There’s quite a song and dance made in science classes at school and university to the tune that the scientific process is superior to non-scientific processes of inference and belief because scientists are sceptical by nature, keen even to challenge their own theories and methodologies in an ever-going quest for the truth, and open always to the probing questions of their peers. And, moreover, if they weren’t then the system is self-correcting because their peers could easily show up any poor or dubious work simply by repeating the claimed experiment for themselves. This is of course a complete load of nonsense. Scientists are prone to exactly the same conceits as the rest of us, and would rather agree that the new Star Wars films are better than the originals than be proved wrong; as for helping you to do that for them … forget about it! But repeating the work of most modern experiments and analyses is so time-consuming that it’s completely infeasible to hold any result to accountability without this cooperation.
Last month I noticed a paper by Philipp Lang and the CANDELS team on bulge-disk decomposition of high redshift galaxies. Having worked on bulge-disk decomposition of local galaxies I know first hand how much of a dark art this game is. In particular, i have pointed out the lack of believable fits with high bulge-to-total ratios: if you plot histograms of the inferred distribution of disk inclination angles for low B/T systems it usually looks much like what we’d expect for a population of randomly oriented disks, but if we do the same for high B/T systems we get a distribution massively biased towards face on disks. The most likely explanation being that the high B/T “disks” are just artifacts of the model fitting process not representing real structures of ordered motion.
Having noticed that Philipps’ paper had a lot of very high B/T fits i emailed him at the start of Feb to ask if he’d looked at this issue. After two weeks with no reply i emailed again. After another two weeks i decided to email his second coauthor, Stijn Wyuts. Agan i waited over a week with no reply. So on Thursday i tried the third coauthor, Rachel Sommerville, and i guess I’ll just have to keep on going down the list until someone can be bothered to get back to me.
Yeah, modern science.