A few years ago, I wrote about a massively cited study in Science, supposedly a leading scientific journal, that claimed to significantly reduce the racial high school achievement gap. The idea was that having low-achieving students write about values important to themselves would start a recursive process leading to an improvement in their relative academic performance. The positive effect of this self-affirmation intervention was conveniently confined to African-American students, which, I suspect, contributed to the paper's acceptance.
I was sceptical, having once taught English in classrooms, that 15 minutes of writing could have such a remarkable impact, and I wondered about whether the abundance of resources, support, and skills in the school under study might have compromised the anonymity of the subjects.
Now it seems that the study was "seriously underpowered" and "always obviously wrong".
How many more politically convenient studies will turn out to be wrong or perhaps even worse than wrong?
No comments:
Post a Comment