The impact of self-construal and message frame valence on reactance: a cross-cultural study in charity advertising

Abstract

In the context of charity advertising, this two-study design project aims to contribute at the intersection of three literatures: psychological reactance theory (PRT), messaging framing, and self-construal theory. Using a survey with student samples from the US and China, Study 1 demonstrated that self-construal affected reactance, such that independent self-construal was more associated with reactance. Further, it outperformed cultural background in predicting reactance. People with predominantly interdependent self-construals showed higher intention to purchase a product with social causes. Study 2 was a 2 (framing: gain vs. loss) × 2 (self-construal: independent vs. interdependent) between-subjects experiment using a non-student sample in the US results on the associations between self-construal and reactance and the intent to donate were consistent with findings of Study 1. Study 2 also indicated that compared to gain-framed appeals, using loss-framed appeals in charity advertising generated more reactance, the gap was more pronounced among individuals with relatively higher interdependent self-construals (i.e. moderate and high levels of interdependent-independent self-construal). This project offers important theoretical and applied implications and provides a robust avenue for future research. Limitations were also outlined.

Keywords:

• Charity advertising
• framing
• self-construal
• reactance

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

Notes

1 When implementing cause-related advertising campaigns, companies sometimes increase prices to partially cover some of the costs of the campaign. Recent research suggests that consumers perceive that modest increases as fair and are willing to bear some of the costs (Koschate-Fischer, Huber, and Hoyer 2016). In Study 1’s scenario, T-shirt B is priced slightly higher (10%) than T-shirt A. T-shirt B also has a label indicating 10% of the profits will go to the charity cause, which balances out the slight price difference between these two. T-shirt A and B are equivalent in customers’ style and color. More importantly, Study 1 is a survey, which doesn’t afford any causality evidence.

2 The original scale (Gudykunst et al., 1996) has 29 items. The current study used a shorter-version containing 12 items, all of which loaded 0.50 or greater in Gudykunst et al.’s (1996) study. Gudykunst and Lee (2003) concluded that these 12 can be used to construct the short version, which is ‘consistently reliable’ (p. 267). Rubin et al. (2009) endorsed this short version in their book on communication research measures. Also, the author used this 12-item version to temper participant fatigue and attrition given the total number of variables included in the study and better ensures conceptual equivalence using the back translation method.

3 Extant literature has indicated that fighting hunger in Africa and helping children in need are not only two very popular causes among various age groups in different cultural groups, but also have been incorporated in cause-related marketing campaigns by many corporations (Basil, Ridgway, and Basil 2006, 2008; Chang and Lee 2010; Das, Kerkhof, and Kuiper 2008; Kim and Johnson 2013; Koschate-Fischer, Huber, and Hoyer 2016). Using these two different charity causes as conceptual replication also enhances the external validity of this project in totality.

4 As in the main study, participants in the pretest also rated the extent to which the ad highlighted ‘the positive consequences of donating to ChildFrontier’ or ‘the negative consequences of not donating to ChildFrontier’ on a 1–5 scale. They perceived more positive consequences (gains) in the gain condition (M = 3.38, SD = 0.90) than in the loss condition (M = 2.80, SD = 1.01, t(78) = 2.68, p = 0.01). They perceived more negative consequences (losses) in the loss condition (M = 3.20, SD = 0.82) than in the gain condition (M = 2.83, SD = 0.78), t(78) = 2.09, p = 0.04.

Additional information

Funding

This research was supported by a grant from Villanova University’s Waterhouse Family Institute for the Study of Communication and Society (WFI).