Stooges Statistics

A bop on the head, a punch in the stomach, a poke in the eyes. These were staples of the slapstick comedy of The Three Stooges. Now they've become elements of a novel approach to teaching experimental design, sampling, data gathering, and statistics.

I can remember watching Three Stooges episodes on television when I was a kid, typically on Saturday mornings. In 1963, I was part of a record crowd of more than 35,000 people who jammed Grandstand Stadium in Toronto during the Canadian National Exhibition to see the knockabout trio in person.

The Three Stooges starred in 190 short films, starting in 1934. In the 1950s and 60s, these early black-and-white films were recycled on television, entrancing a new generation of viewers. They remain a popular fixture on cable TV, and you can sample favorite episodes and clips on YouTube. There's even a YouTube-posted 1963 home movie of Moe, Larry, Curly-Joe, and an ape cavorting around an airplane at the airport in Toronto. Last fall, Sony started releasing the complete collection of Stooges shorts, in chronological order.

For mathematicians (and Stooges fans) Robert Davidson and Robert Gardner of East Tennessee State University, the continuing popularity of The Three Stooges and ready availability of the films presented an opportunity. They looked into using these engaging films as data sources for an introductory statistics class.

"People know The Stooges," Gardner says.

Over the 50-year career of The Three Stooges, Moe Howard and Larry Fine were the constants. Curly (Moe's brother) appeared in the first 97 shorts but suffered a stroke in 1947. His brother Shemp took over for 77 episodes between 1947 and 1955. Joe Besser appeared in the last 16 films.

In these shorts, Moe typically doles out the violence, and Larry and especially Curly (and later Shemp) are the unfortunate recipients. When Joe Besser appeared on the scene in 1955, his contract specified that he suffer no slapping or other bodily harm while filming the shorts. Davidson and Gardner wondered whether the level of violence in the films changed as a result.

Data gathering required watching segments of a random sample of shorts starring Curly, Shemp, and Joe. As they watched, the viewers tallied the number of violent acts perpetrated by Moe against his hapless victim.

Davidson and Gardner first tested the null hypothesis that "the average number of violent acts by Moe against Curly per episode is the same as the average number of violent acts by Moe against Shemp." Statistical tests on the tallies suggested that Shemp may have suffered more than Curly did, but the difference wasn't statistically significant enough (only 81% confidence) to warrant rejection of the null hypothesis. The mathematicians needed more data to settle the question.

Comparing Curly and Joe, Davidson and Gardner found that they could safely reject the null hypothesis and affirm that "the average number of violent acts by Moe against Curly per episode is greater than the average number of violent acts by Moe against Joe."

This exercise raises some interesting questions about data collection in the real (or reel) world. In the context of the Stooges films, for example, what constitutes an act of violence? If Curly gets slapped three times in rapid succession, is that one or three instances? What about an accidental poke?

Students "experience the difficulty in designing an experiment, the problems inherent to data collection, and analysis of the collected data," Davidson and Gardner note.

There are plenty of other questions about The Three Stooges that may be worthy of statistical attention. Does the level of violence vary from director to director or from writer to writer? Did Larry suffer less than, as much as, or more than Curly or Shemp? Did that change during the Joe Besser years? You don't even have to focus on violence. What roles did women or minorities play the films? How did that change over time? What sort of language did The Stooges use (or misuse)?

Davidson and Gardner will describe using The Three Stooges as a data source for motivating statistics students at the 15th Georgia Conference on College & University Teaching, Feb. 15, 2008, at Kennesaw State University.