A new reading program is being tested. Parents are asked whether they would like to enroll their children, and 50 children are enrolled in the program. There are 45 children whose parents do not choose to enroll their children. At the end of the school year, the children are tested. Of the 50 children who participated in the program, 38 are found to be reading at grade level. Of the 45 children who did not participate, 24 were reading at grade level. Explain why these data should not be used to test the hypothesis that the proportion of children reading at grade level is higher for those who participate in the program.