From the article:
The first is that the PSL schools play by different rules. There is a cap of 65 on most of their class sizes, for example, which has prompted allegations that some operators are turfing out less clever pupils. That would be unfair and against the rules of the pilot. But even if it were happening, it would not alter the results of the evaluation. Justin Sandefur of the Centre for Global Development, the research group leading the trial, notes that operators will be held accountable for the results of all children originally at the pilot schools—including any who were later turned away.
Read the full article here.