arithmetic test for each grade from 4 to 8, with some overlapping of items across grades (sample item at grades 6, 8 "A gentleman gave away 1/7 of the books in his library, lent 1/6 of the remainder, and sold 1/5 of what was left He then had 420 books remaining. How many had he at first?"). Rice supervised the administration of the tests to about 6,000 students from 18 schools in seven cities. His most striking finding was the large differences in achievement between cities. although there were some large differences between schools within cities as well. The average scores for the schools ranged from 80%, and the variation was even greater within some grades Rice could not account for the differences between schools in terms of class size, average age of the students, or socioeconomic level of the neighborhood or city Some schools in the slums had higher average scores than the aristocratic" schools. Rice scored not only the number of problems the students got correct but also the number that would have been correct if a mechanical error had not been made. He found relatively few mechanical errors. The schools that had the highest number of problems correct in principle had the fewest mechanical errors, which prompted him to conclude that "the mechanical side of arithmetic has shown itself to be very closely related to the thought side" (pp. 291-292)