Commercial rice, Oryza sativa L., fields in southeastern Texas were sampled during 2003 and 2004, and visual samples were compared with sweep net samples. Fields were sampled at different stages of panicle development, times of day, and by different operators. Significant differences were found between perimeter and within field sweep net samples, indicating that samples taken 9 m from the field margin overestimate within field Oebalus pugnax (F.) (Hemiptera: Pentatomidae) populations. Time of day did not significantly affect the number of O. pugnax caught with the sweep net; however, there was a trend to capture more insects during morning than afternoon. For all sampling methods evaluated during this study, O. pugnax was found to have an aggregated spatial pattern at most densities. When comparing sweep net with visual sampling methods, one sweep of the “long stick” and two sweeps of the “sweep stick” correlated well with the sweep net (r² = 0.639 and r² = 0.815, respectively). This relationship was not affected by time of day of sampling, stage of panicle development, type of planting or operator. Relative cost-reliability, which incorporates probability of adoption, indicates the visual methods are more cost-reliable than the sweep net for sampling O. pugnax.