First the link to this week’s complete list as HTML and as PDF.
***
Some of the scientists involved in Boyajian et al. speculate that all the natural explanations examined so far come up short in one way or another and we ought to consider an extraterrestrial civilisation at work here. Maybe so. Let us wait and see and them come up with some evidence. What they published so far is entirely valid.
***
I think the skeptic students cited in Dance are wrong. Inspiring role models in films need to be superheroes – think of Indiana Jones for archaeology, or about the engineers in Jules Verne’s novels, who identify minerals on sight, know their composition by heart and build a hydrodynamic power station from scratch including the smelting of the metals and drawing of wires. I know people like MacGyver, only they don’t come up with these solutions weekly but rather once a life and tell about them once a week. They are the ones who make you think, if he can do it, so can I.
***
Efferson et al. is one rare example of a study from the social sciences I can’t find any fault with and that yields a meaningful and important result.
***
I do not see how Hainmueller et al.’s answer relates to their question at all. The latter is, or at least should be, whether opening up naturalisation and making it easier to apply helps migrants to integrate. Their answer looks at rejection, but totally ignores the undoubtedly large effect that having applied and been personally rejected will always have on anybody compared to not being induced to apply in the first place. So their answer, if valid, is probably meaningless. Secondly I’m not at all convinced of it being valid in the first place. They explain their method in figure 1, lower left. If you take a small number of noisy data, separate them in the middle and calculate regressions for the left and the right half, those regressions will just about never meet in the middle however continuous the ensemble from which the date were drawn may be. Figure 1 upper left is a spurious exception.
***
Nature has three articles, Nuzzo, MacCoun & Perlmutter, and Silberzahn & Uhlmann, on bias and data blinding. I remember always feeling a bit uneasy about knowing when I ran the ignition I believed in and the one I didn’t in the run up to my diploma thesis, as edging an engine exactly onto the point of its lean limit can’t be done without a remaining rest of subjectivity. I’d have felt better not knowing.
***
I can’t see Sauchyn et al.’s conclusion in their data. According to their figure 5 the variability has been unchanged over the last 600 a and the most recent data lie near the bottom of the oscillations. There was a low decade 650 a ago and an even lower flow lasting 50 a 850 a ago and these can be expected to happen again some time, but there is nothing in the data to suggest it’s going to be in the next century or two.