The influence of position, highlighting, and imbedding on warning effectiveness

Warnings on consumer products are typically the only devices used to elicit safe product use. Manufacturers have a legal duty to warn about predictable product dangers. and currently the legal system heavily emphasizes the importance of such written product warnings in determining product liability issues. The legal system's reliance on such warnings rests on the assumption that these warnings are effective in influencing user behavior. An e1periment was conducted to provide behavioral data pertaining to the effectiveness of such warnings; by investigating the effects of position. highlighting, and imbedding warnings on detection. recall and compliance with warnings through actual user behavior. Three warning positions (top. middle. and bottom) were varied on the label of a consumer product. as were two warning highlighting conditions (inverted and not inverted) and two warning imbeddedness conditions. The imbeddeding consisted of begining the Warning setion with either the critical warning information (the unimbedded condition) or a superfluous warning statement followed by the critical information (the imbedded condition). An additional "control" group condition vas used in which the critical safety information vas included in the Directions section on the label. with no formal Warning section appearing at all. In the experiment each of 1~ subjects was given a product and a situation in which they had to actually use it. Through direct user observation the influence of the warning on user behavior was directly measured and follow-up questions provided insights into the underlying factors involved. The follow-up questions were specifically focused to determine if the user 1) noticed the warning. 2) then read it. and 3) followed it. The purpose being to find out where along this three step process people dropped out (if at all) and more importantly, if they did drop out, why. The results showed a consistent decline over all experimental conditions in terms of the number of people who first noticed, read and then complied with the warning. Surprisingly, the only variable to have a differential effect on warning compliance was that of imbedding. Imbedding the critical warning information significantly reduced warning compliance, as subjects stopped reading the warning prior to reaching the critical safety information. A further unanticipated finding was that placing the critical safety information within a separate Warning section. as opposed to placing it within the Directions. dramatically reduced the number of subjects who read and fully recalled the warning yet failed to comply with the warning because of forgetting. This suggests that the utility of a Warning section may be that if the material is read, it becomes more salient and less forgettable to the user. Even in the best condition. however. warning compliance vas only 37% this suggests that steps beyond just warning a user. such as designing the potentially dangerous aspects out of the system to begin with. are needed.