Monday, July 01, 2013

NAEP Long-term Data Confirms, American Schools Have Never Been Better

This from Kati Haycock in the Huffington Post:

New Data, Real Progress

Most teachers and principals don't have a lot of time to pour over the latest data from the National Assessment of Educational Progress long-term trend assessment. That's too bad, because the results from this "gold standard" national examination provide overwhelming evidence that the blood, sweat, and tears that many have poured into helping their children achieve at higher levels has paid off spectacularly for their students. Despite the constant cry that American schools are in decline, these results suggest they have never been better. At the elementary and middle school levels, overall results are up substantially, and at all ages, they are up for every group of children: white, black, and Latino.

The long-term trend assessment has been given since the 1970s. Despite the fast-changing demography of our country, though, reading and math performance for 9 and 13-year-olds has increased significantly. At all ages, gains have been largest among students of color.

And make no mistake: The gains are meaningful.
  • In math, for example, African American and Latino 9-year-olds are performing today about where their 13-year-old counterparts were in the early 1970s.
  • Moreover, while it might have seemed impossible 25 years ago for black 9-year-olds and Latino 9 and 13-year-olds to reach the proficiency levels that white students then held, they have indeed reached those levels in math.
As a result, even though the performance of white students has improved, too, we've made significant reductions in long-standing gaps between groups. The reading gap between African American and white 9-year-olds has nearly been cut in half. In math, the gap between Latino and white 13-year-olds has narrowed by 40 percent. And among 17-year-olds, the black-white and Latino-white gaps in reading and math have narrowed by about half.

Moreover, there's been progress across the achievement spectrum for younger students, from those at the low end of the performance distribution to those at the high end. In math, for example, the lowest performing 13-year-olds in 2012 scored 27 points higher than did the lowest performing 13-year-olds in 1978. And the highest performers in 2012 scored 16 points above the highest performers at the beginning of the trend. Though the gains in reading weren't quite as large, both low and high-end 9- and 13-year-olds made strides, there, as well.

If we have a crisis in American education, then, it isn't that our schools are getting worse. The crisis, quite simply, is this: We aren't yet moving fast enough to educate the "minorities" who will soon make up a "new majority" of our children nearly as well as we educate the old majority. At best, students of color are just now performing at the level of white students a generation ago in math. And in reading, they're still performing worse than their white counterparts did in any year.

By now, of course, we educators know those patterns aren't inevitable. All around the country there are schools -- including both regular public schools and public charter schools -- that are producing extraordinary levels of achievement among low-income students and students of color. Our challenge is to take what we can learn from them -- and from a careful analysis of the data -- and put it to work far more broadly.

But what, in fact, can these particular long-term trend data tell us? It's a fool's errand to make claims about causality, so I won't do that here. But as we ask the question, "How can we accelerate progress for the very students who will soon be our new majority?" it is instructive to look at rates of progress over time.

It's clear, for example, that some of the biggest gains for black and Latino students took place in the 1970s and early 1980s, a time when policymakers were beginning to confront problems inside of the educational system -- including segregation and deep funding inequities -- and problems outside of it.
Results since the late 1990s -- those that coincide with efforts to raise achievement and close gaps through standards, accountability, and public reporting -- show gains as well; but, like the gains in the late 1970s, they also suggest a worrisome slowing down in the most recent years. Take African American 9-year-olds, for example. From 1994 to 1999, math scores actually declined by one point, or by about 0.2 points a year. But between 1999 and 2004, just as accountability and public reporting efforts took hold nationwide, scores increased by 13 points, or roughly 2.6 points per year. Since then, however, the rate of improvement has slowed. This pattern of steep gains between 1999 and 2004, followed by slower rates of improvement, is consistent across all groups of 9-year-olds.

And there is yet another message clear in the data: We are not getting the job done in our high schools nearly as well as we are at earlier grades. Our students are entering high school better prepared, but along the way that advantage is not being translated into proportionate gains among 17-year-olds overall.

What lessons should we take from all of this?

First, improvement and gap closing is not just a theoretical possibility, it is happening: Longstanding gaps between groups are getting ever smaller, though not nearly fast enough for either the kids or for our collective future.

Second, if we are going to speed up the rate of improvement, we need a full-court press to improve opportunity and outcomes for students of color--one that confronts rising economic inequality and isolation outside of schools, as well as continued unequal opportunities within schools. And we need to stop pretending that, if we get things right in the early grades, our high school problems will take care of themselves: High schools need attention and resources, too.

Third, when it comes to policies governing our schools, standards and accountability spark improvement, but alone are insufficient to generate the kind of sustained gains that we need.

Accountability has made a difference, and, especially as Congress works to reauthorize NCLB, we must hold the line on high expectations for all groups of students. But we can't stop there. Meaningful accountability must be coupled with powerful standards; rich curriculum; effective, well-supported teachers and leaders; and extra supports for out-of-school challenges.

In the meantime, though, we ought to stop and say thanks to the teachers and principals whose work produced all of this progress, and whose continued efforts will be essential to taking us the remaining distance.


Anonymous said...

So how is it that we are actually demonstrating long term increases in acheivement, including gap reduction, but our leaders and media seem to think we are behind the curve?

High Schools are simply going to have to be retooled to be something completely different from the current seat time, pre-determined credit aquisition model which mimics the past. One thing that folks are going to have to let go of is the romanticized and over emphasized place which athletics plays in high schools - we simply can't afford them at the level they currently function. I don't want to beat up on sports but KHSAA keeps lumping more sports onto the possible list of activities which parents can petition to have - bass fishing, bowling, etc. At the same time schools have less and less money to spend on these activities and staff who are already overburdened with growing daily instructional expectations. I say let the AAU/traveling/club teams have it all. I spend more and more time dealing with athletic manusha, not to mention the countless hours of supervision as a school principal and I have an athletic director. I know this wont' be popular in small counties which have centralized their community identity and social behaviors around HS athletice for decades but I am tired of trying to find money for equipment, salary for new sports, increasing officials expenses and transportation. How can we expect our students take a broader and more involved view of their education when we spend so much on time and money on something which frankly doesn't directly contribute to student academic growth?

Skip Kifer said...

John Tukey, esteemed statistician and member of the original NAEP technical committee, insisted that any statistic presented should have along side it a defensible standard error. Else, how is one to know whether a 27 point difference is large or small. On the SATs, for example, a standard error (there could be others) is about 30 points, so 27 points is no big deal.

This Haycock piece is a marvelous example of the worst kind of cherry picking of results, combined with their indiscriminate presentation. How about this:

"In math, for example, the lowest performing 13-year-olds in 2012 scored 27 points higher than did the lowest performing 13-year-olds in 1978. And the highest performers in 2012 scored 16 points above the highest performers at the beginning of the trend."

I am not sure how she defines lowest scoring and highest scoring (and apparently she is not going to tell us), but NAEP has notoriously bad measurement of the extremes of the proficiency distribution, especially on the low end. That 27 points could be just fluff.

There are some noticeable differences in the profiles of performance, for example, in mathematics. Both samples, age 9 and 13, score much, much higher on items measuring data analysis and statistics. On other subscales, there has been little improvement. A trend analysis substantiates both of the above statements.

But, her's is a blog. No need to worry about defending idiosyncratic interpretations of results.