Hawthorne effect

Based on Wikipedia: Hawthorne effect

The Factory Where Everything Seemed to Work

In the late 1920s, something strange happened at a factory outside Chicago. Researchers turned up the lights, and productivity went up. They turned the lights down, and productivity went up again. They gave workers more breaks, and output increased. They took the breaks away, and output still increased.

It was as if the workers had decided to mock the very idea of scientific research.

The Hawthorne Works was a massive Western Electric manufacturing plant in Cicero, Illinois, where workers assembled telephone equipment. The company wanted to know something simple: would better lighting make workers more productive? It seemed like a straightforward question with a straightforward experimental design. Change the lighting, measure the output, draw your conclusions.

What they discovered instead would reshape how we think about human behavior, workplace psychology, and the very act of observation itself. Or did it? The answer to that question is considerably more complicated than most psychology textbooks would have you believe.

The Illumination Experiments

The lighting study ran from 1924 to 1927. The setup was classic experimental design: an experiment room where lighting levels would be manipulated, and a control room where everything stayed the same. The hypothesis was elegant in its simplicity. More light equals better working conditions. Better working conditions equals higher productivity.

The results were baffling.

When researchers cranked up the lights in the experiment room, productivity rose. So far, so good. But productivity also rose in the control room, where nothing had changed. The researchers tried dimming the lights instead. Productivity went up in both rooms again. They kept dimming. Still up. Only when the light levels dropped to approximately the brightness of moonlight—when workers could barely see what they were doing—did productivity finally begin to fall.

The researchers had stumbled onto something unexpected. Whatever was driving productivity, it clearly was not the light bulbs.

The Relay Assembly Test Room

In 1927, a new phase of experiments began. This time, researchers selected two female workers and asked them to choose four colleagues to join them in a special test group. These six women would assemble telephone relays in a separate room, isolated from the regular factory floor, while researchers tinkered with their working conditions.

The measurement system was clever. Each finished relay dropped down a chute, where it was automatically counted. To establish a baseline, researchers began counting two weeks before the women even knew they were being studied. Then the experiments began.

The researchers tried everything. They introduced two five-minute breaks. Productivity increased. They switched to two ten-minute breaks. Productivity increased further. They tried six five-minute breaks, and productivity actually dropped—apparently, too many interruptions broke the workers' flow. They provided morning soup and coffee with sandwiches. Productivity went up. They shortened the workday by half an hour. Up again. They eliminated Saturday shifts. Up once more.

Here is where things got truly strange. When the researchers reverted any variable back to its original condition, productivity often increased anyway. It was as though the mere act of changing something—anything—was enough to boost output.

Enter Elton Mayo and the Birth of a Concept

Elton Mayo, a Harvard Business School professor, became the most prominent interpreter of these puzzling results. His explanation centered on something that had nothing to do with lighting or break schedules or soup. Mayo argued that the workers were responding to attention itself.

The six women in the relay assembly room were not just workers anymore. They had been chosen. They were special. They worked in their own room with a sympathetic supervisor who discussed their performance with them. They formed a team. According to Mayo, they "gave themselves wholeheartedly and spontaneously to cooperation in the experiment."

In Mayo's telling, the workers felt seen. They felt valued. They felt like they had some control over their situation rather than being anonymous cogs in an industrial machine. And this feeling, not any particular change in working conditions, was what drove their improved performance.

The term "Hawthorne effect" was not actually coined until 1953, when a researcher named John R. P. French needed a name for this phenomenon. Henry Landsberger, a sociology professor at the University of North Carolina at Chapel Hill, later refined the definition: the novelty of being research subjects, combined with the increased attention from researchers, could lead to temporary increases in productivity.

What the Effect Actually Claims

The Hawthorne effect, as it came to be understood, describes a specific kind of behavioral reactivity. When people know they are being observed, they modify their behavior. Not because of any particular change in their environment, but simply because of the observation itself.

This is different from related phenomena, though the boundaries can blur. The placebo effect occurs when people experience real improvements because they believe they are receiving treatment, even when they are not. The Hawthorne effect is broader—it does not require any belief about treatment, just awareness of being watched.

