BACKGROUND: Colorectal cancer screening is underused, and primary care clinicians are challenged to provide patient education within the constraints of busy practices. OBJECTIVE: To test the effect of an educational video, mailed to patients' homes before a physical examination, on performance of colorectal cancer screening, particularly sigmoidoscopy. DESIGN: Randomized, controlled trial. SETTING: 5 primary care practices in central Massachusetts. PARTICIPANTS: 938 patients age 50 to 74 years who were scheduled for an upcoming physical examination, had no personal history of colorectal cancer, and were eligible for lower-endoscopy screening according to current guidelines. INTERVENTION: Participants were randomly assigned to receive usual care (n = 488) or a video about colorectal cancer, the importance of early detection, and screening options (n = 450). MEASUREMENTS: Baseline and 6-month follow-up telephone assessments were conducted. A dependent variable classified screening since baseline as 1) sigmoidoscopy with or without other tests, 2) another test or test combination, or 3) no tests. RESULTS: Overall screening rates were the same in the intervention and control groups (55%). In regression modeling, intervention participants were nonsignificantly more likely to complete sigmoidoscopy alone or in combination with another test (odds ratio, 1.22 [95% CI, 0.88 to 1.70]). Intervention dose (viewing at least half of the video) was significantly related to receiving sigmoidoscopy with or without another test (odds ratio, 2.81 [CI, 1.85 to 4.26]). Recruitment records showed that at least 23% of people coming for periodic health assessments were currently screened by a lower-endoscopy procedure and therefore were not eligible. LIMITATIONS: The primary care sample studied consisted primarily of middle-class white persons who had high screening rates at baseline. The results may not be generalizable to other populations. The trial was conducted during a period of increased health insurance coverage for lower-endoscopy procedures and public media attention to colon cancer screening. CONCLUSIONS: A mailed video had no effect on the overall rate of colorectal cancer screening and only modestly improved sigmoidoscopy screening rates among patients in primary care practices.