The concept of survivor bias, if you don’t know it, is well worth being aware of. It’s most often used in finance, where it refers to a “tendency for failed companies to be excluded from performance studies” (thanks, Wikipedia). Think of the Dow Jones Industrial Average, which indexes the stock prices of 30 of the largest and most important U.S. companies — until, that is, one of said companies does so poorly that it is booted from the index and is replaced by a company that’s doing better.
Over time, therefore, the DJIA reflects a different reality than many people presume. It is biased toward survivors — or, if you want to think of the concept more broadly, toward winners.
This winner’s bias, if you will, shows up in pretty much every realm imaginable: academics, medicine, politics, etc. I don’t mean to sprinkle skepticism all over your inherently positive thoughts about the world, but I do think it’s worth keeping winner’s bias in mind whenever you read (or write) something about the performance of a given group or institution or coalition.
Winner’s bias is perhaps especially pronounced in sport. The behaviors of winners are remembered and dissected far more thoroughly than those of losers, and given greater weight, even if the outcome was decided by a tiny margin.
In the sports section of yesterday’s USA Today, I came across not one but two subtle examples of this phenomenon, in two articles on the very same page. Both articles discussed the fourth-quarter comeback of a much-favored N.F.L. team in a Monday night game; both contained quotes about a star player urging his team on to success. First, the Patriots came back to beat the Bills:
“When Tom [Brady] came in the huddle at 5:32 and said, ‘Guys, let’s get it going. We’re going to win this game,’ everything just went positive from there,” [Randy] Moss said.
And then the Chargers came back to beat the Raiders:
“L.T. [LaDainian Tomlinson] came over and said, ‘We’re going to win this football game,’” right tackle Jeromey Clary said.
“That kind of got us going again,” [Darren] Sproles said.
I have a simple question: does anyone really think that Brady and Sproles don’t say that kind of the thing in the huddle when their teams are behind and fail to come back?
Attributing causal effect to such a statement might seem natural for a player whose team just earned a big win, or for a sports reporter soaking up the happy locker-room vibe. But if you think about it, this kind of thinking is essentially nothing more than superstition. I’m not saying that people shouldn’t traffic in superstition if it turns them on (yeah, I’ve got a few of my own hoodoos, thanks). But when you’re trying to measure performance, we should remember to not actually believe in it.
and 
. You can also watch it on 
If I may theorize why this type of survivor bias is so common; it’s because I believe that there is a causal effect between a negative attitude and a negative performance. Actually, I haven’t seen any data in the sporting world confirming this, but Robert Sutton (Stanford prof) does highlight this effect for organization change efforts (see his book,”Hard Facts, Dangerous Half-Truths, and Total Nonsense”). Sutton’s summary of the studies on change is that negative attitudes towards change efforts tend to be a self-fulfilling prophecy. So while the opposite may not be true, it’s easy to see why people draw a linkage between attitude and performance.
Yea agreed, I don’t really think this should be classified as survivorship bias at all.
You mean the people who lose games or wars weren’t defeated because our side believed more in God?
This is definitely survivor bias. You don’t get to the level of professional football — the top thousand people in the country at it — without (among other things) being really, really confident that you’re always going to win, that you’re always going to come back, that you’re always going to be better the next time. No one in the NFL slumps to the huddle when his team is down by less than two touchdowns and says, “Geez, guys, I dunno… feel like packing it in? I mean, it’s a long bus ride back to the hotel. We could beat the crowds.”
IMO, the most famous example of survivor bias is evolution. People say, “it’s impossible that the elephant evolved without help from a higher power!” Yet how many species died out? We see only the ones that made it. The dead ends are not counted, although they are far more numerous. In that context, the elephant is not amazing at all.
Also concur with #5. How about some statistics on how many make it/fail? My guess is less than 5% get to the top.
While I’m a fan of Dubner, he’s wrong on this point. DJIA is not a case of survivor bias as the performance of the index reflects stocks that are falling until they are removed.
A better example of survivor bias is a mutual fund company that brags that 90% of its funds beat their respective index over the last 10 years, perhaps ignoring that fact that underperforming funds were closed or merged into outperforming funds.
Has anyone ever done a study to look at the performence of the companies listed in the DOW of any particular year?
Not to be picky, but since Moss and Brady have been on the same team (excluding last season when Brady was injured all year) the Patriots have won 19 out of 20 games. The one loss, the Super Bowl against the Giants, was lost by the Patriots defense after Brady seemed to win the game late (with a TD pass to Moss no less). From this perspective, in Moss’ mind, Brady has never failed to follow through on winning the game!