Implicit-explicit ambivalent attitudes and health behaviors

Date of Completion

January 2006


Psychology, Behavioral|Psychology, Social|Health Sciences, Public Health




Most research assumes that health-related attitudes are positive or negative rather than positive and negative, yet research in other domains shows that attitude structures can often have both positive and negative components (ambivalence). Not only can these ambivalent attitudes impact people's subsequent behaviors, but implicit attitudes have been shown to be highly predictive of subsequent marijuana use. The purpose of this study was to (a) explore dimensions of implicit-explicit ambivalent marijuana attitudes, (b) examine the relation between ambivalent attitudes and marijuana use, and (c) investigate whether a conditioning task is able to improve automatic marijuana associations. Specifically, this research tested the hypothesis that people with ambivalent marijuana attitudes engage in more marijuana use because their decision-making processes are impaired by simultaneous positive and negative cognitions and that an unobtrusive conditioning task can be used to improve participants' ambivalent marijuana attitudes. A sample of 191 marijuana-using undergraduates with ambivalent or unambivalent marijuana attitudes were randomly assigned to an intervention (I), intervention plus daily health report (I+DHR), daily health report (DHR), or a control (C) condition and were assessed at two time points, 6-weeks apart. During the first session, participants in the I or I+DHR conditions were exposed to a conditioning task that paired marijuana and negative images. For the 40 days between measurement occasions, participants assigned to the I+DHR or DHR conditions completed a daily report of their health behaviors and ambivalent attitudes. Overall findings show that ambivalence, but not implicit attitudes, attenuated the intention-behavior and attitude-behavior relations. Within the theory of planned behavior, the moderating role of ambivalence was examined; findings show that perceived behavioral control-intentions and perceived behavioral control-behavior relations were stronger when participant held ambivalent compared to unambivalent marijuana attitudes. Condition assignment did not impact participants' marijuana use or behavioral intentions but did impact participants marijuana attitudes; participants in the I+DHR condition significantly reduced their positive attitudes toward marijuana use over the 6-week period. Growth curve analyses showed that DHR participants were less ambivalent about marijuana use over time than I+DHR participants. Explanations for these findings and implications for health promotion research are discussed. ^