"Research done by the Free University of Bozen-Bolzanohas in Italy concluded that happy software developers are better at solving analytical problems. 'Even simple and short activities', the researchers note, 'may impact the affective states of software developers.'
Many large software companies have been providing various perks to developers, hoping that they will become more productive. Based on a study of 42 students from the Faculty of Computer Science, this research seems to validate that practice. Its findings suggest that 'the happiest software developers are more productive in analytical problem solving performance.' This is in contradiction to previous studies, most of which concluding that negative affective states foster analytic problem-solving performance.
Physical (assembly line, etc) productivity has been extensively studied, going back well over a hundred years, here's a bibliography starting with Taylor,
http://www.oxfordbibliographies.com/view/document/ obo-9780199846740/obo-9780199846740-0027.xml [oxfordbibliographies.com]
I've seen a confounding result too (sorry, can't remember the reference)--when workers were studied (at least in the early days), they responded to the additional attention by increasing their productivity. This led to a number of false positives in terms of modifying the work environment for higher efficiency. Once the study and the researchers left the factory with an updated production process, productivity returned to the original level.
At a later date, workers were wise to researchers with clipboard and stopwatch. They might sandbag (slow down) to give a false baseline for the study.
Are there any parallels to the present time, with respect to mental/analytical work?
You make a good point. Motivation is one factor that's hard to test and control, and goes beyond environmental/procedural considerations.
This is a whole other kettle of fish, but I think progressive mutualisation is probably a good way to motivate people (ie a small proportion of an employee's salary is paid in equity. Over time it reaches some reasonable maximum level of ownership per-employee. So over time more productive work means more profit). If the balance was right it would probably work out more profitable for the original investors as well, due to increased productivity and loyalty of the workers.
I think that's the Hawthorne Effect [wikipedia.org], where workers improve their performance in response to any change in their environment (positive or negative), simply because they know they're being studied.
A sort of Heisenberg principle for social experiments...
(Though it seems that some of Hawthorne's initial results may have had other explanations.)
Bingo, this section of the Hawthorne entry was particularly interesting, Interpretation and Criticism [wikipedia.org].