Líf- og umhverfisvísindastofnun

Combatting dogma in biostatistics

Dagsetning: 
Thursday, February 14, 2019 -
12:30 to 13:15
Nánari staðsetning: 
Stofa 131

Michael B. Morrissey, Royal Society University Research Fellow and Reader in Evolutionary Biology, University of St Andrews

 

Title: “Combatting dogma in biostatistics”

 

Michael mostly conducts theoretical and empirical evolutionary quantitative genetic studies. He also has an interest in how basic statistics are used across ecology and evolution. This talk will seek mostly to clarify what multiple regression is. It may seem that this relatively simple, and extremely widely-used, method would be well understood by biologists and writers of biostatistics textbooks. Michael will argue that this is not the case, and that confusion about this analysis has led to very complex solutions to non-existent problems. Taking a step back can make things very simple. After talking about multiple regression, Michael is likely to rant a little bit about the validity of linear models vs. generalised linear models, and ideas about variable selection.

See a recent paper.

Morrissey, MB & Ruxton, GD 2018, 'Multiple regressions: the meaning of multiple regression and the non-problem of collinearity' Philosophy, Theory and Practice in Biology, vol. 10, 3. DOI: 10.3998/ptpbio.16039257.0010.003
 
Friday biology seminar series - spring schedule.
Þú ert að nota: brimir.rhi.hi.is