The performance of a variety of areal geophone arrays was evaluated using seismic data recorded on a dense receiver grid in a walkaway survey conducted in the Permian Basin of west Texas. Surface waves, trapped waves, and scattered energy have long been recognized as a significant noise problem in this area. Arrays were formed by extracting sets of traces from the main data set and stacking them to produce individual traces of a receiver gather. We calculated semblance of each receiver gather to evaluate array performance. High values of semblance indicate that an array effectively removes surface waves while preserving reflections. Differences in data quality associated with variations in geophone-array design are often subtle and difficult to discern through simple inspection of field records. By calculating frequency-dependent semblance, we were able to detect and quantify differences in array performance.