As a researcher, you travel a lot. It’s usually a lot of fun, but… Umeå is so far from everything. It’s therefore no surprise that over 90% of the carbon dioxide emissions of Umeå University is caused by traveling employees. With a better world in mind, I therefore spend my last two days in a train, traveling from Umeå-Sundsval-Uppsala-Stockholm-Lund-Copenhagen-Hamburg-Ostnabruck-Hengelo-Zwolle-Leiden. From white to grey to brown to green and greener.
Of course I wanted to take advantage of the undisturbed time in the train and work on proposals, read literature and have great thoughts. So far so good, but after some hours, batteries got low and I realized that my train was not equipped with power outlets. Neither was the next one nor the next and the next, etc.
So I ended up staring out of the window, sticking to ‘having great thoughts’, and doubting about whether I need to take up holidays to compensate for the not being able to work or not. However, soon I found my brain wandering off. It simply can’t stop being a researcher, so it started analyzing the slowly changing landscape: Snow gone after Uppsala, the fist daffodils around Copenhagen and ice-cream-eaters in Hengelo. As the weather changed from snow to rain to sunshine, my hypothesis-addicted brain quickly found an explanation. More difficult to explain was: the stones (flyttblock) being rounder in Lund than in Umeå, and the grass being greener in Amsterdam than around Lund. Because I had nothing better to entertain my brain, I found myself playing the game of thought experiments.
There are three basic designs that underlie most experiments. The first one is based on comparing groups. The second approach extends on that, and looks for relations over a gradient. Instead of comparing two averages (are stones really more round in Lund than in Umeå), one can than draw lines and see trends (As I did along the climate gradient). However, both these approaches encounter problems when your locations differ in too many things. The classic example of this pitfall is that you can find a strong relation between number of crime victims and numbers of hamburgers eaten. Of course there is no causal effect there, but both hamburger consumption and number of crime victims may both be driven by city size. To overcome this ’co-founding’ problem, you can do a ‘manipulative experiment’, and for instance add nutrients to see if it affects the greenness of the grass, or that temperature also is important there.
I assume that many researchers play this game of thought experiments and if you find yourself sitting in a train with low batteries and nothing better to do, you are welcome to contribute. Overall, I enjoyed my long, long journey south, as it all trains ran smoothly in time and I met many interesting people. But next time, I’ll be better prepared so I can keep my brain entertained with work rather than with stones, the weather and grass.