Subscribe to the OSS Weekly Newsletter!

One of Last Century’s Most Influential Social Science Studies Is Pretty Bad

Research done at the Hawthorne Works in the 1920s and 30s, which yielded the “Hawthorne effect,” is a mess of sloppiness and misogyny
Image by

We love putting names to things, especially if those names are scientific. Just look at the variety of phenomena people love to refer to as the Dunning-Kruger effect: the idea that other people (not me!) overestimate what they know, the sighting of someone being aggressively wrong, or simply the belief that dumb people don’t know they are dumb.

The fact that science has studied a phenomenon and plastered a name over it feels good. But sometimes, when we dig into the origin of these scientific stories, we discover they have been heavily distorted in the telling. For example, the creation myth of modern placebo research, which has Dr. Henry Beecher running out of morphine while treating World War II soldiers and improvising with a saline solution to surprising results, simply crumbles when investigated. There is no primary source to support this story, but what a story it is!

Today, yet another scientific myth must come down. You may have heard that, a long time ago, factory workers were studied and their work environment was changed in different ways. No matter what the scientists did, the workers’ productivity increased. It’s the idea that being part of a study changes our behaviour, perhaps because of the attention we are receiving. This is known as the Hawthorne effect and it played a role in the inception of the human resources departments we have today.

But when we transport ourselves back a century to study what actually happened at the Hawthorne plant, we discover a series of experiments whose designs would not pass muster if a modern-day high schooler proposed them.

An easy explanation

West of the downtown core of Chicago lies the suburb of Cicero, Illinois, where Al Capone lived in order to stay outside the reach of the Chicago police. Originally, however, Cicero was named Hawthorne, and for nearly a century the town was the site of a factory complex known as the Hawthorne Works. In 1994, it was demolished to make way for a shopping centre parking lot, but at its peak, it employed over 40,000 men and women, many of whom were European immigrants. The complex belonged to Western Electric, the sole supplier of telephone equipment to AT&T, which held the monopoly on the telephone industry at the time. From 1924 to 1932, experiments were conducted at the Hawthorne Works, beginning with changes to the artificial lighting, to see how productivity could be improved. While these experiments were never properly documented and written up, their conclusions made their way into a trade journal. In 1953, the phrase “Hawthorne effect” appears to have been used for the first time in reference to these studies. It was printed in a textbook chapter written by J.R.P. French, who defined is as “marked increases in production which were related only to the special social position and social treatment” the workers received.

From there, the Hawthorne effect made its way around the globe, positioning itself in first-year student textbooks as one of those classic experiments that taught us so much about human nature, like the Milgram experiment on delivering electric shocks and the Stanford prison experiment on dehumanization. Because, these textbooks told us, workers at the Hawthorne plant had always worked harder no matter what the experimenters did, the very participation in a study changed someone’s behaviour. But why that behaviour had changed was explained in slightly different ways depending on who was telling the story. It was because the workers had received special attention, said one. It was because they were supervised differently, said another. It was because they were simply aware that they were being studied, perhaps, or maybe they had formed an idea as to what the object of the study was and they were trying to play along.

The phrase “Hawthorne effect” has come to mean all of these jumbled-up things, and as many critics have pointed out in the academic literature, it’s an easy label to paste over any kind of weird finding you get when you experiment on people. “The workers were more productive when we dimmed the lights? That’s strange. Must be the Hawthorne effect!” Case closed.

Given that the data from the original Hawthorne studies were never published, we could simply get lost in speculations over what exactly happened. But luckily, modern researchers have found some of the data from these experiments and what they have exposed is gross oversimplifications, sloppy methodology, and a sociopolitical landscape that gets ignored when this creation myth is perpetuated.

“She’s gone Bolshevik!”

It all began with the National Research Council deciding to study if better lighting improved worker productivity. This was the first series of studies conducted at the Hawthorne Works, and the result is often said to be that no matter if the artificial lights inside the factory were brighter or dimmer, workers worked harder. In 2011, however, John A. List and Steven D. Levitt (the latter of Freakonomics fame) published their analysis of these illumination experiments, the data from which they had just discovered on microfilm in university archives. It’s fair to say that this string of studies, which took place between 1924 and 1927, was very messy. There were other explanations for what happened. Lights were changed on Sundays, which was the only off day for these factory workers, and productivity did often increase the following Monday. Of course, when you have been working six days straight and finally get a day off, is it any wonder that you’re more productive after this break?

The Hawthorne Works had its own power generator and the artificial lighting varied in intensity because of frequent voltage fluctuations. Combined with the natural light that was coming in from the windows and which varied from season to season, the idea that strict changes to the light bulbs had a significant impact on overall productivity is far from having been proven. Moreover, there is the issue of control groups. There were three illumination experiments, with the first and last lacking a control group to compare the increase in productivity to, and the middle experiment was quite telling: productivity increased in the room where the lights were changed, but also in the rooms that were not part of the experiment. Even though the Hawthorne studies are, according to Levitt and List, “among the most influential social science research of the twentieth century,” the illumination experiments which launched these investigations yielded mixed results, at best.

Then came the relay studies.

Relays were electrically operated switches used in telephone exchanges. At the Hawthorne Works, women assembled them by conducting 32 distinct operations with each hand. The completed relay was dropped into a chute and recorded. These women had to assemble these relays every minute for up to nine hours a day, up to five and a half days a week, 300 days a year. It was monotonous work.

Five women were taken off of the main factory floor and into a test room where they had to continue assembling relays while the conditions of their work were changed. From 1927 to 1932, the men overseeing the study played with the number of weekly days these women had to work, the length of their breaks, where they were sitting, and the incentives they were given. The idea that the women’s productivity went up no matter what the experimenters threw at them is not true, according to Stephen R. G. Jones from McMaster University, who reanalyzed the data in the 1990s.

Once again, the experiments were messy. In science, two things may look like they are linked, but the link actually comes from a third variable known as a confounder. Drowning may be linked to sales of ice cream, for example, but of course the confounder here is the summer season, which causes increases in both. Researchers who have dug into exactly what happened during these relay studies at the Hawthorne plant have found plenty of potential confounders. For example, the working conditions changed not just because of what the experimenters did, but also because of the times. The Roaring Twenties came to an end in late 1929 when both the London Stock Exchange and Wall Street crashed, ushering in the Great Depression. At the tail end of the Hawthorne studies, an employee might have worked harder to avoiding finding themselves unemployed in a bleak financial landscape. Also, two of the women in the group of five were replaced during the experiment, allegedly because they were too chatty, or perhaps because of animosity between them and the other study participants. Rigorous these experiments were not.

More worryingly, the sanitized tale of the Hawthorne effect, in which women happily work harder because “stuff is happening around them,” dehumanizes the workers and hides important sociopolitical substance. Part of the reason why this small group of women worked harder in the special relay test room had to do with the power they received. They were less supervised than the women in the main room and they were consulted with regards to how their conditions would be changed. Some of them were interviewed fifty years later: they had been working so hard because they did not want to be returned to their former department, where their supervisor was said to be very harsh. Also, when asked what they had liked about being in the test room, one of them immediately said, “We made more money in the Test Room.”

In a commentary on the Hawthorne Studies, Edwin Gale sums up the fable of the Hawthorne Works studies as “the story of investigators who wanted to make the sweat-shop conditions of factory life in the 1920s more humane and yet more profitable.” Western Electric forbade union membership, but the company was also weary of strikes and employee unrest. The academic consultant who was brought in to oversee the experiments, Elton Mayo, did not view this unrest as a problem of work conditions but as one of psychopathology. The problem was bad apples, not the system as a whole. In fact, he wrote of one of the women in the relay studies who had to be replaced that she “was reported to have ‘gone Bolshevik’ and had been dropped.” When confronted with the political undertones of this conflict, he wrote that, actually, the woman had been anemic at the time, which produced “paranoid preoccupations.”

It is ironic, then, that these famous experiments at once inspired a more human approach to employer-employee relationships—moving managerial thinking away from treating their workers as robots in need of optimization and paving the way for human resources departments—and offered a bigoted, dehumanizing view of workers. As Michiel Kompier wrote in his discussion of the myth, there is something cynical about positioning the recognition that workers, like their employers, also have feelings as “a scientific breakthrough.”

The shallow myth of the Hawthorne effect endures for a variety of reasons. It’s a great story that university freshmen are inculcated with at the beginning of their studies, which makes it stick when they go off into their career. It provides managers with a helpful tale of productivity being solely tied to social concerns, hence alleviating the need for increased pay or improved working conditions. It also endures because even academics cannot always escape from intellectual laziness. If a textbook says it’s true, it must be.

Obviously, people may change their behaviour when they are being watched. Researchers, however, need to decide exactly what type of behaviour change they want to control for in their studies. Are they worried about participants trying to please them? Are they concerned about participants realizing they are being observed? Might participants give different answers because they are ashamed of the truth? These are all different effects and studies need to better control for them specifically as opposed to gather up any strange results and slap a “Hawthorne” sticker over them.

A 1989 study of 86 studies that tried to isolate this Hawthorne effect found no evidence for it. There is no one Hawthorne effect. Given the complexity of what happened at the Hawthorne Works and the sloppiness of these experiments, it may be time to retire the phrase “Hawthorne effect.”

Take-home message:
- The Hawthorne effect is a confusing name given to the idea that people in a study change their behaviour simply because they are being studied
- The effect comes from sloppy research done at the Hawthorne Works factory in the 1920s and 30s, but the increase in worker productivity that has been reported to follow whatever the scientists did to change the workers’ environment can be explained by a number of other factors
- Instead of explaining away any strange finding as “the Hawthorne effect,” scientists must do a better job at controlling for specific effects that may occur due to participation in a study, like wanting to please the researchers or feeling observed


Back to top