What matters more for mixed effects models, the random effects structure (maximal vs. not), or the fitting procedure (frequentist vs. Bayesian)? A new blogpost exploring ManyBabies data. Spoiler: It's the random effects, stupid!
This article compares the two and outlines advantages and disadvantages of each: Meanwhile, had a recent blog post noting that the R package brms reliably converges in cases where lme4 fails, with no obvious costs: