Georgia's New Teacher Evaluation System Problematic in Pilot

Georgia’s New Teacher Evaluation System Problematic in Pilot

The trial run of Georgia’s new teacher assessment system is not returning the results that lawmakers and education experts were hoping for, the Atlanta Journal-Constitution reports. According to preliminary results made available earlier this month, only a small percentage of teachers evaluated under the new system were found to be performing below par – making [...]

The pilot covered about 5,800 teachers from around the state, and fewer than 1% received the worst scores while nearly 20% were graded as exemplary – the highest score possible. According to state education officials, this is a strong indication that the way the program is being implemented needs to be adjusted to make sure that the outcomes are realistic by the time it rolls out to the entire state during the 2014-2015 academic year.

Fran Millar, the chair of the Senate Education Committee, pointed out that the low percentage of poorly performing teachers identified in the pilot doesn’t make sense in light of the student achievement data compiled over the past few years.

Teachers in DeKalb, Gwinnett, Cherokee, Clayton, Henry and 21 other school districts participated in the pilot program, which ran from January to May 2012 and resulted in 0.32 percent of teachers being classified as ineffective, 5.95 percent as developing/needs improvement, 74.4 percent as proficient and 19.3 percent as exemplary, according to a state Department of Education report obtained by The Atlanta Journal-Constitution under the Georgia Open Records Act.

However, the numbers might seem out of whack because the ones released by the Georgia DOE don’t include the largest component of the new system – standardized test results. The formula that derives ratings from the scores is still being studied and fine-tuned. Education officials say that those numbers will be released later in the year.

The new teacher evaluations are only one of the major changes currently shaking up the state’s education system. Another proposal recently released to the public – and authored by Georgia’s Chamber of Commerce – calls for a complete overhaul of the way schools are funded by putting more power over money allocated towards education by the state into the hands of individual districts and even directly with the schools.

The individual school board or principal would control financing and have the power to offer larger paychecks to attract and retain sought-after teachers, make decisions on what supplies to purchase for classrooms and determine overall school material needs. Currently, only the state can make such changes and it must make them for all schools in the state. That process has been in use for 27 years and advocates of the proposal say that’s long enough.

Comments

Jeff Pence

January 9, 2013 at 10:34 am

Could it be that teachers are actually qualified, competent, and doing great work?

Reply
Fed up

January 9, 2013 at 11:35 am

So if only 20% of the teachers (1 out of 5) are exemplary-that is too many? If the test finds 74% of the teachers proficient- that is too many? The instrument must be flawed-right? The teachers can’t possibly be that good -right? So lets find an instrument that gives us an accceptable ratio of failure to success that we can sell so we can reform the system whether it needs it or not. It can’t possibly be the kids not doing their job? Of course, they are failing too because the standardized tests have been tweaked annaully to provide evidence that kids are failing. When their scores get too high the test has to be made more difficult. It is a no win situation for students and teachers. No matter how proficient a teacher is at methodolgy and pedagogy, content or procedures they should fail if the student fails the test or decides to blow off the test, or plays connect the dots, or whatever. What a system! I would like to know what percentage of failing teachers would be acceptable to them to justify their evalautions?

Reply
- Nathan C.
  
  January 9, 2013 at 1:42 pm
  
  I think you are misunderstanding the process by which the teacher evaluations are created – or at least the statistical aspect of it.
  
  The teacher evaluations are derived FROM the test scores of the students. If the teacher is bad, the students will (probably) have low test scores, and the teacher will have a bad evaluation. If the STUDENTS are bad (unmotivated, bad behavior, etc.), the students will have low test scores, and the teacher will still have a bad evaluation. The statistical method by which the teacher determinations are assigned assumes causality on the part of the teacher. In other words, the teacher evaluations are constructed directly from the performance of the students, and the performance of the students is ASSUMED to be a result of the quality of the teacher.
  
  Now, if the performance of ALL STUDENTS in the State of Georgia is known (which it is), and the parameters of the statistical model are known (which they are), then it is possible to produce an ESTIMATE of how many teachers should fall into each evaluation category. It’s not about adding cosmetics to the numbers so that a certain percentage of teachers fall into each category – that’s not the goal here. It’s simply making sure that the actual results match up with the predicted results, and being concerned when they don’t.
  
  For example, consider the simple equation 10 + X = Y. Now assume that X is a random number between -1 and 1, and Y is the sum of 10 and that random number. If our result, Y, turns out to be 100, we will know that something went terribly wrong because we were expecting the result, Y, to be somewhere between 9 and 11.
  
  To summarize, the problems with the system mentioned in the article aren’t about deeming a certain percentage of teachers to be failing expectations. It’s about validating the results of the evaluations with what was expected, and then being concerned when the results did not match what was scientifically expected, given the values of the data input into the model.
  
  Reply