Search Google

5/28/15

Shifting ground below paradoxical NAEP scores

A few weeks ago, the National Assessment of Educational Progress (NAEP) released the 2014 national results of eighth-graders’ tests in Civics, Geography, and US History. In absolute terms, these test results are not encouraging. However, if you dig a little deeper, and look at progress over time by student race and poverty, as Chad Adelman, an associate partner at Bellwether Education Partners, illustrated, you’ll find some arguably good news. If you dig even deeper, to find the rapid and huge changes in the racial and poverty composition of US students, you may see the ground shifting.

Overall, NAEP results show that eighth-graders did not excel in social science in 2014. In Geography, about 27% of eighth-graders scored proficient or above (think grades of A, B, and C) which was higher than the 25% scoring “below basic” (think F); this is not encouraging. In US History and Civics, more students scored below basic than scored proficient or above; this is downright discouraging.

Any good news? Perhaps certain subgroups of students are improving. Take NAEP Geography as an example. Since 2001, the overall geography test scores have gone up 1 point (not a significant change); during the same period, white, black, Hispanic, and Asian students gained 4, 7, 9, and 9 points, respectively. This pattern of slight overall increases alongside larger increases for all subgroups can be found over 10-12 years in the US History and Civics results, as well as prior NAEP subject tests.*

But wait, how can that be? All subgroups are improving while overall growth is flat?

As Adelman points out, the key that unlocks this puzzle is called Simpson’s Paradox, and it happens when the composition of the total changes over time. Consider a football analogy. Suppose in 2014, players are stronger than ever, and weigh 10 pounds more than those playing the same positions in 2001. But the overall average players’ weight was the same in 2014 and 2001. The key difference over time is that in 2014 teams had more receivers (think lighter and faster players) and fewer lineman (think big and heavy) than in 2001. By 2014, players had gotten bigger, but teams had fewer “big men” and more “speedsters.”

Like the mix of players in the football example, the mix of US students by race and poverty has changed dramatically since 2001, and the importance of these shifts is hard to overstate (cue ground shifting). In 2014, 50% of eighth-grade students were white, down from 69% in 2001. During the same period, the proportion of Hispanic students grew 230%, from 11 to 26% of eighth-graders. Since Hispanic students made up a larger portion of all students in 2014 than in 2001, and scored about 25 points below their white peers, the overall scores look flat. Student poverty also radically changed, with about 1 in 3 students receiving free lunch in 2001 compared to 1 in 2 in 2014, with an associated test score gap of 26 points.

There are three points to consider in light of these NAEP results:

First, overall, these NAEP results are pretty dismal. Students, and the parents, teachers, and policymakers that support them, have much work to do to improve achievement in these and other subjects.

Second, interpreting complex education data demands nuance. Whether these and other NAEP results should be described as flat, or whether they can be called encouraging, or not, is a discussion worth having. (For example, don’t miss the comments on Adelman’s post by Jay Greene from the University of Arkansas.) Improving scores across subgroups is certainly encouraging, but whether or not such improvement is satisfactory is an open question. Debate aside, gauging how productive US public schools have been requires much more than a glance at the headlines.

Finally, don’t miss the big story here. US students in 2014 have a very different racial/ethnic and socioeconomic makeup than they did in 2001. These are big changes that happened very quickly. Such seismic shifts in the student population should play a huge role not only in interpreting NAEP and other standardized test data, but in other big-ticket issues like Elementary and Secondary Education Act reform, early childhood education policy, and how billions of Title I dollars are distributed. We ignore the shifting ground at our peril.

* You can explore the NAEP results yourself using the NAEP Data Explorer. For the whole picture ask for “Average scale scores” and “Percentages” under “Statistic options.”



from AEI » Latest Content http://ift.tt/1PQhxqS

0 التعليقات:

Post a Comment

Search Google

Blog Archive