The rate of nonresponse of schools to the survey was quite high, as Bowles and Levin state. In part, the rate of nonresponse can be attributed to administrative difficulties created by extreme time pressure; in part it was due to a more serious problem. Of all Western countries I know, only in the United States is there such suspicion, bordering on hostility, of the federal government on the part of some local and state school systems, not only in the South but in the North and West as well. For example, the executive secretary of the American Association of School Administrators recently sent a memorandum to Association members, school superinten-dents, and principals throughout the country urging them not to participate in the program of national assessment that is presently being planned under the auspices of the Carnegie Corporation. As long as such attitudes exist, no one conducting serious, nationally sponsored research will be able to obtain a representative sample of schools. I do not suggest that the fault lies with those local and state school systems that refuse to cooperate with a federally sponsored survey any more than it lies with the federal government. I do suggest that something is wrong with the relations between the federal government and the states and localities in the area of education. But given this, and given the fact that educational policy decisions are now being made and will continue to be made, such decisions must necessarily be based on evidence from incomplete samples or on no evidence at all. Bowles and Levin next point out that the rate of nonresponse of students to certain questionnaire items, especially parents' education, was quite high. They neglect, however, to mention other points which place the nonresponse to this item in quite a different perspective. The procedures used to insure that family background was measured adequately included the following:(1) Having the teacher query the child and complete the questionnaire herself at grade 1, and having the teacher read each question aloud at grade 3.(2) Using a number of questions which the child can answer about current conditions in the home, such as the number of brothers and sisters he has, the existence of selected items in the home related to economic status (eg, telephone, refrigerator, vacuum cleaner), the existence of specific reading material in the home (eg, a daily newspaper, magazines, books, dictionary, encyclopedia).(3) Combining items so as to form eight measures or indices-in particular," urbanism of background,"" structural integrity of home,"" size of family,"" economic items in home,"" reading material in home,"" parents' education,"" parents' interest,"" parents' desires for child's education."(4) Never using any of the indices alone to measure the effect of family background, but using either six or eight taken together. The unique contribution of any one, such as parents' education, to variance explained in achievement is small. Table 3.221. 6 in the Report shows that parents' education accounts uniquely for less than 1 percent of the variance in achievement on the average at grade