What Is the Hawthorne Effect? The Famous Factory Data Nobody Checked for 80 Years
The short version: the Hawthorne effect is the idea that people change their behaviour because they know they're being studied. It's named after lighting experiments at the Hawthorne Works factory near Chicago in 1924-1927, which supposedly showed output rising with every change of light, even dimming. The data were never properly analysed and were thought destroyed. When economists found them, the textbook pattern turned out to be fiction: output also rose in rooms where nothing changed, dips matched the summer season, and the Monday jumps happened on every Monday, change or not. There are faint signs of a small effect, nothing more.
HR has repainted the third floor a colour called Productive Sage. Output in March was higher than in December. HR has announced a success and explained it with science: the Hawthorne effect. People work harder when they feel noticed, whatever you change, as the famous factory proved.
Kevin's boss nodded. He knows the story. Everybody knows the story. Brighter lights, more work; dimmer lights, more work. It's in every management course.
Nobody in the room had seen the data. This turns out to be a long tradition.
The story everyone tells
From 1924 to 1927, engineers ran experiments at Western Electric's Hawthorne Works near Chicago, a plant of about 35,000 people making telephone equipment. The question was narrow: does better lighting make workers more productive? In three departments, women who assembled relays and wound coils had their lighting changed while their output was counted.
The legend says that every change raised output, even making the room darker, so the light didn't matter and the attention did. Later experiments at the same plant fed the same story, and the Hawthorne effect entered every textbook: people behave differently when they know they're being studied.
There was one problem. The lighting results were never written up, and the data were believed destroyed. For around 80 years, the story was retold from a few paragraphs in a trade journal.
What the data actually showed
In 2009, two economists traced the original records to archives in Milwaukee and Boston and analysed them for the first time. Their verdict on the textbook pattern: entirely fictional.
- The first year looks good. Output in all three rooms rose to about 10% above the starting level, while the plant as a whole improved by about 1.4% a year.
- But untouched rooms rose too. In the second round, only room 1 had its lighting changed. Room 2, and to a lesser degree room 3, also produced more, with nothing being done to them.
- The dips were summer. Output fell whenever the experiments paused, which looked like workers missing the attention. But the pauses happened in summer, and output fell in all rooms, experiment or not.
- The Mondays. Lights were always rearranged on Sunday, when the plant was closed, so every change started on a Monday. Output did jump on those Mondays, but by the same amount on Mondays with no change. Saturdays were slow and Mondays picked up, every week.
What's left after all that is faint: output was about 3-4% higher while experiments were running than before they started, and workers seemed to react a little more to light changed by the experimenters than to natural daylight. Possibly a small effect. Not a law of nature.
Why nobody noticed
A before-and-after comparison with no control group credits the experiment with everything else that changed at the same time: the season, the day of the week, workers getting better with practice. The Hawthorne data had all three. The rooms that weren't being experimented on were the control group, and they told the real story, but nobody looked, because the story was too good to check.
Statistics for noobs: the calendar is a variable
Many numbers have a rhythm: weekdays, months, seasons, holidays. If you make a change and measure before and after, you also measure the rhythm. Compare with the same period last year, or with a group that didn't get the change, before you credit the paint.
The Hawthorne effect itself may be real in some settings, such as people tidying their habits while being observed. But how big it is depends on the situation, and the study that gave it its name is very weak evidence for it.
Where it costs you money
"We switched to the new strategy in January and it's up since" is the same mistake. January to March might be up in most years. Before crediting a change, compare it with what the market did over the same weeks and with a random choice made on the same day. That's why every test on this site has a comparison group.
The moral
Kevin pulled up last year's numbers. Output rose from December to March last year too, with the old paint. HR considered this and concluded that last year's rise was also a Hawthorne effect, caused by the announcement that the floor would be repainted. The story is now complete and cannot be disproved by any data, which is how you know it has become management theory.
What this can't tell you
- The data were rebuilt from graphs. The original records survive as charts, which the economists turned back into numbers.
- Three rooms. A small study; a subtle effect could be hiding in the noise. Not finding the famous pattern is not proof that attention never matters.
- Only the lighting tests. The later Hawthorne experiments (rest breaks, working hours) are a separate question, and their data have their own problems.
HR, the paint and Kevin's boss are an illustration. The figures come from the published reanalysis of the original illumination data (National Bureau of Economic Research working paper, 2009; American Economic Journal: Applied Economics, 2011). Educational content, not financial advice.