Statistical Inference & Techniques

This Flash based applet simulates data from a case study of treatments for tumor growth in mice. This simulation allows the user to place mice into a control and treatment groups. The simulation then compares the difference in the groups based on this haphazard selection to those of a truly random assignment (the user may also create multiple random assignments and examine the sampling distribution of key statistics). The applet may be used to illustrate three points about random assignment in experiments: 1) how it helps to eliminate bias when compared with a haphazard assignment process, 2) how it leads to a consistent pattern of results when repeated, and 3) how it makes the question of statistical significance interesting since differences between groups are either from treatment or by the luck of the draw. In this webinar, the activity is demonstrated along with a discussion of goals, context, background materials, class handouts, and assessments. Key Note for Instructors: The data are drawn from a real experiment with an effective treatment but where the response is correlated with animal age and size (so tumor size will tend to be smaller in the treatment group when measured at the end of a randomized experiment but animal age and size should not be). Typically people choosing haphazardly will tend to pick larger/older animals for the treatment group and thus create a bias against the treatment.

It is now proved beyond doubt that smoking is one of the leading causes of statistics. Quote found in "Reader's Digest" (December, 1961) by journalist Fletcher Knebel (1911-1993)

Joke from "The Little Black Book of Business Statistics", by Michael C. Thomsett (1990, Amacom) p. 117. also quoted in "Statistically Speaking" compiled by Carl Gaither and Alma Cavazos-Gaither.