# Statistics -- An allegory (with apologies to Phil Runkel)

[From Bruce Gregory 960409.1145 EDT]

A social scientist decides to test the ability of people who have
just seen a movie to recall the plot as they emerge from the theater.
He finds that 20% of the sample he interviews can recall the plot.
This finding is robust. Samples interviewed on different days
support the 20% finding. The social scientist can predict with some
confidence the results of a survey to be conducted on some future
date. Only 20% of the people interviewed leaving a theater can recall
the plot of a movie just shown in the theater

The scientist's associate is not convinced that the plot of movies is
so difficult to understand. She conducts her own sample. In her
protocol she first asks each interviewee which movie he or she has
just seen (the movie complex has five theaters). Only if the person
has seen the movie whose plot she is asking about, does she proceed
with the interview and count the person in the survey. She finds
that 98% of the people in her survey can recall major features of
the plot. The second social scientist is accused of introducing
biases into her results because her sample consists of "self-
selected" individuals.

Bruce G.