FAIRFAX County, Virginia, where I live, is a large (160,000 students), high-achieving district. Within Fairfax County, a magnet school, the Thomas Jefferson High School for Science and Technology, educates about 1,600 of the brightest of the county's bright (and a few exceptional students from neighboring districts). At TJ, virtually everyone takes the SAT, and in 2001 the school's average was 725 on the verbal and 745 on the math. For Fairfax County as a whole, the numbers were 540 and 553 respectively; for the nation, 506 and 514. So TJ's students, who score collectively more than two standard deviations above the norm, should be set as as getting into the colleges and universities of their choice, right? Maybe not, according to Paul Attewell of the City University of New York, writing in the October 2001 issue of the Sociology of Education. Attewell points out that parents are concerned (some might say obsessed) with getting their children into prestigious colleges. They try to maximize their children's chances by moving into areas perceived to have superior by sending their children to prep schools, or by sending them to public such as Bronx Science, Boston Latin, Lowell High School in San Francisco, or the Thomas Jefferson High School for Science and Technology in Virginia. Sometimes the effort backfires. Attewell argues three points: 1. The reputation of a star school becomes tied to the attainments of its academically strongest students, but [that reputation] may be bought at the cost of reduced opportunities for the talented graduates who are below the top. 2. Attempts to game or manipulate the selective college entrance system have led to modifications in the grading systems that are used by many star high schools that further distance their top students from all the other students. 3. In some star this effort has led to changes in curriculum that lower the involvement of highly talented students in rigorous courses, especially in math and science and in Advanced Placement (AP) courses. Attewell finds that the SAT scores of students admitted to selective colleges from star schools are substantially higher than the average SAT scores for those colleges. However, students from star schools who are rejected by selective colleges had SAT scores that were far higher than the average of students who were admitted from other high schools. The principal culprit here is not the students' SAT scores or their grade-point averages (GPAs). It is class rank. I have pointed out repeatedly in connection with international comparisons that small differences in scores can yield large differences in ranks among countries. That is because many countries have similar scores. Forcing ranks onto a set of similar scores creates artificially large differences among nations. Similarly, at a school, ranking students can make a student with an SAT math score of 770 look much more able than a student who has an SAT math score of 750. In recent years, star schools have tried a variety of tactics to deny admissions officers access to class ranks. Some report only in terms of deciles. Colleges then get around this by assigning pseudoranks at the median of a decile. Thus a student from a class of 400 who is in the first decile is assigned a class rank of 20 (in a class of 400, the first 40 ranks are taken by the top decile). This puts anyone whose class rank is actually above 20 at a disadvantage. Some schools stopped reporting ranks altogether. Colleges then calculated pseudoranks using the distribution of GPAs that high schools report in the they send to college admissions offices. Some schools stopped sending fact sheets that included frequency distributions of grades, which led admissions officers to think the schools were hiding something. …