Understanding the biases on the use of statistical methods
Mostly as part of my synthesis and collaborative work, I am interested in understanding methodological and statistical biases present in the empirical literature. The ultimate goal is to understand the nature of these problems, and, if possible, provide solutions to tackle them.
I have worked with Petri Niemelä and Alfredo Sanchez-Tojar in a project about "understanding terminology in animal personality research". We designed a survey to gather researchers' opinion, and then, performed a literature review to contrast published evidence with the collected opinions. Results show that there is a mismatch between the statistical definition of repeatability and what researchers report in the questionnaire and publications, pointing towards a potential problem in how researchers in animal behaviour understand repeatability (Sanchez-Tojar et al., 2022 Proc B). This finding is in line with a recent meta-analysis where we (Petri Niemelä, Kate Laskowski and I) found that only 3% of studies reported among-individual level estimates, although the publications were explicitly about individual differences in behaviour (Moiron et al., 2020 Ecol Lett).
|
Main collaborators:
Petri Niemelä, Helsinki University (Finnland)
Kate Laskowski, UC Davis (US)
Alfredo Sanchez-Tojar, Bielefeld University (Germany)