Growth mindset interventions: why the gap between meta-analyses and actual replications

There's something almost poignant about watching a field discover its own methodological blind spot in real time. I've spent enough years reading education research to know the pattern: meta-analysis shows moderate effect sizes, pre-registered replication with proper controls finds d = 0.08, everyone acts surprised, then we move on to the next thing. The honest answer is probably boring. Publication bias + file-drawer problem + the fact that older studies often had looser designs. Meta-analyses aggregate whatever got published, which skews upward. Newer, pre-registered replications by people like Sisk and Dweck's own lab can't just stop the study when results look promising. They have to finish it. That's not a failure of mindset research specifically—it's the entire edifice of how we've been doing social science, and growth mindset just happened to be famous enough that someone actually bothered to check. What's interesting to me is what happens next. We could say "effect is real but small, still worth doing." Or "effect is mostly context-dependent, need to figure out which contexts." Instead I see people either retreating to "mindset interventions don't work, period" or quietly using old studies in their lit reviews and moving forward. Neither is exactly wrong. I'm just struck by how rarely anyone sits with the uncomfortable middle position: it might work for some kids in some conditions, but we have no honest way to know which yet, and the intervention as a scalable one-size-fits-all is probably oversold. That's not a publishable finding, though.