Download Testing the Validity of the Extended Parallel Processing Model ... and more Slides Marketing in PDF only on Docsity!
Environmental Protection: Testing the Validity of the Extended Parallel
Processing Model (EPPM) when a Threat and Proposed Solution
Require Collective Action
June Marchand
Université Laval
Marlyatou Diallo
Université Laval
Cite as:
Marchand June, Diallo Marlyatou (2020), Environmental Protection: Testing the
Validity of the Extended Parallel Processing Model (EPPM) when a Threat and
Proposed Solution Require Collective Action. Proceedings of the European Marketing
Academy , 49th, (53108)
Environmental Protection: Testing the Validity of the Extended Parallel
Processing Model (EPPM) when a Threat and Proposed Solution Require
Collective Action
Abstract: The Extended Parallel Process Model (EPPM) posits that a fear-inducing message will be effective if it presents a serious and relevant threat. To generate favorable reactions, a strong threat combined with a high-efficacy recommendation is advised. By contrast, a strong threat with a low-efficacy recommendation will trigger counterproductive reactions. This research tests these predictions using concepts of collective threat and collective efficacy adapted to fight climate change. The results were partially consistent with the EPPM. As expected, a message with a strong collective threat and a high-efficacy collective recommendation was the most persuasive. Also, a message presenting a strong collective threat and a low-efficacy collective recommendation generated more negative reactions than the one with a strong threat and high-efficacy recommendation. However, contrary to our expectations, a message presenting a weak threat did not elicit the weakest reactions. Keywords: EPPM, Environment Social Marketing, Fear Advertising. Track: Advertising & marketing communication
impact, it must be adopted by as many people as possible. In the present study, we integrate concepts of collective threat and collective efficacy to test EPPM hypotheses. Our objective is to determine whether the predictions of the EPPM will be confirmed or invalidated when a population faces a collective threat requiring collective action. Figure 1: The Extended Parallel Process Model (Witte et al., 2 001)
2. Hypothesis The hypotheses tested in this study are based on those of the EPPM. However, notions of threat and efficacy are expressed as collective rather than individual. Hypothesis 1a : The group exposed to a low-threat message ( LT ) will have weaker threat- control reactions (attitude and intention) than the high-threat and high-efficacy group ( HTHE ). Hypothesis 1b : The group exposed to a low-threat message ( LT ) will have lower fear-control reactions (defensive avoidance and reactance) than the high-threat/low-efficacy group ( HTLE ). Hypothesis 2 : The group exposed to a message with high threat and low efficacy ( HTLE ) will have higher threat-control reactions than the other groups (higher scores for reactance and defensive avoidance). Hypothesis 3 : The group exposed to a message with high threat and high efficacy ( HTHE ) will have more favorable danger-control reactions than the other groups (higher scores for attitude and intention). 3. Material The three messages used in the present study were created by modifying a Greenpeace poster. The original poster presents an image of an hourglass in which an iceberg in the upper
bulb appears to be melting, drop by drop, to flood a house in the lower bulb. This image is accompanied by the slogan: “Help us fight the melting of the ice.” This image appealed to us for two reasons. Firstly, the hourglass image illustrates some of the consequences of global warming, in particular the connection between melting ice and flooding. Secondly, it recommends energy-saving actions, a behavior advocated in our study. To adapt the poster to our study, we created three different versions of the poster (Figure 2). Message 1 Low threat/High efficacy ( LT ) Message 2 High threat/Low efficacy ( HTLE ) Message 3 High threat/High efficacy ( HTHE ) Figure 2: The three posters used in this experiment The first version presents a low threat. It should be recalled here that the two components of a threat are perceived severity and perceived vulnerability. To mitigate the threat, we played with the aspect of vulnerability. As we knew our audience would predominantly be Canadian, we played with geographical distance. Thus, in the first version of our poster, we replaced the more familiar image of a house with an image of flooded huts in Bangladesh. This image was accompanied by the message: “With global warming, the AVEC LE RÉCHAUFFEMENT CLIMATIQUE, LE NOMBRE D‘INONDATIONS DÉVASTATRICES POURRAIT CROÎTRE SUR LA PLANÈTE! DES ORDINATEUR GESTES SIMPLES AU LIEU COMME DE LE METTRE LE FAIT END’ÉTEINDRE VEILLE PEUVENT SON AIDER À RÉCHAUFFEMENT ÉCONOMISER L’ÉNERGIE. D’APRÈS ET HYDRO À LUTTER-QUÉBEC, CONTRE LES LE APPAREILS MIS EN VEILLE AU CANADA CONSOMMENT L’ÉQUIVALENT EN ÉLECTRICITÉ DE 300 000 MÉNAGES. P NONRENDRE UTILISÉS TOUS PENDANT L’HABITUDE PLUS D’ÉTEINDRE D’UNE HEURE, LES ORDINATEURSEST UN GESTE SIMPLE QUI PROFITE À LA FOIS À VOTRE APPAREIL, À VOTRE FACTURE ET AU CLIMAT! ÉTEIGNEZ VOS APPAREILS AU LIEU DE LES METTRE EN VEILLE! AVEC LE RÉCHAUFFEMENT CLIMATIQUE, LE NOMBRE D‘INONDATIONS DÉVASTATRICES POURRAIT CROÎTRE AU QUÉBEC! ÉTEIGNEZ VOS APPAREILS AU LIEU DE LES METTRE EN VEILLE! AVEC LE RÉCHAUFFEMENT CLIMATIQUE, LE NOMBRE D‘INONDATIONS DÉVASTATRICES POURRAIT CROÎTRE AU QUÉBEC! DES ORDINATEUR GESTES SIMPLES AU LIEU COMME DE LE METTRE LE FAIT END’ÉTEINDRE VEILLE PEUVENT SON AIDER À RÉCHAUFFEMENT ÉCONOMISER L’ÉNERGIE. D’APRÈS ET HYDRO À LUTTER-QUÉBEC, CONTRE LES LE APPAREILS MIS EN VEILLE AU CANADA CONSOMMENT L’ÉQUIVALENT EN ÉLECTRICITÉ DE 300 000 MÉNAGES. P NONRENDRE UTILISÉS TOUS PENDANT L’HABITUDE PLUS D’ÉTEINDRE D’UNE HEURE, LES ORDINATEURS EST UN GESTE SIMPLE QUI PROFITE À LA FOIS À VOTRE APPAREIL, À VOTRE FACTURE ET AU CLIMAT! ÉTEIGNEZ VOS APPAREILS AU LIEU DE LES METTRE EN VEILLE!
