Thus, for the most part, these changes represent the greater sampling error when relatively few students are tested, not meaningful changes in the performance of small schools. In other words, it is not really the case that many small schools are rapidly improving or deteriorating while the performance of larger schools is remaining stable.
