Permit me to interrupt your search for the market's cheapest stocks with a brief discussion of an important topic. Are you familiar with Jeremy Siegel's Stocks for the Long Run? It's a seminal investing text about Siegel's study of stocks all the way back to 1802 -- more than 200 years of data!

I've long referred to his findings, because it's useful for us investors to know that over long periods, the stock market has tended to average about 10% in annual gains. (Of course, during your specific investing time frame, the average likely will be higher or lower than that.)

There's a problem with such long-term studies, though: survivorship bias.

A mutual problem
Survivorship bias affects all kinds of long-term studies, and mutual-fund statistics are a prime example. For example, you might learn that a certain class of mutual funds has averaged this or that over time, but has the research taken into account the many funds that failed and closed their doors over that time? Probably not. So the numbers you're looking at reflect only funds that have not crashed or closed -- and that means that the average performance figures you see are probably rosier than they would have been if they had included those stinker funds.

Take this as a warning to be wary of certain mutual fund statistics.

Stocks are more complicated
Some will also accuse stock indexes of survivorship bias. Consider, for example, the Dow Jones Industrial Average (known as the Dow). It came to life in 1896, priced at 40.94. Today, 113 years later, it's around 9,400. Here are the 12 companies in the index when it began -- see if you notice anything:

  • American Cotton Oil
  • American Sugar
  • American Tobacco
  • Chicago Gas
  • Distilling & Cattle Feeding
  • General Electric (NYSE:GE)
  • Laclede Gas
  • National Lead
  • North American
  • Tennessee Coal & Iron
  • U.S. Leather pfd.
  • U.S. Rubber

That's right -- they're probably not familiar, except for General Electric. Today, the Dow contains 30 much more familiar companies, such as Caterpillar (NYSE:CAT), Hewlett-Packard (NYSE:HPQ), and AT&T (NYSE:T). So the Dow's rise from 40.94 to 9,400 has not purely reflected the progress of those initial 12 companies. Some of them went out of business, while others merged or were otherwise transformed into companies familiar to us today. The American Cotton Oil company, for example, morphed into Bestfoods, which was later bought by the Dutch giant, Unilever (NYSE:UL).

It's the same with indexes such as the S&P 500, which contains companies big and less big, from Microsoft (NASDAQ:MSFT) to RadioShack. Companies are added and dropped from the list all the time. Recently, for example, General Motors was dropped and DeVry added, and CIT Group (NYSE:CIT) gave way to Red Hat. Poor performers end up dropped -- but often only after they plunge, with that performance incorporated into the index's results.

With stocks, though, survivorship bias can understate actual returns. For instance, some good performers end up merging with other healthy companies, and leaving an index for that reason, instead of poor performance. And if a stock like Red Hat goes into an index only after several successful years of good returns, then the index misses out on it, arguably reducing the index's return.

So the survivorship bias in indexes can swing both positive and negative.

Don't stop looking at historical data -- it's still useful. But do keep in mind the possibility that sometimes survivorship bias is at work, skewing results one way or the other.

More Foolishness: