A big thanks to David McRaney of You Are Not So Smart for namechecking me and TOP in his TEDx Talk (at around the 9:30 mark). There goes another minute of my allotted 15 minutes of fame, for sure.
The talk is about Survivor Bias and is excellent, especially if you're not already familiar with the idea. His story about the bomber and the statistician is a great one.
I always think of Survivor Bias when they interview centenarians for the secret of living so long and they say things like "you should smoke a cigar every day" or "I drink one shot of bourbon first thing every morning." We learn a lot more by studying what killed all the people who don't live to 100.
David did a nice job with the talk, too. I wish I could do that. I'm not good at public speaking.
(Thanks to Doug Nelson)
Original contents copyright 2015 by Michael C. Johnston and/or the bylined author. All Rights Reserved. Links in this post may be to our affiliates; sales through affiliate links may benefit this site.
(To see all the comments, click on the "Comments" link below.)
Featured Comments from:
Kenneth Tanaka: "'Survivorship bias' is also a hazard in financial market analyses. Someone back-testing an equity investment model, for example, might find themselves inadvertently working with a data set composed only of stocks (companies) that remained in existence throughout the period of the test. Of course this is a rookie error, but the ranks of 'financial analysts' are continuously brimming with shiny new rookies. What's more, easy and relatively inexpensive access to datasets via the Internet often makes it easy to overlook such a pitfall. Fun facts to know and tell."
Wayne: "As statistics and quantification have become the rule of the day in virtual everything in life these days, I purchased a text, written in the '40s, by a gentleman named Miller, on the basics of statistical analysis. I did not get too far into the book when I stumbled on Mr. Millers explanation that statistics essentially have their genesis in the fact that operations do grow to the point where management through personal/individual attention is no longer possible. I still reflect on that statement.
"This TEDx talk congeals Miller's statement to some degree for me: when applied to human events, statistical analysis is simply a record of available/gathered history; the same human shortcoming that brought about the need for statistics can still exist within the statistics. Big stuff may be missing. Mastery of the math is no guarantee to success.
"Sorry, this is a big subject for me, I may never fully grasp it.... But it fascinates me. Thanks for the link."
DavidB: "Years ago I took an avalanche awareness course. The instructor commented that it was important, when in the moving snow, to 'swim' upwards to the surface. I asked how you know which way is up? He responded that all the survivors he had interviewed knew which way was up. I asked, what about the ones that didn't survive? He was astonished at my question as he had never considered it."