One of the possible adverse effects of transgenic insecticidal crops is the unintended decline in the abundance of nontarget arthropods. Field trials designed to evaluate potential nontarget effects can be more complex than expected because decisions to conduct field trials and the selection of taxa to include are not always guided by the results of laboratory tests. Also, recent studies emphasize the potential for indirect effects (adverse impacts to nontarget arthropods without feeding directly on plant tissues), which are difficult to predict because of interactions among nontarget arthropods, target pests, and transgenic crops. As a consequence, field studies may attempt to monitor expansive lists of arthropod taxa, making the design of such broad studies more difficult and reducing the likelihood of detecting any negative effects that might be present. To improve the taxonomic focus and statistical rigor of future studies, existing field data and corresponding power analysis may provide useful guidance. Analysis of control data from several nontarget field trials using repeated-measures designs suggests that while detection of small effects may require considerable increases in replication, there are taxa from different ecological roles that are sampled effectively using standard methods. The use of statistical power to guide selection of taxa for nontarget trials reflects scientists’ inability to predict the complex interactions among arthropod taxa, particularly when laboratory trials fail to provide guidance on which groups are more likely to be affected. However, scientists still may exercise judgment, including taxa that are not included in or supported by power analyses.