Scientists are supposed to be trained to examine and make conclusions based on evidence, however, this is widely ignored when it comes to themselves rather than their object of study. A 2016 poll on the Nature website showed that about 70% of academics work more than 50 hours per week. The lack of a work-life balance was chosen as the biggest challenge for early career scientists by 19% of respondents and almost two thirds have considered leaving research. This is similar to previous results in a 2011 Nature poll where 65% of post-docs said that they worked over 50 hours a week.
In contrast to this behaviour, current research does not support the idea that longer working hours are more productive. One study, based on factory workers during WWI, showed that productivity is proportional to hours worked only to a certain point. Above 48 hours worked per week, productivity per hour sharply decreases. Because of the decreased productivity per hour over longer work weeks it can be that working fewer hours will cause an increase in productivity.
It is alarming to consider that the behaviour of a majority of scientists is in direct conflict with the evidence of the effect of working hours on productivity. Perhaps more alarming, though it will not be considered here, is that longer working hours are also detrimental to workers’ health. For example, longer working hours (defined as 55 hours or more per week) is associated with an increased risk coronary heart disease and stroke and women who worked over 60 hours per week had triple the risk of “diabetes, cancer, heart trouble and arthritis.” Also not considered here is the, frankly insulting, notion that one’s identity should be derived solely from one’s work.
Many will no doubt object to the conclusions presented here as being irrelevant for intellectual work or somehow limited to repetitive tasks. But these results also stem logically from the fact that people become tired over time, leading them to work slower and make more mistakes. It shouldn’t be surprising that longer hours will have negative effects. Again, the research supports this view.
Studies with medical interns have shown that reducing the length of their shifts leads to a reduction in medical errors. Performance of Dutch schoolchildren on standardised tests decreases for every hour later in the day when the test is taken. There is a small recovery in performance after a 20 – 30 minute break. Even the possibility of a successful parole hearing decreases throughout the day, also rising again the judge has a meal and a break. This should hopefully satisfy critics that the effect of longer hours leading to lower productivity is not limited to repetitive factory work but also affects intellectual abilities.
Furthermore, on a global scale, there is no relationship between the average number of hours worked and GDP. Greece, for example, has one of the longest average working hours in the OECD countries while Germany has one of the shortest average working hours but Germany’s productivity is much higher than that of Greece.
So if, as the evidence suggests, working longer hours is not actually benefiting anyone, why do we have it? One likely reason is that people naïvely assume that productivity remains constant and so increased working hours will lead to increased productivity. However, I think there might be a more troubling explanation.
In The Greatest Show on Earth, Richard Dawkins described a forest of trees. Trees need to grow taller than their competitors to reach the sunlight. If they do not grow tall, they will not have the requirements they need to survive. But, on the other hand, if the trees all agreed not to grow beyond a certain low height, then they would get just as much sunlight as if they all grew tall but they would not have to waste energy in unnecessary growth. The problem is two-fold; the trees cannot agree to a set height and any tree which grows taller than the others has an advantage.
Similarly, at work, we would all benefit from shorter working hours but anyone working longer has a benefit. It is not a benefit in production, as we saw earlier with the poll answers and the study on productivity, average working hours in academia are far beyond where they could lead to any increase in productivity, but the benefit is in signalling. Working longer hours is a signal of motivation and dedication to one’s work and supervisor, even if the actual effect of those hours is negative. Working long hours ceases to be about doing more but about signalling more and improving one’s social standing, just like how a peacock’s tail signals his fitness.
This is illustrated really nicely by a study done in 2005 on the working hours of consultants in one firm. (Harvard Business Review article here and an interview with the author here.) What he found was a culture where someone working 60 – 80 hours a week was seen as the ideal worker. Many workers did not want to work like that though and asked for ways to reduce their working time so they could meet family obligations. These people were marginalised in the firm and didn’t get promotions. Other workers found ways to cut down their hours while maintaining the appearance of long hours. These workers were not penalised and were often seen as an example to others.
It’s unsurprising that those who reduced their hours while maintaining the illusion of working long hours did not have a reduction in work quality as they were still working around 50 hours per week. So we already know that the 30 hours that some were working to reach 80 hours a week were, essentially, wasted. This clearly illustrates that longer working hours do not make a noticeable performance boost but that the signalling of the longer hours does create a social benefit. That is also the reason why it is important to enforce laws about working hours, it can help prevent the negative repercussions of losing that signalling and prevent the wasted energy that we saw with the trees.
I’m not the first one to note that this is a problem. It has been brought up in the scientific literature itself with the 2013 article in BioScience entitled How to Exploit Postdocs. Before that there was the the more directly titled Exploitation of Junior Scientists Must End published in Nature in 1999. That these issues are still a problem 18 years later speaks to a massive failure in the scientific and academic fields. There has been a failure of government to enforce labour laws in research, a failure of research institutions to follow the laws, a failure of supervisors to properly mentor their charges and a failure of those that are meant to represent PhDs and postdocs to correct institutional behaviour and a failure of individuals to enforce their contracts.
Changing the way things are done in research and the scientific field will not be easy. Some people have made it to where they are with long working hours and they believe that is the secret to their success. Changing people’s beliefs and habits is difficult. But if we can bring about such changes, it will benefit everyone. It will improve the quality of work that is produced, benefiting the institutes and funding agencies, and it will improve life satisfaction of researchers, directly benefiting them but almost certainly feeding back to improving their work. As scientists we should be basing our decisions on evidence. In this case, the available evidence shows that longer working hours as a general principle is bad.