MORE NCLB MADNESS….Guess what? In today’s education news, we get yet another report that compares the results from state tests to the results from the “gold standard” NAEP test. This one is from the National Center for Education Statistics, and here’s the nickel summary: the standards used by states to measure compliance with NCLB are all over the map.
A picture is worth a thousand words, so take a look at the chart on the right. It shows how state standards compare to the NAEP “basic achievement level” for 32 states. This one is for fourth grade reading, and the state standards vary from 161 to 234.
Does this seem like a lot? Well, hold on: the rule of thumb for NAEP is that ten points is about equal to one grade level, which means that Mississippi at the bottom has a passing standard seven grade levels lower than the passing standard for Massachusetts at the top. Overall, more than half the states had passing standards a full grade level lower than the NAEP “basic” level.
So improved scores on state tests probably need to be taken with a grain of salt. Or a grain of statistics. Like this one:
There is also a negative correlation of -0.88 (with a standard error of 0.094) between the estimated NAEP score equivalents and the statewide percents proficient; that is, the larger the NAEP score equivalent, the lower the percent of students in a state deemed proficient.
Now, unlike yesterday’s poor correlation between states that showed improvement on their own test vs. states that showed improvement on the NAEP test (suggesting that the improvements we’re seeing in state test results might not be very meaningful), 0.88 is a very strong correlation. You won’t see one much better in the social sciences. And what that paragraph means in plain English is that the easier the state test, the more students who pass it. No surprise.
None of this means that testing is useless or that NCLB has nothing to recommend it. But it does mean that glowing reports about soaring test scores should be greeted very cautiously. Caveat emptor.
