Abstract

Evidence is presented to show that on-product warnings (OPWs) are not seen or are ignored. The author maintains that there is no sound empirical evidence that OPWs influence behavior. He describes evidence, however, that they might be dangerous. It is suggested that OPWs are being used too pervasively. Use of warning labels should be reserved for only those products where it can reasonably be determined that labels are an appropriate method for communication. The author provides a qualitative protocol for making that determination. The protocol includes consideration of six elements of a communication model: sender, receiver, message, medium/channel, objective, and environment.