In assigning student growth ratings, GreatSchools works to ensure that all schools which serve similar grade students are compared equally across a state. We believe that when measured over time, student growth is as important to academic achievement as test scores.
While student growth rating methodologies vary from state to state, in every state we normalize the student growth rating into a percentile-based ranking per school, before assigning a 1-10 rating based on that percentile. For example, a school in the top 10% across the state for student growth may be assigned a 9/10 student growth rating (although this is not a guarantee). Student growth percentiles are compared between schools across an entire state before a GreatSchools student growth rating is applied.
Depending on the state-wide level of student growth, it is possible for a school with high test scores to have a lower student growth rating when compared to other schools. If a school's year-on-year student growth percentile drops, then their growth rating may drop too, if other schools have maintained their level of growth.
Here’s an example. In the 2013-2014 school year, a student attending School A achieves 1100 on a single test score. In the 2014-2015 school year, the same student achieves 1200 on the same test - growth of 100 points.
A student attending School B – which is similar in student makeup, has the same grades tested and so on - achieves 900 on the same test in the 2013-2014 school year. In the 2014-2015 school year, the same student achieves 1100 on the same test – growth of 200 points.
If this trend is reflected across other students at School B, even though their test scores may be in a similar range to School A, the overall growth rating at School B is higher. Remember, the two schools are similar in all other ways.
In the end School A is in the 40-50 percentile range for student growth ratings for the state. This becomes a 5/10 student growth rating on GreatSchools.
School B – even though our theoretical student got lower test scores – has seen a higher rate of growth year-on-year. In theory, the same student attending School B will see better academic growth than if they were attending School A. School B ultimately is placed in the 70-80 percentile range for student growth ratings in the state. This becomes an 8/10 student growth rating on GreatSchools.
Using the student growth rating, we combine it with the test score rating to give a final GreatSchools overall rating for a school. Again, test score ratings are comparable across similar schools in the state.
In the case of School A – which had slightly higher test scores than School B – they have a test score rating of 9. School B had a test score rating of 8.
To get the final GreatSchools rating, we calculate the average like so:
Growth Rating + Test Score Rating / 2 = Final GreatSchools rating
School A: (5 + 9) / 2 = 7
School B: (8 + 8) / 2 = 8
As you can see, in this instance our overall rating for School B is slightly higher, based on their higher growth rating. However, it’s important to remember that a ‘middling’ student growth rating at a school (such as School A) does not mean that the school is average, or even poor. As this rating compares across the entire state and to similar schools, in comparison to those schools specifically, the school may be lower than some, but outperforming others. It is possible that many schools are 'clustered' in a similar range for test scores across a state, and this can mean student growth ratings can be similarly clustered.
Student growth ratings are a new metric for GreatSchools, and as such, they will become increasingly useful over time, as we gain more data to compare against. However, we still believe that growth ratings are a very important measure of student success and should be considered when making choices about a child’s school.