I appreciate this perspective on theory evolution - but most importantly, the reason why - a strict focus on the “best practices” that then get applied (or mandated) for broad segments of students is often a misapplication that doesn’t require a rejection, but a revision - more attempts - and not strict replications, but variations based on the same theoretical principle.

The incredible number of variables in education are what make hard science in education so hard, and why many studies who attempt to include those many variables end up rejected for sloppiness. In order to produce the “robust” and replicable research in education, we have to control for (or ignore) so many applicable variables to isolate the impact of a certain instructional approach. Then, when a replicable finding gets rolled out and applied as a strategy for learning, there are inconsistencies with the original “data” as the replication attempts encounter the variables that has been controlled or unaccounted for.

But if we don’t judiciously attempt instructional moves based on promising research that DO produce results in certain contexts under certain conditions, we risk missing a better revision that may account for that exposed variance or gap.

Totally agree, failure is the way we keep improving teaching and learning, but also helping school-level instructional folks see themselves as researchers of their practice, too - providing a feedback loop to researchers interested in developing theories that are flexible enough to be used to make better learning happen. A teacher-as-researcher stance also threatens systems of compliance, so the relationship between those developing theories that add value and those testing them out in the wild (if they are even able to) is too far apart for better data to inform either.