Since Governor Malloy released Senate Bill 24, we have noticed a lot of anxiety over the idea that teacher evaluations will be partly informed by standardized tests.  S.B. 24 relies upon the teacher evaluation guidelines that were unanimously approved by the Performance Evaluation Advisory Council (PEAC), which calls for 22.5% of a teacher’s evaluation to be based on standardized testing.

Anxiety over being evaluated on the performance of students who might already be behind is a valid concern, but there is a model for using standardized testing to evaluate teachers fairly. You see, there is a difference between measuring student learning against an absolute performance standard and measuring growth-based performance (which measures each student’s growth against his or her own baseline).  While the details about the proposed evaluation model are still being worked out, we wanted to take a minute to talk about what we mean when we say “growth-based” student performance measures – because we know this phrase isn’t very clear and can be alarming.

Measuring student learning based on “growth” means that each student’s performance is measured against his or her own historical performance and the historical performance of a similar peer group, so that a predicted performance “growth” rate can be set for each student.  Then, each student’s actual growth in learning is compared with the expected performance that was determined for him or her.

Essentially, annual growth measures are determined by comparing groups of students with similar characteristics and starting points, and measuring each student’s progress relative to his or her own starting point.  According to the U.S. Department of Education definitions,  “effective” teachers will have students who demonstrate at least one grade level of academic growth, every year. So HOW can “growth” be measured fairly? (This article by The Education Trust does a good job explaining it.) Basically, if you wanted to set annual growth goals for a single student, let’s say his name is Michael, you would use data on Michael’s past performance and the past performance of students with similar characteristics to Michael (for example, third grade, low-income students).  You would use this information to set expectations for Michael’s academic growth.  Under this model, an effective teacher would help Michael to learn and demonstrate at least one year of academic growth during one school year.  Again, using the U.S. Department of Education’s definitions, a highly effective teacher would help his or her students to achieve high rates of student growth by exceeding the predicted academic growth for one school year.

It’s important to recognize that a “growth” based model treats teachers fairly and does not blame or penalize teachers for teaching students who may start out behind or have diverse learning needs.  In fact, these are the very factors that are taken into consideration when determining academic growth measures for students. It’s also important to remember that under the PEAC guidelines student achievement, as measured by a standardized test, counts for 22.5% of a teacher’s evaluation, with other measures of student learning, professional practice and professional activities comprising the majority of the evaluation.

A number of states, such as Tennessee, Louisiana and Colorado have adopted or implemented similar “growth” based systems to determine a teacher’s impact on student achievement.  We think you would agree that it is reasonable to expect that students be able to demonstrate one year of learning and growth for every year they are in school.  Similarly, we think you would agree that teachers need to be able to help students learn so that all students make at least one year of growth, every year.

5 thoughts on “Fair Evaluations for Teachers: Exploring Growth vs. Absolute Measures of Student Achievement

  1. Nadja Young says:

    After attending Gov. Malloy’s New Haven Town Hall meeting on 3/14, I think this is the exact topic teachers need to better understand in order to ease their fears. For so long, teachers have looked at (and been unofficially measured by) a plethora of achievement data. It is a big shift for us to understand how to intepret student growth data. I was a teacher in NC when we went through this same transtion. I now work for SAS Institute, which is TN’s partner in measuring student growth (as you referenced in your blog).

    I wrote a similar blog post back in October defining the differences between “achievement” and “progress.” I hope it can be helpful in this conversation:

  2. Lori Geraci says:

    This model does not seem to account for children who don’t yet have a diagnosis but have a disability or those who have a diagnosis but are removed for so many special classes that they miss the regular school work and assignments. It also does not account for those whose parents don’t participate actively or ignore the teacher and school reccommendations or children who are medicated or not medicated and that impact on their learning abilities. There are too many grey areas that will impact the outcome to hinge someone’s salary and job status.

    • Lyndsay says:

      I agree. Children are not machines where if you input X you will get an output of Y. There are so many other factors involved that would prevent a child from making one year of progress in a school year. I am not saying it is imposible, but I am saying that congress should stop looking at children so one dimensionally. If any of them spent any time in a classroom, they would understand this.

      • Mike says:

        It is true that that growth models are not perfect (which is why they are not advocated to be the sole determinant of teacher evaluation systems), but in most cases it seems reasonable to expect a classroom of students to know more after leaving a classroom than they did coming into it. That’s the fundamental concept behind growth. And the models are sophisticated enough such that the issues mentioned by Lori do not render them useless.

  3. Mary says:

    In our school all of the special education students are put in two classrooms. Then they put the at risk children in these two classrooms(not id yet) for extra help. Now let’s add in ESL and behavior/emotional issues. The other two classrooms don’t have these students. How will this be accounted for? How about administrator favoritism? Teachers have no control over their class make up.

Comments are closed.