From the start, however, some doubts must have risen about the crisis rhetoric, because in 1990, Admiral James Watkins, the secretary of energy (yes, energy), commissioned the Sandia Laboratories in New Mexico to document the decline [of SAT scores] with some actual data.More here and here.
Systems scientists there produced a study consisting almost entirely of charts, tables, and graphs, plus brief analyses of what the numbers signified, which amounted to a major "Oops!" As their puzzled preface put it, "To our surprise, on nearly every measure, we found steady or slightly improving trends."
One section, for example, analyzed SAT scores between the late 1970s and 1990, a period when those scores slipped markedly. ("A Nation at Risk" spotlighted the decline of scores from 1963 to 1980 as dead-bang evidence of failing schools.) The Sandia report, however, broke the scores down by various subgroups, and something astonishing emerged. Nearly every subgroup -- ethnic minorities, rich kids, poor kids, middle class kids, top students, average students, low-ranked students -- held steady or improved during those years. Yet overall scores dropped. How could that be?
Simple -- statisticians call it Simpson's paradox: The average can change in one direction while all the subgroups change in the opposite direction if proportions among the subgroups are changing. Early in the period studied, only top students took the test. But during those twenty years, the pool of test takers expanded to include many lower-ranked students. Because the proportion of top students to all students was shrinking, the scores inevitably dropped. That decline signified not failure but rather progress toward what had been a national goal: extending educational opportunities to a broader range of the population.
By then, however, catastrophically failing schools had become a political necessity. George H.W. Bush campaigned to replace Reagan as president on a promise to confront the crisis. He had just called an education summit to tackle it, so there simply had to be a crisis.
The government never released the Sandia report. It went into peer review and there died a quiet death. Hardly anyone else knew it even existed until, in 1993, the Journal of Educational Research, read by only a small group of specialists, printed the report.[emphasis all mine]