In this prospective, multiinstitutional analysis of medial collateral ligament sprains in college football players, we categorized 987 previously uninjured study subjects according to frequency of wearing preventive knee braces, studied the patterns by which 47 of 100 injuries occurred to unbraced knees, and identified several extrinsic, sport-specific risk factors shared for both braced and unbraced knees. The attendance, brace wear choice, position, string, and session of each participant were recorded daily; medial collateral ligament sprains were reported whenever tissue damage was confirmed. Both the likelihood of wearing braces and risk of injury without them was highly dependent on session (games/practices), position group (line, linebacker/tight end, skill), and string group (players/nonplayers). Subjects wearing braces often faced a high injury risk to their unbraced knees, a finding compatible with the opinion that braces were a necessary evil, best worn when concern over danger of injury outweighed desire for speed and agility. It is concluded that to avoid misinterpretations due to the confounding influence of brace wear selection bias, accurate investigation of daily brace wear patterns is required. Then, before considing the impact of preventive knee braces, a repartitioning of the data base is essential to assure that only similar groups will be compared.