On Throwing Out Data

Because of the recent election crisis in Florida, the issue of throwing out data has become a very important one that we are seeing acted out right before our very eyes. Notice that this submission was by a theory student. We cannot compartmentalize our learning into separate classes. I hope this discussion will help you all clarify the ways in which data are discarded, and when that is a reasonable thing to do.

On Wednesday, November 29, 2000, Samara Kenney wrote:

Hello Jeanne! This is Samara Kenney, I'm in your theory class. This is an email to let you know that I am prepared for that following exercises: Prepared! Multiple Interpretations Elections:Presidential Elections.

For this exercise I answered the question about the effect of throwing out data on the results. I believe that if the data that were thrown out are not included in the end results, the end results are ambigous.

On Friday, December 1, 2000, jeanne responded:

Hi, Samara. This is a good example to work with. You have identified an important concept: that of collapsing or throwing out data. A fundamental understanding of this concept will help you as you try to evaluate data analysis. And in the Florida election results we certainly had to do that.

But your e-mail doesn't tell me much about how well you understand the concept. First, it's conclusionary: "I believe that if the data that were thrown out is not included in the end results, the end results are ambigous." You've concluded that the results would be ambiguous, which was a good choice of words, so that those who disagree can't claim you're on the wrong side. But you haven't given me any steps in the argument. So I can''t follow your reasoning.

Now, as I recall, this practice was done in statistics class, in which I had already taught "recoding," not that all those students remembered the concept of recoding. But because we did have a foundation, I didn't try to define "throwing out data" for all of you who were tackling this issue of data analysis outside the stat class. Lecture Notes on the 2000 Elections.

Here's what I think you need to do:

• Review the material on throwing out data
• Then see me, or e-mail me, and make it clear for me that you now understand the process, and that you see how that links to adversarialism and mutuality in social theory.

Some hints for linking the elections to social theory:

• A major theme of our theory course has been that of adversarialism and mutuality. How does the issue of throwing out data relate to this theme? Consider that the winner takes all electoral votes.
• Consider the possibility that we decided our election on popular vote. Suppose a candidate wins by one vote. Would that make him President? How would that relate to the concept that winner takes all?
• Then consider the possibility that Fellman might be right, that our real problem might be adversarialism and "winner takes all."
• Then put each of these together to build an argument, and then conclude that the election results might be ambiguous. Then I can follow your argument and see whether or not I agree with your conclusion. But whether or not I agree doesn't matter. I will have been able to see your thinking.

---

---