[yummyfajitas]

I've read the research. It doesn't back up the claims made by the OP.

Last time you you were trying to claim 40-60 hours was an 'optimal' working time for manual labour and linked research yourself which showed that it wasn't.

The research linked to shows that weekly product (which is not the same as hourly productivity) can be maximized by anything from 40-60 hours, depending on the nature of the work and the duration of the project.

For example, look at Figure 16 which has numbers:

    100% productivity x 40 hours = 40
    90% productivity x 50 hours = 45
    83% productivity x 60 hours = 49.8
    69% productivity x 70 hours = 48.3

Output is maximized by 60 hour weeks.

Figure 6 shows similar results, based on eyeballing the 5 and 6 day weeks (10 hours/day each) as being about 83% productivity. (Data is in a bar chart, numbers not listed.)

http://www.danzpage.com/Construction-Management-Resources/Ca...

In any case, any paper discussing averages (i.e., every source I've ever found) simply does not address the question of whether the author of the piece can productively work 60 hour weeks. That's like saying "the mean American wage is $50k/year or so, therefore it's impossible for a software engineer to earn $100k".

[hox]

What about figure 9 that shows sustained 60+ hour work weeks drop productivity to 45% that of a 40 hour work week after week 4?

Additionally, you are never going to find research that discussed the magical unicorns you describe and not averages. That's not how research works.

[yummyfajitas]

What about figure 9 that shows sustained 60+ hour work weeks drop productivity to 45% that of a 40 hour work week after week 4?

What about it? My only claim is that the answer is complicated. I'm not claiming that 60 hours/week is a magical long term answer perfect for everyone everywhere in every industry. It's the 40 hours/week people who are making that claim.

Additionally, you are never going to find research that discussed the magical unicorns you describe and not averages.

Yes, it clearly requires a wizard riding a narwhal to compute quintiles.

Without the quintiles/ventiles/etc it's impossible to conclude that an individual can't work 60+ hour weeks. That's a simple fact about distributions. If the 100+ years of research don't provide that, it simply means that more work needs to be done because we don't know the answer.

[mattmanser]

The quote accompanying those figures:

the majority based on 10-hour workdays and an overtime schedule of four consecutive weeks

FOUR WEEKS. Did you read that bit? Productivity continues to drop after 4 weeks (with some parts of it suggesting the productivity drop stops around week 16ish).

You're a lost cause.

[shrughes]

You're the one claiming there's a huge body of research disproving his claims. Your link upthread points to stuff that supports his argument and other stuff that counts as weak evidence against. Nowhere is there anything remotely like disproof. You can also find evidence (e.g. in http://cmdept.unl.edu/drb/Reading/overtime2.htm from the link above) that how you structure your work and rest periods can improve your production. After all, somebody doing office work 60 hours a week with some regular exercise is probably going to outperform somebody like me, who's fat and doesn't exercise. I certainly outperform myself on days that I walk to the office instead of driving. There's clearly some variance involved in individual employees' age, health, height, diet, and attitude, so it should be obvious that you can't just talk about averages over the construction industry and apply them to programming, and to all programmers, with the degree of certainty that you hold.

[yummyfajitas]

Yes, most of this data is short term because it's for the construction industry which works on projects. Depending on how long the recovery period is, it could easily be the case that the long term maxima is achieved by long hours + short recovery periods. (I.e., 7 65 hour weeks + 1 30 hour week.)

And again, basically all the research either you or I have found is in construction. Construction != Programming.