numbers game

By Diogenes ( articles ) | Mar 09, 2007

If you followed the news coverage of the Discovery Channel special, "The Lost Tomb of Jesus," you know that the producers cited a statistical analysis, claiming that there was only an infinitesimal probability that another Jewish family tomb would show the same names as the ones found in what was allegedly the tomb of Jesus' family. You might have wondered how those statistics were calculated.

The "Numbers Guy" in the Wall Street Journal wondered, too. So he contacted the University of Toronto statistician who furnished the figures. It turns out that-- surprise!-- the statistician began his calculations with the assumptions that the filmmakers gave him: that "Mariamene" was an unusual name referring to Mary Magdalen, rather than a variation on the very common "Mary;" that "Yose" was an unusual name rather than a form of the popular "Yosef;" and so on. Without those assumptions, the connection between this tomb in Jerusalem and the family of Jesus would have been "statistically not significant."

"When I was doing the calculation, I was naively unaware of the extent to which the filmmakers might be depending on the ultimate result of it," the Toronto professor told the Journal. "I did carry out the calculation in every good faith."

In short, the calculations were accurate. The data that went into the calculations... Well, you know the old principle of data analysis by computers: Garbage in, garbage out.