Objective:​ There exists a need to better understand the biomechanical forces associated with head impacts in American football. The purpose of this study was to investigate whether or not differences in head accelerations existed between different player positions and different event types in collegiate football. We also sought to identify whether or not any associations existed between high-magnitude impacts and location of head impacts.

Methods:​ We conducted a prospective field study in which accelerometers were embedded in the football helmets of 72 collegiate football players. Linear accelerations of all head impacts sustained over the course of the 2005 and 2006 National Collegiate Athletic Association football seasons were collected. One-way analyses of variance and Χ² tests of association assessed positional, event type, and location of head impact differences.

Results:​ Football players consistently sustained head impacts between 21 and 23 g. Positional differences were identified within our sample. Impacts sustained during helmets-only practices were greater than those sustained in games or scrimmages. There was an association between position and high-magnitude impacts, as well as between high-magnitude impacts and location of head impact, with the likelihood of impacts to the top of the head much higher than those to the front, back, left, and right sides.

Conclusion:​ Less than 0.35% of impacts exceeding theoretical injury thresholds resulted in concussion. More injury data are required before any theoretical thresholds for injury can be confirmed. Coaches and sports medicine professionals should recognize that head impacts sustained in helmets-only practices are as severe as games or scrimmages;​ there seem to be no "light" days for football players.