The Hawthorne Effect Probably Wasn't Real

From 1924 to 1933, researchers at the Western Electric Hawthorne Works outside Chicago ran a series of experiments on what makes factory workers productive. They turned the lights up; output rose. They turned the lights back down; output rose again. The conclusion in the published write-ups was that workers responded to being studied, not to the lighting — an effect named for the plant and cited in almost every management textbook since.

Steven Levitt and John List went back to the original records, which Harvard Business School had archived but never fully re-examined, and published their reanalysis in 2011 in the American Economic Journal: Applied Economics. The raw data told a different story. The illumination experiments had been poorly controlled, sample sizes were tiny, and the output increases correlated more strongly with the day of the week and the onset of a new experimental session than with any manipulation of lights. When Levitt and List corrected for those variables, the "Hawthorne effect" largely disappeared from the data that had given it its name.

The folklore is now almost a century old and cited in thousands of papers as a known behavioral phenomenon. A real effect — people acting differently when observed — does exist and is well-documented in more rigorous studies. What Levitt and List showed is that the specific experiment that coined the term didn't demonstrate it.

The story survives partly because it's useful. Managers like to explain away productivity bumps as short-term attention effects. The idea that workers briefly work harder under scrutiny is a convenient ceiling on reform. And the term "Hawthorne effect" is a handy shorthand even when the original evidence is bad — which, by a certain irony, means the Hawthorne effect is itself a case study in the Hawthorne effect's failure mode: published, memorable, and slightly less true than everyone thinks.