Growth Results and Why They Don't Grow

The 2014 overall scores released early this morning include growth results, identified by comparing students' reading and math performance to that of other students with similar scores last year. 

The state Briefing Packet for this year's accountability release says:

The student growth percentile model (SGP) is based on a normative distribution of academic peer group. Since SGP model uses a normative distribution, the percent of students statewide scoring at the typical or higher level will be consistent from year to year at approximately 60 percent. At the individual school level students scoring at typical or higher level range from 14 percent to 89 percent.

I can explain this a little further.   Kentucky's student growth percentile model identifies groups of students who had similar scores on the previous year's assessments (creating those "academic peer groups").  Then, with the current year's results, the bottom 40% within the group are noted as not making typical progress, while the rest are identified as making typical growth.  For all the groups combined, and for the state as a whole, that means the growth component will always come in right below that 60% ceiling.

An individual school can have more than 60% of students making the progress defined as typical for that year--so long as some other school has less than 60%.  Or one school can have less than 60% if another has more.  But when looking at statewide results, it's not a fluke that the results are going to be very similar every year: that's exactly how the model is designed to work.

--Posted by Susan Perkins Weston