It is also distinct from what psychologists call demand characteristics, where research subjects try to figure out what the experimenter wants and then provide it. Hawthorne subjects were not necessarily trying to please anyone. They might simply have been more alert, more engaged, more careful, because they knew someone was paying attention.

The opposite of the Hawthorne effect might be described as behavioral entropy—the gradual decline in performance that occurs when no one is watching, when routine sets in, when workers become invisible even to themselves.

The Experiments Nobody Talks About

Most textbooks that mention the Hawthorne effect focus almost exclusively on the illumination studies and the relay assembly room. But the research at Hawthorne Works was far more extensive, and the later experiments produced results that complicated the tidy narrative considerably.

Between 1928 and 1930, researchers conducted a mass interviewing program involving twenty thousand workers. Initially, they used direct questioning—yes or no answers about company policies and supervision. This proved nearly useless for understanding what was actually happening. When they switched to a more open-ended approach, simply listening to what workers had to say, they discovered something important: individual behavior was profoundly shaped by group dynamics. Workers did not operate as isolated units responding to management. They operated as members of social groups with their own norms and expectations.

The bank wiring observation experiment, conducted between 1931 and 1932, drove this point home even more forcefully. Fourteen men who assembled telephone switching equipment were studied to see how payment incentives affected productivity. The expectation was simple: pay people based on individual output, and they will produce more.

Productivity actually decreased.

The researchers discovered that workers had formed informal groups—cliques—with their own unwritten rules. These groups developed mechanisms to enforce conformity among their members. When managers asked questions, clique members gave identical answers, even when those answers were not true. The workers had a shared fear: if productivity went up too much, management would either raise the expected baseline or use the higher numbers to justify layoffs.

The men were more responsive to the social pressure of their peer groups than to the financial incentives dangled by management. This was not the Hawthorne effect as popularly understood. This was something closer to collective resistance.

The Skeptics Arrive

Richard Nisbett, a prominent psychologist, has called the Hawthorne effect "a glorified anecdote." His criticism is pointed: "Once you have got the anecdote, you can throw away the data."

This is harsher than it might sound at first. What Nisbett means is that the story of the Hawthorne effect has become more important than the actual evidence behind it. People remember the narrative—workers perform better when observed—without examining whether the data actually support that narrative.

And when you examine the data closely, problems emerge.

The relay assembly room study, the most famous of the Hawthorne experiments, was methodologically weak. Six subjects is a tiny sample. There was no proper control group. Many variables were changed simultaneously, making it impossible to isolate which ones mattered. The study ran for only about a year, and the researchers who conducted it had no way to know whether the productivity gains were permanent or temporary.

J. G. Adair, who reviewed the literature on the Hawthorne effect, found what he called "gross factual inaccuracy" in most secondary publications. Many studies that cited the Hawthorne effect had not actually found evidence for it in their own research. Adair argued that when the effect did occur, it was better understood as a variant of experimental demand effects—subjects trying to figure out what was expected of them and behaving accordingly.

The Hunt for the Original Data

Here is a remarkable fact about the most famous study in the history of industrial psychology: for decades, no one could find the original data from the illumination experiments.

The economists Steven Levitt and John List—Levitt is best known as the coauthor of the book Freakonomics—spent years searching. They finally located the data in 2011, on microfilm stored at the University of Wisconsin in Milwaukee.

When they reanalyzed it, they found slight evidence for a Hawthorne effect over the long run. Slight. Not the dramatic, transformative effect that had been claimed for nearly a century. Their findings supported earlier work by S. R. G. Jones, who had examined the relay experiments in 1992 and reached similar conclusions.

Despite this, List has said he remains confident that the Hawthorne effect is real. The effect may exist, in other words, even if the original study that supposedly demonstrated it did not actually demonstrate it very well.

What Else Might Have Been Happening

If the Hawthorne effect was not as strong as claimed, what else might explain the productivity changes observed in the 1920s?

One possibility is surprisingly mundane: the workers got better at their jobs. This is sometimes called a learning effect. When you do the same task day after day, you improve. The experiments ran long enough that simple skill development could explain some of the productivity gains.

