The sweep net is a standard sampling method for adults of the western tarnished plant bug, Lygus hesperus Knight (Hemiptera: Miridae), in cotton (Gossypium spp.). However, factors that influence the relationship between true population levels and population estimates obtained using the sweep net are poorly documented. Improved understanding of these factors is needed for the development and application of refined treatment thresholds. Recent reports of significant among-sampler differences in sweep net-based population estimates of the adult tarnished plant bug, Lygus lineolaris (Palisot de Beauvois), seem to preclude meaningful comparisons of population estimates collected by different samplers. We used a mark—release—recapture method and the standard sweep net to evaluate among-sampler differences in population estimates of L. hesperus adults. Adult lygus, marked with fingernail polish to facilitate identification and prevent flight, were released into 10-m sample rows on the evening before 10-sweep samples were collected the following morning. The experimental design was a randomized complete block with three replications of three treatments (sampler). Separate experiments were conducted in two plantings each of Pima (Gossypium barbadense L.) and Acala (Gossypium hirsutum L.) cotton. Collections of marked bugs from each study were evaluated for effects of sampler, sample date, and their interaction. Although differences in lygus collections were observed among sample dates in some tests, no differences were detected in the population estimates by different samplers. These results demonstrate that the sweep net technique can be sufficiently standardized to allow direct comparison of population estimates obtained by different samplers.