Are the Talpiot statistics neutral?
Today's post by James Tabor is very interesting. He seeks to address the role of the statistician in doing math on the Talpiot. The Discovery Channel reported that 600 to 1 the tomb belongs to Jesus of Nazareth. This means that 599 times out of 600, this conclusion will be true.
Tabor writes, "What the statistician is asked to do is to determine the probability of the cluster of names based on name frequency data in late 2nd Temple Jewish Jerusalem. It is a relatively simple task. That is it." Tabor is seeking to defend the stats that have been arrived at by Andrey Feuerverger, the statistician, in my opinion, because this is the only piece of evidence that supports the conclusion to which Tabor leans.
However, something is amiss in Tabor's defense. He writes, "Nothing in Feuerverger's math determined or informed the next step of the analysis, namely the issue of whether these names, rare as they have been shown to be in this cluster, are 'highly appropriate' as names for the Jesus family." This is true. The math does not make specific assignments, but it already assumes that certain names are highly appropriate and that the (quoting Feuerverger) "results of any such computations are highly dependent on the assumptions that enter into it. Should even one of these assumptions not be satisfied then the results will not be statistically meaningful." What are some of these assumptions (found at the above link, too)? That names on the ossuaries (the Mariamenou e Mara is appropriate for Mary Magdalene; that Yose is appropriate to Jesus' brother Joses), are highly appropriate for people we know associated with Jesus. In this way, the conclusions are indeed worked into the math before the historians work it out the conclusions. It assumes that certain names are highly appropriate for certain people we know associated with the person whose identity we are trying to figure out, which makes it highly probable that the person will end up being attached to the people we are already assuming. Let me say it like this. Suppose I know a person named James Doe, who has a best friend named Jill Smith and a brother named John Doe. I find a piece of paper that is addressed to "James" and signed "Love, Jill and John." If I assume that Jill is highly appropriate to Jill Smith, and John is highly appropriate to John Doe, then I will automatically increase the probability that the "James" to whom the paper is addressed is James Doe. Now, this is not a parallel situation, as whether or not the assumptions made on the tomb names are appropriate is an important dialog to have. But they are assumptions in the original calculations.
The stats that Tabor is defending just aren't so neutral.
Tabor writes, "What the statistician is asked to do is to determine the probability of the cluster of names based on name frequency data in late 2nd Temple Jewish Jerusalem. It is a relatively simple task. That is it." Tabor is seeking to defend the stats that have been arrived at by Andrey Feuerverger, the statistician, in my opinion, because this is the only piece of evidence that supports the conclusion to which Tabor leans.
However, something is amiss in Tabor's defense. He writes, "Nothing in Feuerverger's math determined or informed the next step of the analysis, namely the issue of whether these names, rare as they have been shown to be in this cluster, are 'highly appropriate' as names for the Jesus family." This is true. The math does not make specific assignments, but it already assumes that certain names are highly appropriate and that the (quoting Feuerverger) "results of any such computations are highly dependent on the assumptions that enter into it. Should even one of these assumptions not be satisfied then the results will not be statistically meaningful." What are some of these assumptions (found at the above link, too)? That names on the ossuaries (the Mariamenou e Mara is appropriate for Mary Magdalene; that Yose is appropriate to Jesus' brother Joses), are highly appropriate for people we know associated with Jesus. In this way, the conclusions are indeed worked into the math before the historians work it out the conclusions. It assumes that certain names are highly appropriate for certain people we know associated with the person whose identity we are trying to figure out, which makes it highly probable that the person will end up being attached to the people we are already assuming. Let me say it like this. Suppose I know a person named James Doe, who has a best friend named Jill Smith and a brother named John Doe. I find a piece of paper that is addressed to "James" and signed "Love, Jill and John." If I assume that Jill is highly appropriate to Jill Smith, and John is highly appropriate to John Doe, then I will automatically increase the probability that the "James" to whom the paper is addressed is James Doe. Now, this is not a parallel situation, as whether or not the assumptions made on the tomb names are appropriate is an important dialog to have. But they are assumptions in the original calculations.
The stats that Tabor is defending just aren't so neutral.
posted by Aaron Perry at
9:36 AM
1 Comments:
I think you have slandered the paragon of academic virtue yet again. Fortunately for you, you are not alone. If the lawsuits start coming, surely Amos Kloner, Joe Zias, Ben Witherington, and even Warren Kinsella (!), will be ahead of you.
anonymous.
PS. I didn't like your tone. Anyone who disagrees with Prof. Tabor is a slanderer and decidedly atonal.
Post a Comment
Subscribe to Post Comments [Atom]
<< Home