62 children, between the ages of 9 and 18 months old, were observed in an instrumented walker to measure the peak horizontal pull forces. These pull forces were later used to evaluate an infant walker with a braking system that would stop on the top step of the stairs before falling down the stairs. This brake system is activated when part of the walker crosses over the edge of the top step. Using the range of horizontal pull forces generated by the 62 children, the horizontal brake system for walkers would not always prevent the walker from falling down the stairs. Four floor surfaces were compared: carpet, vinyl, glossy wood, and unfinished wood. The walker brake system did not always stop the walkers on these floor surfaces. Using the measured weights and horizontal forces of the 21 nonwalking children between 9 and 13 months old who represent children who typically fall down the stairs in a walker, a simulation procedure was completed to represent the worse possible force condition, the peak horizontal force, for each of the 21 children. During this simulation, brakes would have failed all the time for 18 of the 21 children, and at least half the time for the remaining 3 children. These brake systems may provide false security to parents who use these walkers, since there are no published standards regarding the performance of brake systems for infant walkers as a safety device.