Friday, October 12, 2007

The Tyranny of the Norm (or, it’s normal to be different)

According to the Boston Globe (“Student takes his C to federal court,” October 4, 2007), a student at the University of Massachusetts-Amherst has gone to court to challenge a grade in one of his courses. The student determined he had done “A” work in the course but, because the final grade was computed “on a curve,” he got a “C” for his final grade. Lots of high school and college teachers grade “on a curve.” It’s more or less assumed that the bell curve provides a natural description of most human behaviors, including academic performance. The idea that, in general, human behavior distributes “normally” is a powerful idea that stands behind whole-class instruction (most kids cluster around the average), special education (students who depart significantly from the norm require a “special” education), educational testing (students are often compared to the “average” student at their grade level), educational research (statistical comparisons between groups or instructional interventions are based on the mean), and, sometimes, grading practices.

One of my doctoral students and I have been researching the history of the normal curve. Almost everyone assumes that the normal curve is an accurate model for representing variation in human behavior, but does the evidence actually support this common-sense assumption? As it turns out, some physical characteristics like height do distribute along a bell-shaped, normal curve. The vast majority of human behaviors do not, however, distribute normally. Weight doesn’t. Running speed and reaction time don’t. It’s not clear that intelligence or academic achievement distribute normally, either. Standardized tests are designed to produce normal distributions so the degree to which traits like intelligence and academic achievement distribute normally says more about the skill of the people who construct tests than it does about the human condition (and, even then, actual test scores often do not produce normal distributions). So, as it turns out, normal distributions are not the norm and this has been apparent to a few statisticians and social science researchers for over 100 years.

If the normal curve is a myth, as we believe that it is, then beliefs and practices based on the assumption of normality must be challenged. One particularly troubling practice that emerges from the “myth of normality” is the use of means (or averages) to represent groups of people, especially school children. The problem of using averages to represent the performance of children in our schools is that averages obscure the natural variation that characterizes the behavior of human beings. The claim that boys do less well in school than girls does not consider the fact that many boys do very well in school (better than most girls) and many girls do poorly in school (worse than most boys). Similarly, research on fourth grade readers or students with learning disabilities, for instance, obscures the individual needs of real children who are not statistical averages.

The assumption of normality has led to a range of educational practices and research that ignore the variety of ways children learn and the wide range of experiences they have in and out of school. We need to replace the notion that it’s normal to be average with the idea that it’s normal to be different. This will lead us to shift the focus of research and instruction from the mythical average student to the needs of individual learners.

Friday, October 5, 2007

Remedying the Achievement Gap

The recently released results of the National Assessment of Educational Progress (NAEP) reveal a persistent achievement gap between Black and Hispanic students and their White counterparts. The data also indicate that the achievement gap has held fairly steady since the results of the first NAEP were released 15 years ago.

Why has the achievement gap proven to be such an intractable problem? Presumably, No Child Left Behind (NCLB) was drafted specifically to ameliorate the chronic underachievement of poor Black and Hispanic children, especially in reading.

My sense is that NCLB has promoted teaching practices that largely sustain the achievement gap. NCLB and related programs like Reading First promote circumscribed, skills-focused reading curricula that deny students in under-performing schools the rich reading experiences routinely provided to students in more affluent, higher-performing schools. It’s no surprise that students in poor urban and rural schools are plagued by a “fourth-grade reading slump.” An obsession with discrete skill instruction to the near exclusion of reading connected text virtually insures that many underachieving students will struggle when the expectation changes from learning to read to reading to learn.

I don’t understand why many policy makers think that students in underachieving schools – places overpopulated by poor Black and Hispanic children – require qualitatively different reading curricula from students in high performing schools. If students in high performing schools have lots of opportunities to read and discuss engaging texts – and they do – then this is what students in low performing schools need, too. If high achieving students have time for sustained engagement with various kinds of texts through practices like sustained silent reading – and they do – then this at least as important for low achieving students.

The antidote to the achievement gap then is high expectation curricula informed by the practices and opportunities common in high achieving schools. The evidence that the rich and varied curricula found in high achieving schools “work” is obvious: the students in these schools do very well academically. The evidence that low expectation curricula common in underachieving schools do not is equally obvious.