We will summarize our conclusions about their three claims. Do we unfairly expect that remediation should raise outcomes of remedial students above those of college-level students? The conclusions from the regression discontinuity studies suggesting that remediation is not effective because it fails to raise the outcomes of developmental students above those of similar college-ready students may indeed sound puzzling. Since one tends to think that developmental and college-ready students are two distinct categories, it is often difficult to understand a methodology based on the idea that some students in these two apparently distinct groups are in fact very much the same (at the time of assessment). It is this pool of virtually identical students on each side of the cutoff that are examined in the regression discontinuity studies, and for them it is appropriate that we expect remediation to raise outcomes for "developmental" students above outcomes for the "college-ready" students. The large majority of the outcomes in these studies find no improvement or even negative results, although there are a small number of positive results. Focusing only on success in the first college-level course does not yield a more positive general result. Do we ignore studies with positive outcomes? The most positive findings that Goudas and Boylan (2012) cite are based on comparisons of developmental education completers and college-ready students. Focusing on the completers makes intuitive sense: It is worthwhile knowing the effect of a service for those who experience it, and indeed we frequently hear college personnel state that their developmental education completers do as well as their college-ready students. But unfortunately, this comparison cannot differentiate between the positive academic effects of remediation, which surely some students experience, and the screening effect, which eliminates from the comparison many of the weaker students who fail to complete remediation.