Wednesday, June 3, 2009

MCAS Scaled Scores

Over at Gotham Schools, Skoolboy, the resident educational researcher and statistician, has been talking a lot about how the most statistically valid way of comparing test data is to look at average scaled scores, instead of proficiency rates. He argues that proficiency rates can be misleading because they don't tell you if kids are near the top or bottom of each category range. In New York, they publish both proficiency rates and average scaled scores for each test in each grade.

In Massachusetts, the DOE publishes the percentage and numbers of students at each score range, which they also compile into a number called the CPI. I've been looking around on the DOE website to see if I can find scaled score data with no luck. If someone knows how to find this data, it would be great if you could post it below.

Update: From Bob Lee, who is MCAS Chief Analyst at the MAESE, responded on the comments page:

"We don't have an interval scale in Massachusetts. In other words the difference between a 216 and a 218 is larger than the difference between a 222 and a 224 so our Technical Advisory Committee (and I agree) doesn't want us to compute average scaled scores. Use the CPI, it has 5 intervals that are equal. Otherwise you have to use medians or the scaled score equivalent of the average scaled score."

Update II: My understanding of CPI is that is separates kids into 5 groups (Advanced and Proficient, High Needs Improvement, Low Needs Improvement, High Warning and Warning). The problem that I see is that it does a good job of accounting of sorting the kids and schools on the low end of the spectrum but does not allow for differentiating schools on the top end. Here's an example: School A has 80% of kids scoring proficient and 20% advanced on the 8th grade math test. School B has 20% of kids scoring proficient and 80% advanced on the same test. Which school and kids are performing better at math? CPI says they are equal; I say they're not. CPI sets an important, but low bar. As a teacher, I want all my kids to be proficient or advanced, but I want the advanced section as high as possible, too.

2 comments:

  1. Sorry Skoolboy,
    We don't have an interval scale in Massachusetts. In other words the difference between a 216 and a 218 is larger than the difference between a 222 and a 224 so our Technical Advisory Committee (and I agree) doesn't want us to compute average scaled scores. Use the CPI, it has 5 intervals that are equal. Otherwise you have to use medians or the scaled score equivalent of the average scaled score.

    Bob Lee, MCAS Chief Analyst

    ReplyDelete
  2. Thanks Bob for the quick info.

    FYI, Skoolboy writes over at Gotham Schools. I'm "mathteacher."

    ReplyDelete