In June 1993 the "Wall Street Journal" carried a table of data prepared by the Heritage Foundation that listed the states in order of the average amount they expend on each public school student. The table also contained each state's rank on the average score on the Scholastic Aptitude Test (SAT) and the rank of each state in the average performance of a random sample of eighth graders on the mathematics portion of the National Assessment of Educational Progress (NAEP). The obvious inference drawn from this table was that spending money on education does not improve student performance, since some states that spent a great deal did not have high performance indices. Any such broad inference is almost surely specious. Setting aside such considerations as the differing cost of living, one issue is examined: self-selection. Average performance on the SAT does not represent study proficiency in that state, since only a small percentage of students take the SAT in some states. The uncertainty within a state about who takes the SAT is too great an obstacle for statistical adjustment to overcome. NAEP scores, which are closer to being the sort of random sample that is required for meaningful comparisons, are a more reasonable measure, and when they are examined with expenditures, there is a small relationship that indicates that spending money does improve performance. For every thousand dollars spent a state's NAEP ranking improves by two places. The reproduced table and two figures illustrate the discussion. (SLD)