Actions Speak Louder than Words: How Characters’ Effectiveness as Message Sources Depend on Their Story Experiences

ABSTRACT

A television (TV) character’s actions and the consequences of these actions in TV storylines can shape the audience’s own behavioral intentions, especially if the audience identifies with that character. The current research examines how storylines depicting positive versus negative consequences of drinking affect youths’ drinking intentions, and whether post-narrative intervention messages delivered by story characters alter these influences. Results indicate that a post-narrative intervention can correct drinking intentions shaped by a pro-alcohol storyline, but the effectiveness depends on the source: a peripheral character is more effective than the main character at delivering a corrective message. This research pinpoints the role of identification with the main character as a key driver of stories’ influence and a key focus of health intervention efforts to correct these stories’ potentially undesirable impact on vulnerable audiences.

Acknowledgments

The first author also acknowledges support through a fellowship from the European Commission’s Marie Skłodowska-Curie Actions.

Notes

1. Participants reported current alcohol consumption by indicating how often they currently drink each of each of seven types of alcohol: beer, wine, strong alcohols, pre-mixes, cider, champagne, other alcohol. Answers to these items were coded as never = 0; rarely = 1; about once a month = 2; about once a week = 3; daily = 4 and were averaged (α = .89) to create a current drinking indicator.

2. There were no effects of post-narrative intervention source on drinking intention in the negative storyline condition.

3. To check whether the manipulations evoked differences in identification with the focal character, we conducted a 2 (story valence: positive, negative) x 3 (post-narrative intervention: control, Tom-delivered, Katie-delivered) ANCOVA on identification with Tom, controlling for gender and age. No effects were significant (Fs < 2.92), with the exception of gender (F(1,312) = 5.89, p = .02), such that males were more likely to identify with Tom.

4. Analyses including the negative conditions (and each of the two- and three-way interactions with valence) showed that valence was not significant. Differences in drinking intentions emerged only in the positive condition. We thus focus the analysis on the positive conditions in the interest of ease of presentation.

Additional information

Funding

This research was funded by grant for project IMAJE (Impact des Médias Auprès des Jeunes) from the French National Cancer Institute [INCA – DEPREV 2013-149] to HYGEE centre - Lucien Neuwirth Cancer Institute ICLN.