The easiest way for a statistician to have minimal impact on improving research practices is to suggest solutions that require other scientists to think and care about statistics as much as they do.
Funny, such a common failure mode across humanity. I've sat with groups of X discussing "How can we get people to care more about X, it's a serious issue that people don't know enough about X", for X=lawyer, artist, scientist, entrepreneur, etc. Near-isomorphic conversations.
-
-
Yes, exactly. A very interesting conversation emerges with a reset of assumptions -- Assume people are never going to care about X at all, now what are you going to do to advance X? Instead of changing hearts & minds, solutions go after structures and incentives. Much better.
-
Yes! In the Netherlands the gov once tried to persuade ppl to separate their bio-waste by having guilt-inducing TV ads. Didn't work at all. Then they started charging $ for general waste. Worked immediately.
End of conversation
New conversation -
-
-
Answer, also care about Y and appeal to a symmetry argument
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I think you have to abstract it away. For example, as much flak as the p-value gets - it still was a standardized thing where basically any scientist could run it without understanding it, and get a good idea if their experiment worked or not.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Well-observed! It is human to be egocentric, thinking «I’m good at X -> X is important and undervalued -> everyone should care more about X»
-
So I agree it’s more likely to achieve behavior change w structural interventions + when a new X is helpful for pre-existing/intrinsic goals
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.