Complex morphological structures, such as skulls or limbs, are often composed of multiple morphological components (e.g., bones, sets of bones) that may evolve in a covaried manner with one another. Previous research has reached differing conclusions on the number of semi‐independent units, or modules, that exist in the evolution of structures and on the strength of the covariation, or integration, between these hypothesized modules. We focus on the avian skull as an example of a complex morphological structure for which highly variable conclusions have been reached in the numerous studies analyzing support for a range of simple to complex modularity hypotheses. We hypothesized that past discrepancies may stem from both the differing densities of data used to analyze support for modularity hypotheses and the differing taxonomic levels of study. To test these hypotheses, we applied a comparative method to 3D geometric morphometric data collected from the skulls of a diverse order of birds (the Charadriiformes) to test support for 11 distinct hypotheses of modular skull evolution. Across all Charadriiformes, our analyses suggested that charadriiform skull evolution has been characterized by the semi‐independent, but still correlated, evolution of the beak from the rest of the skull. When we adjusted the density of our morphometric data, this result held, but the strength of the signal varied substantially. Additionally, when we analyzed subgroups within the order in isolation, we found support for distinct hypotheses between subgroups. Taken together, these results suggest that differences in the methodology of past work (i.e., statistical method and data density) as well as clade‐specific dynamics may be the reasons past studies have reached varying conclusions.