 Solution

Extinction selectivity means that the probability of extinction depends systematically on biological or ecological traits, rather than being equal across all species. An empirical association needs a specified event, group and taxonomic level; a trait protective in ordinary background turnover may not protect during every mass extinction.

Geographic range: at the end-Cretaceous crisis, geographically widespread marine bivalve genera had lower extinction intensity than narrowly distributed genera. This makes range breadth a demonstrated selective trait for that dataset. The same analysis found no clear selectivity by body size, so that result should not be generalized to every trait or event.^[ref]

Body size: large-bodied mammals suffered disproportionately during late-Pleistocene and later Quaternary losses, including the loss of many giant herbivores and carnivores. A global analysis over the last $125000$ years found strong size-selective losses across continents and trophic modes. Large size can correlate with slow reproduction, high resource demand and exposure to human hunting, but the observed size pattern and its possible mechanisms should be distinguished.^[ref]

Sexual dimorphism and reproductive investment: a study of fossil cytheroid ostracods found much higher estimated extinction rates in species with stronger male–female carapace dimorphism, interpreted as greater male investment in reproduction. This supplies a third empirical trait association; the observed relationship is not a universal claim that sexual selection is always harmful.^[ref]

The examples concern different time intervals and levels, including genera in the first example and species in the third. They demonstrate extinction selectivity without assuming one set of trait effects is universal across all crises.