Conversation

Cedric Chin
@ejames_c
One test I've been toying with for 'smart person attempts to use their considerable intelligence to wave away inconvenient data' is "how many of these different reasons do they use when pushed towards a conclusion they don't want to accept?"
Quote Tweet
Cedric Chin
@ejames_c
3/ It turns out there are basically only 7 ways you can respond to inconvenient data. 6 of them allow you to preserve your existing mental models. See if any of these are familiar to you, before we go through them in order:
Show this thread
Image
2
18
Cedric Chin
@ejames_c
The test applies to yourself, of course. If you catch yourself changing reasons more than 3 times, perhaps you might want to accept the conclusion? Instead of using your considerable intelligence to dig in?
1
3
Cedric Chin
@ejames_c
What isn't captured by the test: the scenario where you accept the data, change your conclusion, and then fail to take action on your updated conclusion, because ... reasons. (Which also regularly happens to reasonably smart people!)
1
2
Cedric Chin
@ejames_c
Perhaps one way to make this test more actionable/acceptable is to devil's advocate the acceptance. That is, don't accept the undesirable conclusion outright, but go "suppose this is true, what might this mean I have to do ...?"
1
3
Cedric Chin
@ejames_c
At which point, if every fibre of your being screams to look away, you know you've probably identified something you should take a closer look at.
1
5
David Golden
@xdg
Replying to
@ejames_c
I like the "suppose this were true" recommendation. What I find disquieting about the overall test is the assumed validity and weight of new, "inconvenient" data. No's 2-6 all seem reasonable enough when the validity of new data is unproven. Not sure how to square that circle.
1
1
Cedric Chin
@ejames_c
Replying to
@xdg
Yeah one of the points that Chinn & Brewer make in the original paper is that most new data is wrong, or is noise, and so it’s perfectly reasonable to wave it away. But if you’ve changed reasons to ignore it 3 times … Let’s see if this works.
1