Message reactance as a mediator of effects of alcohol and cancer risk messages

This study examined whether message reactance mediates the cognitive effects of messages about the causal relationship between alcohol and cancer. We randomly assigned US adults (N = 799) recruited from a commercial online research panel to receive one of two types of messages about the alcohol-cancer link: (1) high-certainty (e.g., “Drinking alcohol causes cancer”), and (2) low-certainty, employing the modal verb “may” (e.g., “Drinking alcohol may cause cancer”). We evaluated whether message reactance mediated the effects of message certainty vs. uncertainty on (1) perceived causal certainty about the alcohol-cancer link, and (2) perceived alcohol-related cancer risk. We also explored whether heavy vs. non-heavy alcohol consumption moderated these effects. Message reactance mediated the effects of high-certainty (vs. low) messages on perceived causal certainty (direct: b=0.54; indirect: b=-0.11; both ps<0.001) and alcohol-related cancer risk (direct: b=0.37; indirect: b=-0.13; both ps<0.001)—suppressing the positive effects of expressed certainty on both outcomes. The suppressive effects of message reactance on perceived causal certainty were stronger for participants reporting heavy (b=-0.50) vs. non-heavy alcohol consumption (b=-0.35; p=.034). Conditional indirect effects for the perceived causal certainty model only remained significant for participants with non-heavy alcohol consumption (b=-0.10; p=.004). Effects of high-certainty (vs. low) messages were negatively mediated by message reactance and moderated by alcohol consumption. Findings suggest that although high-certainty risk messages may increase cancer risk perceptions, their effect is attenuated by inciting negative reactance to the messages, particularly for those who consume a heavy amount of alcohol. These mixed effects should be considered in future messaging initiatives aimed at communicating the alcohol-cancer relationship to the public.