Table 3.

Tobacco Warnings Model versus reasoned action mediators in explaining health warnings’ effect on sugar-sweetened beverage (SSB) purchases

Construct	Warning on mediator	Mediator on behavior	Mediated effect
Tobacco Warnings Model mediators
Attention	●	○	○
Thinking about harms	●	●	●
Negative affect	●	●	●
Anticipated social interactions	●	●	●
Intentions to limit SSBs in store	●	●	●
Intentions to limit beverages with added sugar	○	●	○
Reasoned action mediators
Perceptions of added sugar	○	●	○
Positive product attitudes	○	●	○
Positive behavioral attitudes	○	●	○
Negative outcome expectations	○	○	○
Self-efficacy	○	●	○
Response efficacy	○	○	○
Injunctive norms	●	●	●

Table shows statistical significance of path coefficients (second column: a pathway, regressing the mediator on treatment arm; third column: b pathway, regressing sugar-sweetened beverage purchases on the mediator, controlling for treatment arm) and mediated effects (last column: a × b). Coefficients were estimated in linear regressions controlling for Hispanic ethnicity and overweight status.

● Statistically significant path coefficient or mediated effect, p < .05.

○ Path coefficient or mediated effect was not statistically significant, p > .05.