[unishark]

They reported two facts, statistical significance and effect size, one of which was not replicated. Of course this doesn't prove anything definitively, but it still arouses suspicion. As for whether it's exclamation-point-worthy, at worst it depends. For example, 0.5 standard deviation improvement in math scores might normally require a ton more expensive effort or even have been thought previously impossible, while a 0.2 was easy and not publication-worthy.