By 1997, almost half of all U.S. workers used computers on the job, and such workers generally received higher wages than non-users. However, on-the-job use was less common in nonmetro areas than in metro areas, and wages for nonmetro workers were generally lower. But is computer use instrumental in explaining the metro-nonmetro wage gap? A survey of 48,000 households found that wages for on-the-job computer users versus other workers showed a wage difference of 32 percent in 1997. After controlling for industry and occupation, worker education and skill level, and other worker and job characteristics, a 10-11 percent wage premium was associated with on-the-job use of computers. Taking into account the magnitudes of both the computer use wage premium and the metro-nonmetro gap in use, the computer effect accounts for only about 5 percent of the overall metro-nonmetro wage gap. The wage premium associated with computer use in rural areas is about 6 percent, about half the computer-use wage premium found in metro areas. As returns to computer use are smaller for rural workers, improving their computer literacy may contribute only slightly to reducing urban-rural wage inequality. However, computer literacy programs may improve the earnings of some racial and ethnic minorities, for whom the computer-use wage premium appears to be substantially larger. Appendices present data, methods, and eight detailed regression tables. (Contains 23 references.) (TD)