Ricardo,
The 'best' approach depends on what you mean by 'the effect of size' in your analysis. There are size effects and there are allometric effects of size on shape; the latter being the covariation of shape with size.
Strictly speaking, in Conaway and Adams, we removed the effect of size only, not really allometry. This is accomplished using GPA where size is standardized. Then, we compared the size-free data to size + shape data by adding size back in for an analysis of
essentially form.
The approach in Natale & Slater 2022 accounted for size in the GPA, but also accounted for the allometric effects of size via regression. Using residuals from that regression accounts for both components, and that is the correct approach if one wishes to remove
size-based changes in shape from the data.
Dean
--
Dr. Dean C. Adams
Distinguished Professor
Director, Ecology and Evolutionary Biology Graduate Program
Department of Ecology, Evolution, and Organismal Biology
Iowa State University