To be clear, my understanding of the Hawthorne effect, to which Luther referenced, is very different. Yes, the workers grew more productive with each environment change. But, that was supposedly because they were being monitored and NOT because there was an environment change.
This is one of the big bugbears of all systematic productivity schemes. We use CMMI at work and that mandates, at the highest levels, that you only change process if you can measure the improvement and that the best way to do this is to pilot a change, see if there is an improvement, then, if appropriate, roll it out across the organisation. So we have a situation where a keen good group of engineers who know they are the pilot try something and, surprise surprise, there is an improvement. So we roll the change out across the whole organisation. No one seems to think of trying it on the project that's got low morale, disaffected engineers and so on...
Steve