NAEP changed the way pupils with disabilities were considered
Dick Allington
University of Florida

Much is being made of the rising reading achievement scores on state tests and the National Assessment of Education Progress in the media this past week.

But the NAEP changed the way pupils with disabilities performances were considered (allowing test modifications) and so that 10-15% of the lowest-achieving students looks like they read better but the improvement may be more a function of the changes than improved reading performances. On state tests big improvements in many states have been noted, and have been typically linked to the No Child Left Behind Act. But the NCLB funding and regulations have barely hit the streets at this point.

A better explanation may be what seems to be a massive increases in flunking this year, especially in grades K-1-2. Just five years ago Boston, for instance, failed only 2% of the 3rd grade students but most recently failed over 10%. All this added flunking would likely account for much of the reported increases as lower scoring ids are held back and don't take the grade level tests on schedule (taking the 3rd grade test after 4 years of schooling, K-1-2-3). Any kid who looks like he might not pass the 3rd grade test doesn't get to 3rd grade anymore. By holding lower achieving kids once or twice the school gets additional time to teach them to read, so they should score better.

But that doesn't mean that schools are more effective or efficient. It just means that last year that 8 year old poor reader was in 3rd grade but this year boys like him were still in 2nd grade and their scores weren't included in the 3rd grade scores. So kids who get to 3rd grade might be reading better but taxpayers may have paid for one or more year's additional education (at an average cost of $5000+ per kid) to get that reported improvement. The question is whether we might have spent less than that $5000+ to get the same result? Or put another way, just how much is all this flunking really costing us and is flunking the most cost-effective way to push reported reading scores up?

Note, I've not even touched on the ever-rising numbers of kids identified as pupils with disabilities. Reported scores also rise because these kids results are typically excluded from state reports and state grading of schools (and in some states it is an allowable "modification" to have someone read the reading test to pupils with disabilities!).