Does study group size affect test performance? From a large class of students, 12 are selected to study?

12 is nowhere near enough students to get a statistically significant answer, and you don't seem to have controlled for some people are smarter than others.