Re: Basic rookie question about standard error

By the way, if you remove the zeros, you would be making one of the
same mistakes that doomed Space Shuttle Challenger.
...really? What happened there?

Before the launch, someone was analyzing the data on # ring failures vs. temperature. They decided that the launches with no failures contained no information, and dropped them from the analysis. The resulting plot doesn't show much of a temperature effect. What they missed was that there were almost no zeros at low temperatures. Analysis with the zeros in tells a very different story.

The canonical statistical reference for this is a JASA paper by Dalal, Fowlkes and Hoadley. Feeding the three authors names to Google will get you lots of references.
.

Relevant Pages

  • Re: Basic rookie question about standard error
    ... someone was analyzing the data on # ring failures vs. ... > doesn't show much of a temperature effect. ... > there were almost no zeros at low temperatures. ... > The canonical statistical reference for this is a JASA paper by Dalal, ...
    (sci.stat.consult)
  • Re: Au Gratin Potatoes
    ... My oven doesn't reach that temperature;) so I'll assume you punched ... too many zeros. ... you can't fire porcelain in it. ... Pity! ...
    (rec.food.cooking)