Page 225 - The Handbook of Persuasion and Social Marketing

Empirical Generalizations for Social Marketing 213

Attitudes, Intentions, and Behaviors

The relationships between attitudes, intentions, and behaviors have been
extensively studied in many contexts, resulting in hundreds of studies
summarized in multiple meta-analyses. Table 8.2 shows the results for
miscellaneous topics as well as focused research on blood donation, con-
dom use, health-related screening programs, exercise, smoking, and pro-
environmental behaviors. The correlation between attitudes and intentions
equals .50 on average, slightly higher than the .45 correlation between
intentions and behaviors. Both are higher than the .36 correlation between
attitudes and behaviors. These are medium-to-large effects that are gener-
ally much stronger than the results for the interventions shown in
Table 8.1.
The major exception to the general pattern described above involves
studies on smoking (Topa & Moriano, 2010). The correlation between
smoking attitudes and intentions is r = .16, almost identical to the .17 cor-
relation between smoking attitudes and behaviors. The intentions-behav-
ior correlation is higher at r = .30. The lower values relative to the other
results in Table 8.2 may reflect the greater freedom of choice for the other
behaviors compared to the addictive nature of cigarette smoking (West,
2009).
The overall correlations can be used in a structural model to calculate
the effect of attitudes on behaviors, controlling for the effect of intentions.
The results show that attitudes have an indirect effect on behaviors of .36,
mediated by intentions, as well as a direct effect of .20. (The direct effect
of intentions on behaviors, controlling for attitudes, happens to also equal
.36.) Therefore, social marketing programs may better influence behaviors
by addressing both attitudes and intentions rather than assuming that ad-
dressing intentions alone is sufficient.

Effects of Alternative Persuasive Techniques

Table 8.3 summarizes the findings for various approaches that can be used
in designing persuasive messages: comparative advertising, explicit versus
implicit conclusions, fear appeals, gain versus loss framing, humor, pow-
erful versus powerless language, asking rhetorical questions, testimonial
assertion evidence, two-sided messages, and warning labels. Most of the
effect sizes are small but positive, implying that certain approaches tend to
be more persuasive than other approaches. Because it costs no more to
implement one technique than another, these findings may have useful
practical implications for social marketing programs.