What do other
researchers have to say about the recent trends in learning and schooling?
The decline (and rise) in
student test scores over the last twenty-five years is one of the most
researched and least understood phenomena in education. As yet, researchers have
produced no simple or adequate explanation for the initiallytroubling, then temporarily encouraging,
trends in test scores. The trends appear to be the product of many factors, some
educational but many non-educational. The most important factor, accounting for
perhaps a fifth of the total decline, appears to be a change in the ethnic
composition of the test-taking population.
 American schools were taking in
different kinds of students, students who were more difficult to educate than
students in the past. Influences in the home were also changing. The second most
important cause of the decline and upturn appears to be changes in family size,
with larger families initially hampering achievement and then smaller families
also clear that the decline did not affect all grades equally. The decline was
comprised primarily of worsening scores among students born before 1963, the
"baby boom" generation.
 As these students moved through the schools, test
scores declined, pushing SAT scores down from 1964 to 1979. But as the baby
boomers began to be replaced, around 1970, by a new cohort, the "baby bust"
generation, test scores in the early grades began to climb. By 1980 the younger
cohort, now in high school, was taking the SAT tests, and posting the modest
increases in SAT scores observed during the early eighties. Unfortunately,
further gains have not been posted by subsequent cohorts, leaving achievement
generally below the levels of twenty-five years ago.
significant contribution of so-called compositional and cohort effects to
changes in test scores highlights the importance of factors beyond the control
of schools in producing student achievement. Yet, even when the full range of
non-educational factors is taken into account – alcohol and drug use, and
exposure to environmental lead (both of which had small effects on test scores);
single-parent households, maternal employment, and television viewing (none of
which had any effect on test scores) – no more than a third of the variation in
test scores over time can be explained.
 That leaves a lot of room for
educational factors to make a difference.
researchers have made little progress in identifying significant educational
factors. The most comprehensive study to date, by the Congressional Budget
Office in 1987, found some evidence that schools might have undermined
achievement by watering down the content of courses, assigning less homework,
and using less challenging textbooks.
 But the study found no impact,
positive or negative, from other educational factors such as teachers' test
scores, teachers' educational attainment, or state graduation requirements. The
fact of the matter is, most of the relationship between schooling and learning
over the last twenty-five years remains a mystery.
clues about the relationship can be found, however, in other kinds of research
into student achievement, research that has not focused on test score trends but
on differences in tests among schools at any given time. This research has
reached some fairly strong, though negative, conclusions about the connection
between schooling and learning. This research implies that there is no surprise
in the fact that test scores declined or stagnated while school resources and
conditions improved. A recent survey of 147 statistical analyses of school
performance, for example, found no consistently positive and significant
relationship between student achievement and any of the major factors popularly
assumed to influence achievement: teacher-pupil ratios, teacher education,
teacher experience, teacher salaries, and per-pupil expenditures.
In other words, much of what school systems were doing to turn test
scores around may have no systematic effect on school performance.
Nevertheless, we know
that factors outside of schools do not adequately account for student
achievement either. And we know from casual observation, as well as careful case
studies, that some schools are much, much better than others. The challenge
remains to find out why. The research in which we have been engaged takes up