the Federal dollars with a national standard of evaluation, right? There is my biggest problem right there. I hate the state and federalization of education. It drives me crazy. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 Need to get the funding from somewhere Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (StrangeSox @ Apr 29, 2011 -> 01:53 PM) Need to get the funding from somewhere At worst funding needs to go to the state level for redistribution. There is zero need for school systems to send their tax money to Washington to get it sent back to them. None at all. Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 02:03 PM) Improvement rates would be the biggest way. In general, I think that is a terrible idea to use a standardized test to judge how good of a teacher someone is. I support evaluations being done by in-building administration, just like any other job would be done. Improvement rates is tough also...because studies have shown that the education process really isn't consistent on that either. Fairly often, the results from going through a couple of highly rated teachers won't show up in students for 2-3 years (especially in the middle grades, where most kids have roughly equal success until everyone plummets off a cliff around 8th grade). Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Balta1701 @ Apr 29, 2011 -> 02:33 PM) Improvement rates is tough also...because studies have shown that the education process really isn't consistent on that either. Fairly often, the results from going through a couple of highly rated teachers won't show up in students for 2-3 years (especially in the middle grades, where most kids have roughly equal success until everyone plummets off a cliff around 8th grade). The measure comes against peers, so those ups and downs wouldn't matter. You'd be compared against people in the same boat. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 02:37 PM) The measure comes against peers, so those ups and downs wouldn't matter. You'd be compared against people in the same boat. But, if Balta's correct, it means your performance is really being judged on what the teacher two or three years ago did, and the teacher a couple years ahead of you is getting judged on your teaching. That doesn't seem right, because you'd never actually be judging the person based on their own performance. Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (StrangeSox @ Apr 29, 2011 -> 02:39 PM) But, if Balta's correct, it means your performance is really being judged on what the teacher two or three years ago did, and the teacher a couple years ahead of you is getting judged on your teaching. That doesn't seem right, because you'd never actually be judging the person based on their own performance. How they improve while they are in your classroom is absolutely fair. Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 03:43 PM) How they improve while they are in your classroom is absolutely fair. You clearly missed the point of that argument. Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 Try number 2 with actual data. Regardless of the type of school, income level, race of the student, etc., there is a general trend. Performance on standardized test scores for the average student are essentially flat through 8th grade, at which point, for every class of student, there is a very large dropoff. Now, it's worth noting that there are differences between groups, obviously, but the trend is overall the same. Having a great teacher in 1st grade, 3rd grade, 5th grade, whatever, doesn't produce marginal increases in average performance across a district. The districts performance stays roughly constant until 8th grade until scores go over a cliff. In other words, if you had the best teachers in the wordl in 5th, 6th, and 7th grade, the effect might be to avoid going over the cliff in 8th grade, rather than seeing any improvement in 5th-7th grade. Here is that data for Milwaukee (MPS = Milwaukee Public Schools). This was available this week because the governor of Wisconsin decided he didn't want to publish this data any more since the data didn't show that charter schools were a priori better than public schools. Link to comment Share on other sites More sharing options...
NorthSideSox72 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 01:49 PM) There is my biggest problem right there. I hate the state and federalization of education. It drives me crazy. Mostly agree. The local schools are now heavily dependent on non-local funding, which is a problem all around. State is a little different though, I think schools particularly in poor areas need support to level the playing field funding-wise. Alternately, you re-do the whole funding system state-wide towards that end. But the Feds' role in education should be only to provide: --Basic skills standards, as a purely advisory role --Safety and regulations around schools to ensure all schools are a safe environment at the basic level (safe buildings, basic security, health standards, etc.) All else should be state and local, mostly local. The schools, the students and the government will all be better off this way. Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Balta1701 @ Apr 29, 2011 -> 02:48 PM) You clearly missed the point of that argument. Not at all. When you receive a child in class, they have their score history. When they again test you are able to reflect how they did during the time period in your classroom. That rate of change is compared to all other relative rates of change. Then it is easy to compare relative levels of students anywhere across the spectrum to how other teachers of the same classes/levels did. It would also be really easy to pick out trends such as people from a certian teacher doing better down the road over the course of years in an individual school system setting. It is a super-simple data set really. You could break out any first grade teacher, for example, and plot their kids success over the future, and compare that to another first grade teachers history. It wouldn't even be hard to build an algorithim to picks trails of teachers (having teacher A for K, teacher B for 1st, teacher C for 2nd) to see what combinations are the most successful, and if certian combinations conflict with each others styles, or compliment each others styles more or less than any others. The programming would take zero time to write, and would require only a new data dump with every testing cycle. Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 Oh come on man, first of all, no one is waiting 10 years to build up that kind of data set for every teacher. Teachers wouldn't stay in the position if they were going to get fired for not having been good enough 10 years ago. Secondly, that math problem is very ill posed. You've got 20 data points per year, each one jumbled up, with the additional factor of students taking classes from 5-10 different teachers once they get to middle school. Throw in additional factors like geographic trends, that math is very difficult to discern anything from. Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Balta1701 @ Apr 29, 2011 -> 02:59 PM) Oh come on man, first of all, no one is waiting 10 years to build up that kind of data set for every teacher. Teachers wouldn't stay in the position if they were going to get fired for not having been good enough 10 years ago. Secondly, that math problem is very ill posed. You've got 20 data points per year, each one jumbled up, with the additional factor of students taking classes from 5-10 different teachers once they get to middle school. Throw in additional factors like geographic trends, that math is very difficult to discern anything from. You don't need 10 years to find out that kind of stuff. Trends emerge right away. You can tell very quickly if the initial returns are worse than peers, when all of the variables are put into consideration. The math isn't that complex at all. These kind of algorithms are located on every single trading firms website with way more variables included. Hell the climate change math is a million times more complex than what I am talking about, yet we have to take that as gospel. Anyways, that brings me back to my preferred option, which is in house administrators making these decisions, and then everyone b****es about how teachers can get fired for "nothing". Its either that or keeping it as impossible as possible to fire teachers, which is what the whole point of fighting progress in evaluations is about anyway. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 02:43 PM) How they improve while they are in your classroom is absolutely fair. Not if the improvement in your classroom is largely a signal from the 3rd grade teacher's ability. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 03:06 PM) You don't need 10 years to find out that kind of stuff. Trends emerge right away. You can tell very quickly if the initial returns are worse than peers, when all of the variables are put into consideration. The math isn't that complex at all. These kind of algorithms are located on every single trading firms website with way more variables included. Hell the climate change math is a million times more complex than what I am talking about, yet we have to take that as gospel. Anyways, that brings me back to my preferred option, which is in house administrators making these decisions, and then everyone b****es about how teachers can get fired for "nothing". Its either that or keeping it as impossible as possible to fire teachers, which is what the whole point of fighting progress in evaluations is about anyway. By very definition trends cannot emerge right away. Sample sizes here are small and spread over years and decades Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 04:06 PM) You don't need 10 years to find out that kind of stuff. Trends emerge right away. You can tell very quickly if the initial returns are worse than peers, when all of the variables are put into consideration. The math isn't that complex at all. These kind of algorithms are located on every single trading firms website with way more variables included. Hell the climate change math is a million times more complex than what I am talking about, yet we have to take that as gospel. Anyways, that brings me back to my preferred option, which is in house administrators making these decisions, and then everyone b****es about how teachers can get fired for "nothing". Its either that or keeping it as impossible as possible to fire teachers, which is what the whole point of fighting progress in evaluations is about anyway. As I just demonstrated, no they don't. The other thing that your trading firms and climate sensors have is data. A good teacher in a good school system gives you 20 data points a year. Then you move them around the next year and you get some splitting up of students, and so you wind up with 5-10 students from 1 class moving into other classes. Or, you have a program like the one I was in, where the same group of students stayed togehter for 4 years or so. You just don't get enough students to do the calculation you want to do here and have the results be meaningful. Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Balta1701 @ Apr 29, 2011 -> 03:10 PM) As I just demonstrated, no they don't. The other thing that your trading firms and climate sensors have is data. A good teacher in a good school system gives you 20 data points a year. Then you move them around the next year and you get some splitting up of students, and so you wind up with 5-10 students from 1 class moving into other classes. Or, you have a program like the one I was in, where the same group of students stayed togehter for 4 years or so. You just don't get enough students to do the calculation you want to do here and have the results be meaningful. It'd be super easy to pull data sets from first year teachers in the same relative data settings, with even being able to factor in "successful" or "not successful" previous teachers. Systems where you had less variation of students' teachers would be way easier to evaluation performance in because there would be less variables. Link to comment Share on other sites More sharing options...
southsider2k5 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (StrangeSox @ Apr 29, 2011 -> 03:09 PM) By very definition trends cannot emerge right away. Sample sizes here are small and spread over years and decades Trends can be measured in miniuta. It is done 24/7 in trading. Link to comment Share on other sites More sharing options...
StrangeSox Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (Balta1701 @ Apr 29, 2011 -> 03:10 PM) As I just demonstrated, no they don't. The other thing that your trading firms and climate sensors have is data. A good teacher in a good school system gives you 20 data points a year. Then you move them around the next year and you get some splitting up of students, and so you wind up with 5-10 students from 1 class moving into other classes. Or, you have a program like the one I was in, where the same group of students stayed togehter for 4 years or so. You just don't get enough students to do the calculation you want to do here and have the results be meaningful. Well, that depends on the grade level, but I'd say it's anywhere from 25 for elementary to 80-100 for middle school and HS. This would be a very, very complex analyses because there are a ton of variables involved here. If you're basing it on future performance, you've got to control for future teachers, changing standards, and movement of teacher, student or both to another school or district. You're also going to have major hurdles whenever there's an educational philosophy/methodology switch, such as going from elementary school to middle school to high school. Link to comment Share on other sites More sharing options...
Balta1701 Posted April 29, 2011 Share Posted April 29, 2011 QUOTE (southsider2k5 @ Apr 29, 2011 -> 04:15 PM) Trends can be measured in miniuta. It is done 24/7 in trading. With an enormous amount of data compared to what you're trying to compile with here. Link to comment Share on other sites More sharing options...
Recommended Posts