Nature Human Behavior just published an RCT in Germany of a program that teaches self-regulation skills in 1st grade (“Mental Contrasting with Implementation Intentions”). Quick take: The RCT had a key flaw, but found large effects on reading achievement — so findings are still somewhat promising.

Program & Study Design:

• The study randomly assigned 31 first-grade classrooms, with 599 students, to treatment (five weeks of lessons in self-regulation skills focused on practicing reading and monitoring one’s mistakes) vs control (regular classroom teaching).

Findings:

• 1 year after the program, the study found large effects on students' reading scores and teacher-rated reading ability (effect sizes ~0.4). 3 years later, 76% of treatment students were on advanced academic track vs 63% of controls. These effects were statistically significant (p<0.01).

Comment:

• The RCT was well-conducted in many respects (e.g., low to moderate sample attrition, analyses accounted for the fact that classrooms rather than students were randomly assigned, testers were blind as to who was in T vs C group).

• But the RCT also had weaknesses, including: At baseline, T students scored much higher than Cs in self-regulation (effect size 0.37) & had a 15% point lower rate of language problems. The main analyses didn't control for these differences so may have overstated the impacts.

• On balance, the effects are large enough that I think the findings are moderately promising despite the study limitations.

• I think this would be a good candidate for a replication RCT - to see if the sizable effects can be reproduced in a second study in a different set of schools.