A school district is comparing two reading programs for high school students reading below grade level. Two very similar schools are used to test the programs, a different one at each school. Students participating in the programs are pretested for their reading abilities and the mean scores are similar at both schools. The reading program implemented at High School A is known to be much more rigorous than the program implemented at High School B. The posttest shows more improvement at High School A, but more of the students dropped out of the program. What threat to internal validity does this pose?