In 1981, David Remnick — future editor of The New Yorker — wrote the following in The Washington Post: “… something has to be said for any kind of technology that can interpret hazy Miss America qualities like radiance and congeniality through its transistors.”
Remnick wrote of the statistical research of one George Miller, a Northern Illinois University professor who spent the 1980s irritating the producers of the Miss America pageant. Miller developed a predictive computer model called MISSAM that correctly predicted the pageant winner in five years out of 10 — and came very close the other five years.
Building a data-driven beauty queen
Miller built the model by analyzing data from the pageant’s 1921–1979 competitions. MISSAM analyzed everything from state and talent to education and body measurements to predict each contestant’s chances of winning. Was the contestant’s state Alaska, Delaware, Maryland, Missouri, Montana, Nebraska, Nevada, New Mexico, North Dakota, or Vermont? Too bad — the computer says she won’t win.
The next step analyzed the contestant’s “measurements.” If she fell into a specific range, she moved on; the next step dropped anyone with a common, long, or difficult surname. Next, MISSAM threw out any baton twirlers, along with ballerinas, dramatists, comedians, and mimes. With final considerations of education level, academic major, and eye color (but not hair color), Miller derived a final probability of winning. As Remnick described what he called “Miss Composite America 1981”: “She is 21 years old, 5 feet 6 1/2 inches, 114 pounds, has brown hair, one brown eye and one blue eye. Her favorite hobbies are cooking and baking and she loves to swim and ski. Her best talent is her singing.”
When the algorithm started beating the experts
Much to the producers’ frustration, Miller’s model continued until 1989, when rule changes made it unreliable. “The swimsuit gave me the edge that I needed to make the prediction,” Miller told UPI.
Did you enjoy this installation of SmarterMSP’s Tech Time Warp? Check out others here.
Photo: Shannon Volmar / Shutterstock
