As always with priceless support from @jagiellonskiuni.bsky.social @ioes-ju.bsky.social @unsw.edu.au , with amazing @itchyshin.bsky.social and other non-bsky-ers ;)

data structure (make sure your categorical data is sufficiently balanced and all categories are represented at the key intersection of predictors’ values); data complexity: collapse/merge levels if it’s reasonable to increase parameters identifiability.

unordered ordinal. Categorical data pose many challenges but they are as tractable with comparative methods as more well-behaved Gaussian models. The most important issues to keep in mind are: data availability (attempting to fit multinomial models with small datasets may lead to spurious results);

Follow us (incl @itchyshin.bsky.social and other co-authors not yet on Bsky ;)) for more useful methodological tips applicable to synthetic data. With support of @unsw.edu.au @jagiellonskiuni.bsky.social @ioes-ju.bsky.social @ncn.gov.pl @ualberta.bsky.social )

of heteroscedastic variance to meta-analytical heterogeneity can uncover hidden patterns in variability, including systematic trends in between-study variation. Stay tuned for our future papers where we will keep convincing y’all to adopting location-scale modelling as your default approach to data.

As always with invaluable support from @unsw.edu.au @jagiellonskiuni.bsky.social @ioes-ju.bsky.social @ncn.gov.pl

The models are notably more complex but fittable with modern Bayesian GLMM routines. Think of them as looking at mean and scale (usually variance) evolution along the tree, here in an example of two concurrent Brownian Motion dynamics driving diversification of both variance and means.

By extending the comparative model we can assay the evolutionary signal in evolvability itself and test a number of biologically interesting trait configurations (codivergence, contra-divergence, ceiling effects, phenotypic integrations and more!)

I’ve recently had my Swarovski cleaned, adjusted and resleeved free of charge. I request the front lens covers almost yearly (their rubber attachments break easily) - wondering when they’ll lose patience…