January 31, 2009

Study Biases with Eponyms

Berkson's Bias: selection bias using hospitalized patients as the control.
Berkson's = Bad/sick patient
The history: Berkson was a statistician who recognized that small, selective sampling can cause false correlations.
The problem: "N vs n," statistically.  The data is a poor representative of a greater whole.

Pygmalion effect: Expectations affect the outcome.
Pygmalion = "Pointed": hidden agenda, "a self-fulfilling Prophecy"
The history: Pygmalion is a play in which a professor makes a bet that he can make a poor girl act like a lady. (aka My Fair Lady.)
The problem: unconsciously, researchers and participants conform to predisposed expectations.

Hawthorne effect: a tendency for those studied to affect outcome due to their knowledge of the study
Hawthorne = Heisenberg
The History: Hawthorne Works was a factory that studied the way lighting affected their workers. Productivity got better when the lights were dimmer. It also got better when the lights were brighter. Soon they realized it was because of the novelty of change (and the presence of researchers.)
The problem: observing someone affects the way they act.

1 comment:

  1. Just made a list of ophthalmology eponyms yesterday. They are indeed everywhere. Have a love/hate feeling with them. Not sure if they are just burning med students neurons, are really an efficient way to precisely point something to another "trained" person or just junk terminolgy to hide the science from profanes.