Externally set curriculum-based examinations at the end of high school apparently have pervasive backwash effects on middle school students, their parents, teachers and school administrators. Holding the social class background of students constant, students from Canadian provinces with examination systems were substantially (23 percent of a standard deviation) better prepared in mathematics and 18 percent of a standard deviation better prepared in science than students from provinces lacking such exams. The effect of an exam system on mathematics achievement of 13 year olds is larger in a standard deviation metric than the decline in math SAT scores between 1969 and 1980 that has been such a focus of public concern. Other natural experiments yield similar findings. When adjustments are made for ethnicity, gender and social class of SAT test takers, New York State ranks higher on the SAT than any of the other 38 states where the test is taken by large numbers of students. The mathematics and science achievement of Swedish high school seniors declined in the years following the elimination of high/medium stakes curriculum-based exams. The analysis also found that examination systems had pervasive effects on school administrators, teachers and parents. In the provinces with external exams, schools were more likely to: -- employ specialist teachers of mathematics and science -- employ teachers who had studied the subject in college, -- have high quality science laboratories -- schedule extra hours of math and science instruction -- assign more homework in math, in science and in other subjects -- have students do or watch experiments in science class and -- schedule frequent tests in math and science class. At home students watch less TV, spend more time reading for fun, and are more likely to report their parents want them to do well in math and science. In addition, parents are more likely to talk to their child about what they are learning at school.