Interim Assessment Data: Setting Growth Targets and Understanding Scores

Setting-Growth-TargetsThere are some great conversations that take place in our SPARK Community Forums, and recently there was one question that deserved repeating. We get this question quite a bit from our MAP users and our own Eric Merchant does a great job explaining the answer using our 2011 Norm Study FAQ:

How are growth targets created and why do we see flat growth scores across grades at times?

The 2011 growth norms are, in many grades, flatter across the achievement distribution than the 2008 norms. Most of the flattening can be attributed to changes in the growth estimates for very high and very low performing students. For nearly all grades and subjects, the differences between the 2008 and 2011 growth norms are one point or less for any student performing between the 5th and 95th percentile.

Our point here is that for students between the 5th and 95th percentile, the 2008 growth norms already showed the flattening that is more visible in the 2011 norms. For low students, the 2008 norms, which relied on observed score, gave a lot of weight to students whose fall scores were likely deflated.

For example, in 9th grade math, students with a starting RIT score of 197 (the first percentile) had an average growth of about 5 for fall to spring and a 3 for the corresponding fall to winter. A large proportion of these students, however, do not really perform that poorly. Many of them simply didn’t try. Thus they had larger gains if they gave effort on the spring test. The 2011 norms estimate the growth of low children by using the growth patterns of all children. In doing so, the model gives less weight to low performance that is due to lack of effort.

In this respect, we believe the new model actually better reflects the growth that might be expected from low achieving students who are giving their best effort on both tests. Neither the 2008 nor the 2011 norms provide reasonable growth estimates if the student didn’t give effort on the first test (in these cases one should retest).

As for high scoring students, the growth averages are slightly greater, and we believe the difference may reflect a number of factors. One of these is a concerted effort on our part over the last three years to improve the measurement range of the assessments by adding items at the high end of the achievement spectrum so that students can more accurately demonstrate their learning. We do know that the measurement range of our assessments has increased because of this effort and that seems to be reflected in higher growth averages among high performers.

I like to spend more time focusing on what we do with these growth targets and how we can impact student achievement. When setting individual growth targets for specific students we want to look at our first data point. This is generally fall, so I will use that as our first data point. Take a look at the subject, grade, the RIT score and the percentile rank the student is in and note it. For this example, let’s say the Student (John) is in the 5th grade, we are looking at math, his RIT was 200 which places him in the 19th percentile. The growth target set would be approximately 7 RIT points from fall to spring. If John meets his expected growth target then he will have shown about 1 year of instructional growth, which is wonderful, however, he would start the next school year at the 19th percentile which to me is problematic. In this case we would need to help John achieve a higher percentile rank in the spring by setting a stretch goal.

Let’s say we want John to go from the 19th percentile to the 25th percentile thus improving his overall understanding of math and having him end the school year in a better place than he began. If we apply this thought process to our goal setting it takes the 7 point RIT growth target and creates a stretch goal of 11 RIT points. If we apply this methodology through the test terms with students, especially struggling students, we can help make a profound impact on their overall instructional level. I personally like to see this because I feel it’s important for students to leave my school in a better place in regards to the learning then when they started.