Like most of the TARSKs I will be reporting, the common version of the story is much more powerful than the nuanced truth, so I will start with the way it is normally told.
The Hawthorne Effect describes a phenomenon that when we tell people we are studying them or we tell them we are making changes for them they modify their behaviour.
The term dates back to experiments conducted between 1924 and 1932 at the Hawthorne Works factory in Chicago. The experimenters found that when they increased lighting (for example) productivity went up, but when they tried an experiment where they reduced the lighting, productivity went up! The research found that almost any change that was tried resulted in increased productivity amongst the people being measured. A finding that has implications for any pre-post testing that market and social researchers might be conducting.
The facts themselves are a bit more complicated as the experiments conducted in 1924 to 1932 were not well structured by modern standards, subjects were allowed to self-select under some circumstances and subjects were sometimes replaced during an experiment.
However, the Hawthorne Effect is a useful reminder to researchers that when we measure something we can sometimes cause the results we think we are observing.
I had an interesting example of this, over twenty years ago, when I was working with a UK retailer who was rolling out a new store design. The background story was that the chain had decided it needed a radical change and it was going to close about half of its stores and redesign the other half.
The research was a pre-post and control test. We evaluated customer responses at a number of stores who were scheduled to be amongst the first to be redesigned, and amongst some stores that were not amongst the first to be scheduled for closure, but were also not amongst the first to be redesigned.
After a few months two interesting results emerged:
1) The stores which had been redesigned were performing worse in terms of customer satisfaction than before the change. Now, it is not unusual for a redesign to reduce satisfaction, many people dislike change, but these changes were so big that the conclusion was that there was a real problem with the redesign.
2) Many of the stores which had not been redesigned were seeing improvements in their scores (and in sales)!
When the client looked into the improved scores (and sales) in the unimproved stores they found that some of the store managers had realised that being not redesigned and being measured meant they were being considered for closure, therefore they had been pulling out all the stops to increase sales and to attract more customers, for example: tidying up, offering advice, asking staff to encourage friends to visit, etc.