1. Sample size n=80. You would need a pretty incredible result for a sample size this small to be significant.
2. All students from the same region of France. Will this result carry over to other countries? Other cultures?
3. The treatment groups were chosen from pre-existing classes. Classes whose composition may not be random.
It's a cool result, and it should definitely be tried again elsewhere but pop science articles shouldn't be saying this is anything worth talking any yet.