4. Methodology 4.1 Measured variables Seven (7) variables were measured using a 5-level Likert scale (1 = strongly disagree, 5 = strongly agree), while the attitude variable required the use of semantic items. The internal consistency was measured using Cronbach's alpha and considered adequate. Table 2 presents these results, as well as an example of the item used for each measured variable. Table 2 : Variables measured with Cronbach’s alphas Variables Number of items α Threat variables: Severity (“Climate change is a serious threat” ) Susceptibility (“Some negative effects of climate change will affect Quebec”) 3 3 . 769 . 896 Efficacy variables: Response efficacy (“The effects of climate change can be reduced by saving energy by turning off your computer instead of putting it on standby”) Self-efficacy (“Turning off my computer when I don't use it for a while (an hour or more) is a simple thing for me to do”) 3 3 . 832 . 759 Fear control process variables: Reactance (“The message deliberately tries to take advantage of my feelings”) Defensive avoidance (“When I saw the message, I had the reaction to ignore the information presented in it”) 3 3 . 843 . 828 Threat control process variables: Attitude (“Turning off my computer to save energy is...”) Intention (“From now on, I will turn off my computer instead of putting it on standby”) 7 3 . 922 . 932 4.2 Procedure and participants One hundred and forty-three participants were recruited in their classrooms. Each of them received an envelope containing a consent form, a questionnaire and one of the three versions of the poster. Variables were measured only once, after exposure to a single version of the message. The experiment lasted about 20 minutes. Data collection was carried out for just under one month. The final sample is described in Table 3. 5. Results 5.1 Fear control process (see Figure 1) The results obtained by the three groups of respondents on the reactance and defensive avoidance variables offered support for hypothesis 1a, as well as hypothesis 2. In order to confirm hypothesis 1a, the low-threat group (LT) had to obtain a significantly lower score on the avoidance and reactance variables than the high-threat/low- efficacy group (HTLE). Anova analysis revealed that there were no significant differences
between the three groups for defensive avoidance (F= 0.914, p= 0.403). Thus, no group was more likely than another to reject the message in order to protect itself from its more or less threatening content. Table 3 : Sample characteristics Demographic characteristics Number Percent Age 15 – 19 20 – 24 25 – 29 30 – 34 35 and over 30 91 19 2 1 21
Sex Male Female 61 82
Origin Province of Quebec Other 134 9
Field of study Information & communication Other 129 14
Owns a laptop computer Yes No 136 7
The experiment showed a significant difference in reactance between the three groups (F= 7.308, p = 0.001). The post-hoc test showed that participants in the high-threat group (HT) had a significantly higher mean score than those in the high-threat/low-efficacy group (HTLE) (p = 0.013). This result also invalidates hypothesis 1b for reactance because, contrary to expectations, participants in the low-threat group (LT) were more likely than those in the high-threat/high-efficacy (HTHE) group to feel manipulated by the message. To support hypothesis 2, the high-threat/low-efficacy group (HTLE) would need to obtain the highest score for both defensive avoidance and reactance. However, the results clearly show that the group exposed to the high-threat/low-efficacy message (HTLE) was no more likely than the other groups to adopt the fear control process, whether in the form of reactance or defensive avoidance. Therefore, hypothesis 2 is not supported by our results. 5.2 Threat control process Attitude and intention variables were used to test hypothesis 1b and hypothesis 3. In the first case, Anova results showed significant differences between the three groups for both attitude (F= 147.723, p = 0.001) and intention (F= 60.686, p = 0.001). The post-hoc test revealed that the attitude score was not significantly lower for the low-threat group (LT) than for the high-threat/high-efficacy group (HTHE) (p = 0.167). Thus, the low-threat group attitude score did not confirm the model’s prediction. As for the behavior intention variable,
Scharks, T. (2016). Threatening messages in climate change communication, unpublished Doctoral Dissertation, University of Washington, United-States. Retreived from ProQuest Dissertations & Theses Global : https://search-proquest- com.acces.bibl.ulaval.ca/docview/1821905070/BDB331B27B994505PQ/1?accountid=12008. (Last accessed June 15, 2017). Witte, K, Meyer, G, & Martell, D. (2001). Effective Health Risk Messages: A step-by-step guide. Los Angeles, CA: Sage Publications.
