Mitchell Ricketts, Ph.D. (2007)

Unintentional injuries represent the leading cause of death among Americans aged 1-44 years. While there have been many life-saving advances in engineering, attempts to save lives by changing people's behavior have been less successful. For instance, safety and health communications have sometimes led to increased knowledge and self-reported intentions to comply with recommendations, but traditional efforts to demonstrate changes in actual target behaviors have often failed.

Research in many settings has shown that narrative communications have exceptional power to persuade and affect peoples' decisions. This suggests that safety and health messages might be more effective if they include narratives, such as brief stories about people who have been injured. The purpose of this dissertation is to determine if safety communications that include stories about injuries result in superior behavioral compliance when compared with traditional abstract safety messages.

Teams of two participants assembled a swing set, using written instructions that contained relevant safety messages. Fifty-four teams were randomly assigned to three conditions: story-based safety messages, concrete nonstory safety messages, and traditional abstract safety messages. Compliance with safety messages was defined as the number of compliant components in the finished swing set. After adjustment for covariates, story-based messages resulted in a 20 percent improvement in compliance, compared with concrete nonstory and traditional abstract messages. Covariates included age, gender, (log) childcare experience, equipment assembly experience, presence of observer, and a final covariate related to timing of experimental sessions conducted by different experimenters.

A positive relationship was noted between behavioral compliance and immediate (but not delayed) recall of message content. Narrative transportation was also positively related to compliance, but only within the story-based condition. Behavioral compliance was not related to remindings or judgments about the likelihood of injuries.

The research is important because of its potential for improving safety communications and saving lives. Stories about injuries improved safety behavior even though the stories were brief and not designed to be entertaining or transporting. In contrast, the lack of correspondence between observed behavior and many surrogate measures suggests caution is in order when evaluating interventions using self-report measures, delayed memory, and other common dependent variables.