Do you trust the measurements your bathroom scales provide? If not, you share a classic dilemma faced by scientists, which is whether or not their data gives an accurate picture of what they are trying to study. This is an especially sore point in the debate between climate scientists and their critics who differ, for example, over whether it’s OK to exclude temperature measurements from certain weather stations. One accusation is that scientists are just choosing the figures that support their arguments, a practice referred to as “cherry picking”.
Such disputes made me especially interested to see that Max Planck Institute for Developmental Biology scientist George Wang and his colleagues had specifically chosen a set of temperature measurements for their research. Together with Michael Dillon at the University of Wyoming and Raymond Huey at the University of Washington, Wang looked at how temperatures since 1960 would have affected metabolism of ectotherms – better known as “cold-blooded creatures”. They surprisingly found that, despite temperatures changing more slowly where they live, tropical species would be worse affected than those living in cold areas. Could choice of weather stations have influenced this research improperly? Not according to Wang.
“Scientists will exclude data for many reasons, but fundamentally we do it because blindly leaving data in an analysis can bias results,” he told Simple Climate. “In general, scientists in every field have to use judgement based on experience to detect and remove outliers. It is something scientists take very seriously, and it is an integral part of analysing data.” Read the rest of this entry »