Another explanation involves feedback. In the relay assembly room, workers could see their own output numbers. They watched finished relays drop down chutes into counters. For perhaps the first time, they had real-time information about how they were performing. This kind of feedback can drive improvement even in the absence of any external observation.

A researcher named Parsons argued that the Hawthorne effect might really be about performance feedback rather than observation per se. If you give people information about how they are doing, they adjust their behavior—learning from what works, correcting what does not.

Some researchers who interviewed surviving participants decades later discovered something that should have been obvious: the workers in the special experimental groups were paid significantly better than their counterparts on the regular factory floor. Financial incentives, in other words, may have mattered quite a lot.

Harry Braverman, a Marxist sociologist, offered a more political interpretation. The Hawthorne studies, he argued, revealed not the power of observation but the power of class antagonism. Workers resisted management's attempts to extract more productivity from them. The informal cliques and group solidarity documented in the bank wiring experiment were forms of collective self-defense against exploitation. The whole enterprise of trying to use behavioral science to manipulate workers was, in Braverman's view, fundamentally misguided—not because it did not work, but because it failed to understand what work actually was.

The Effect That Will Not Die

Despite all these critiques, the Hawthorne effect persists. It appears in textbooks, training programs, and research design guidelines. Why?

Partly because it is useful. Whether or not the original studies proved what they claimed to prove, the underlying idea—that being observed changes behavior—is plausible enough that researchers feel compelled to account for it. Clinical trials now routinely consider what is sometimes called the "trial effect," recognizing that simply being part of a study may alter patient outcomes in ways that have nothing to do with the treatment being tested.

The effect also persists because it can explain so many different things. Research by Clark and Sugrue found that novelty effects in educational settings cause on average a thirty percent of a standard deviation rise in performance. This sounds technical, but it translates to roughly a fifty to sixty-three percent score improvement that decays over time—falling to a twenty percent improvement after eight weeks. Whether this is the "Hawthorne effect" or simply the excitement of something new is a matter of interpretation.

Researchers have developed various methods to work around the effect, assuming it exists. Some conduct field observations from a distance, using hidden cameras or one-way mirrors. Others use unobtrusive measures that subjects are unaware of. The goal is to capture behavior in its natural state, uncontaminated by the presence of observers.

A Secondary Effect

Recently, some researchers have proposed an intriguing extension of the concept. If the Hawthorne effect describes how observation changes the behavior of subjects, might there be a secondary observer effect that changes the behavior of researchers themselves?

When scientists work with secondary data—survey results, economic indicators, medical records—they bring their own biases and preferences to the analysis. They make seemingly innocuous choices about which data to include, which statistical methods to use, how to handle outliers and missing values. These choices may be invisible, even to the researchers themselves, but they can shape the results.

In this view, observation is never neutral. The observer and the observed are locked in a relationship that affects both. This may sound like a postmodern truism, but it has practical implications for how research should be conducted, reported, and interpreted.

What We Actually Learned

The Hawthorne experiments, whatever their methodological flaws, changed how people thought about work. Before Hawthorne, the dominant approach to industrial management was something called scientific management, or Taylorism, named after Frederick Winslow Taylor. This approach treated workers essentially as machines. You studied their movements, eliminated inefficiencies, and optimized output. The worker's internal experience—their feelings, their relationships, their sense of meaning—was irrelevant.

After Hawthorne, it became harder to sustain this view. The experiments suggested, however messily, that workers were social beings who responded to attention, recognition, and group membership. Management theorists began talking about motivation, morale, and workplace culture. The field of organizational behavior was born.

Whether this shift was based on solid science or on a glorified anecdote is, in some ways, beside the point. The Hawthorne effect became a useful fiction, a story that captured something true about human nature even if the experiments that supposedly proved it were deeply flawed.

We want to be seen. We perform differently when we know someone is watching. Whether that performance improves or worsens depends on context, on who is watching, on what we think they want. But the fundamental insight—that observation is never passive, that the act of measuring changes what is being measured—remains as relevant today as it was in a factory in Cicero nearly a century ago.

The workers at Hawthorne may not have taught us what the researchers thought they were teaching. But they taught us something. And that something, imperfect and contested and endlessly debatable, has shaped how we think about work, about research, and about what happens when we know someone is paying attention.