Sunday, January 30, 2011

More on where Kentucky really ranks in new federal science testing

Whenever I discuss how Kentucky ranks with other states in federal National Assessment of Educational Progress testing, I always point out that it isn’t appropriate to simplistically compare our scores for all students to scores for all students in other states. The reason is that Kentucky’s classrooms remain about 85 percent white, while most other states now have much higher percentages of minorities in their total enrollment figures. Because minorities score much lower than whites, simplistic comparison of overall scores just provides Kentuckians with a false sense of our real performance.

One way to get around this problem, which even NAEP publications discuss, as I mentioned several days ago, is to break out the NAEP scores by race. With so many whites in Kentucky, it is clearly important to see how our white students do against their peers. I showed you a map of that several days ago here.

However, whenever I show people such maps, they always respond that our whites are poor and want to see a comparison of scores for poor whites only. OK, here that comparison for proficiency rates in the various states for NAEP Grade 8 Science from 2009:


States in green (3) got scores that were statistically significantly higher than we did. States in tan got scores that the rules of statistics declare are basically tied with us. Those states in salmon color got statistically significantly lower scores.

A total of 46 states participated (Note: A news release from the Kentucky Department of Education had a misleading comment that all states participated in NAEP Science. That isn’t correct. All states participated in math and reading only in 2009).

So, given the statistical limitations of this sampled test, what do we know?

Kentucky’s poor whites were only outscored (with a high level of certainty) by three states on NAEP Grade 8 Science. That’s the good news.

Kentucky did outscore 13 other states.

The rest tied us, as close as we can tell with any level of statistical confidence.

So, what does that say about how Kentucky’s poor whites (those who qualify for federal free and reduced cost lunch) rank for grade 8 science as of 2009?

We placed somewhere between fourth and 33rd in the nation. That’s all we can confidently conclude. Anyone who tries to make more precise claims is simply violating the statistical limitations of this sampled testing program.

No comments: