Spatial distribution of boll injury caused by stink bugs to developing cotton (Gossypium hirsutum L.) bolls was studied in five commercial fields (≍22 ha each) in 2011 and 2012 to understand variability in boll injury dynamics within fields. Cotton bolls and stink bugs were sampled weekly from a georeferenced grid of sampling points (one sample per 0.40 ha) in each field, but no samples were taken within 30 m of field edges. The inverse distance weighted interpolation, variogram analysis, and Moran's I were used to describe spatial variability of boll damage within the fields. Boll injury was found to be spatially associated at distances ranging from ≍75 to 275 m with an average distance ≍150 m. An exponential variogram model was selected as the best fitting model to describe the spatial association in four of the five fields. Moran's I indicated that spatial association was significant in three of the five fields. The spread of boll injury from stink bugs was gradual in most fields and always exceeded the treatment threshold during the fourth or fifth week of bloom. Capture of stink bugs using a sweep net was inefficient, strongly suggesting that quantifying boll injury is a better sampling method and predictor of stink bug activity when sampling all but the edges of the field. These data suggest that scouts need to sample boll injury from sample locations separated by at least 150 m to assure independence in the central part of large fields. Second, future researchers who plan to use parametric statistical methods could use a 150-m grid, as opposed to a denser grid that would require greater time and effort.