Keywords: Collaborative, Multi-Disciplinary, Principal Components Analysis, Mixed Models, Canonical Correlation
Academic statisticians can often collaborate with faculty in other fields as a co-author and statistical analyst. Doing so, however, can be fraught with challenges, as often the “standard” approach in a field is not the appropriate one, requiring the statistician to educate their collaborators on the necessary statistical techniques, and further to defend their choice of analysis against what is easy or common.
This talk presents several case studies based on collaborative projects performed with faculty at Le Moyne College: studying how English speakers perceive different Asian Englishes, the effect on climate change on bird migration, and the intersection between smell awareness and sexual orientation. In every case, the original analysis performed was ANOVA, whereas the final analyses performed include mixed models, principal components analysis, multinomial regression and canonical correlation. This talk will discuss the challenges in selecting the appropriate analysis to answer the research questions, and in explaining these analyses to collaborators who are not statistically trained.