No Evidence of Sampling Bias in a Comparison of Two Common Avian Capture Methods

Abstract

A common assumption underlying many biological studies is that the data we collect from a subset of individuals are representative of the broad population of interest. However, in studies of free-ranging animals, different capture methods might skew samples towards individuals with specific morphological, physiological, and/or behavioral traits. For example, in studies of free-ranging birds, bolder individuals might be more likely to enter certain traps commonly used in capture. Many studies require the capture of individuals for sampling, for instance for measurement of morphological or physiological traits, or for marking individuals. Because the capture methods that are employed often differ both within and between studies, there is potential for capture methods to introduce bias into these studies. We sought to explore this possible bias by comparing individual birds sampled using two very different, and commonly used capture techniques. We caught free-ranging black-capped chickadees (Poecile atricapillus) using Potter traps baited with seed and mist nets paired with an audio stimulus (chickadee mobbing calls) and determined sex, body condition, baseline and stress-induced glucocorticoid levels, behavioral response to a novel object, and behavioral response to a predator. We found no significant differences in any of these traits between individuals captured in these two methods. We also found no effect of capture method on the relationship between glucocorticoid levels and risk-taking behavior. We found similar variation of traits within both methods and conclude that selection of either of these two commonly used capture methods can similarly sample the population of interest.