Research in education is based and is validated by test groups involving students. However, teachers are not unified in their outcome. Some consistently produce higher average scores given uniformly allocated student capabilities classes and some reasonable time frame to gather data.
These teacher's methodologies bear a causal role to these results but these results they may not be explained by standard explanations i.e. cognitive or behavior focus or lesson plan delivery and quality.
These kind of research scenarios might also be use to downplay the actual contribution to learning of popular ideas like student engagement or the popular conceptions of engagement i.e. 1/3 of class time, 1/2 of student population participating or surface types of participation or teacher-led participation based on class momentum rather than authentic understanding. Cognitive models with little engagement might prove to contribute more than believed.