The studies in this special section of the "Journal of Educational Psychology" present a variety of social-psychological interventions across large numbers of classrooms and populations. They show notable benefits for many students at risk for low performance. This is the glass half-full interpretation, and we consider the strengths of the articles from this vantage. When viewed collectively, however, the results also raise a number of basic questions, the most pressing of which is, why do the interventions only benefit a particular population, when there are reasons to believe most students might benefit? This is the glass half-empty interpretation. We consider the limitations of each study from this vantage. We argue that although these studies provide important contributions, in all cases better measures and theories are necessary. We identify two related questions for the field: Why do these interventions show selective effects and fail to generalize to all students? Second, why do these very different interventions, presumably involving different psychological mechanisms, consistently result in similar selective effects? To answer these questions, a grand challenge for the field is to develop better instrumentation that can capture relevant behaviors and attitudes over time, and how these vary across context and population. With improved instrumentation, efficacy studies could rest on a much stronger foundation that would yield more confidence as to the broad benefits and replicability of large-scale social-cognitive interventions.