Reports in the scientific literature and lay press have suggested that computer users are at increased risk of upper extremity musculoskeletal disorders (MSDs). Early studies often found elevated rates of MSD outcomes among keyboard users when compared to non-users. Attention soon focused on specific aspects of keyboard work that might be responsible for the observed rate increase. In this review, the epidemiological evidence examining associations between MSD outcomes and computer user posture and keyboard use intensity (hours of computer use per day or per week) are examined. Results of epidemiological studies of posture and MSD outcomes have not been entirely consistent. Reasons for the inconsistency in results include cross-sectional study design (with possible failure to assure that measured exposure preceded health effect), imprecision of posture measures used, and difficulties involved in analyzing multiple related variables. Despite the inconsistencies, it appears from the literature that posture is an independent risk factor of modest magnitude for MSDs among computer users. It appears that lowering the height of the keyboard to or below the height of the elbow and resting the arms on the desk surface or chair armrests is associated with reduced risk of neck and shoulder MSDs. Results of epidemiological studies examining computer use (hours keying per day or per week) are more consistent than those examining posture, although some inconsistency is observed. Reasons for the inconsistency include possible selective survival bias resulting from cross-sectional study design, differences in exposure categorization, and possible interaction with other exposure variables. Overall, the literature shows that daily or weekly hours of computer use is more consistently associated with hand and arm MSDs than with neck and shoulder MSDs.