The application of transgenic (knockout) technology to the study of pain is rapidly expanding. Despite its power, this technique has several shortcomings that complicate the interpretation of the data obtained. Although compensation by other genes is a well recognized problem, issues related to the background genotype of the mutant mice are less well appreciated. This review describes these confounds as they apply to studies of pain and pain inhibition. We show that the 129 and C57BL/6 mouse strains, which provide the default genetic background on which null mutants are constructed, display significant and sometimes extreme phenotypic differences in many assays of nociception, hypersensitivity, and analgesia. Although problems related to the differential responsiveness of the two strains are minimized by placing knockouts onto "pure" 129 and/or C57BL/6 backgrounds, we also illustrate that neither of these strains are particularly representative of inbred mice in general. Procedures to reduce confounds and converging evidence must be used to accurately determine the functions of the targeted genes in pain-related phenomena.