One of the more interesting events in the annals of management psychology involves a serious of studies performed in the late 1920s and early 1930s in a GE plant in Chicago. The workers were polled on various preferences, and working conditions were altered systematically to see what effects there would be on productivity. It turns out that with every variable studied -- pay, shop-floor temperature, length of coffee breaks, etc. -- productivity went up no matter the direction the variable was altered. The only thing that seemed to matter for productivity was the fact that the workers knew they were being studied.
One way to interpret thisis that workers just want to be cared about, that they are like infant children flattered by any kind of attention, and unable, ultimately, to tell the difference between good and bad attention. This seems to be how management theorists have interpreted it, and it leads to all sorts of inadvertant worker punishments like meetings where guys in suits explain to workers how they've tinkered with the corporate mission statement, or special breakfasts with management, where workers are forced to come in early and eat stale bagelettes and be told how much they are appreciated. Management loves this because it makes their self-interested gestures of largesse seem meaningful, signs that they really care, instead of being simply the short-sighted tokens that they really are. Management would love it if they could make up for deficits in fair wage with a surfeit of niddling attention; then they could feel there was something generous about their surveillance. Because that seems the more likely significance of the Hawthorne effect -- when workers know they are being "studied" -- that is, watched -- they are likely to be more productive. Nicey-nice meetings and such are displays of the corporation's soft power, allowing them to monitor without seeming to threaten.
The same results also prove, at Braverman pointed out in Labor and Monopoly Capital, that work performance and pre-existing worker skills aren't necessarily correlated -- workers will adjust their performance in response to the class antagonisms at work in the factory. That is, workers should always regard the demands of management with skepticism, and management's demands are only carried through when there is a palpable threat behind them. This should be remembered anytime Wal-Mart workers "vote" to not form a union.