MASTER THESIS
Using the Self as a Persuasion Tactic
An analysis of self-boosting messages and fear appeals regarding alcohol
consumption and breast cancer
Eveline Maertens Student ID: 11122994
Graduate School of Communication Master’s Program Communication Science Supervisor: Dr. Stephanie Welten
Abstract
The aim of this experiment was to investigate the relationship between fear appeals (strong fear vs. weak fear) and self-boost messages (efficacy vs. self-affirmation vs. no self-boost) and how they influence message persuasiveness regarding alcohol consumption and breast cancer. Previous research explains efficacy to be challenging to increase, which is a problem for persuasive purposes. A solution might be to explore self-affirmation, as research shows it to be a promising boost strategy. Comparing the effects of an efficacy message vs. a self-affirmation message together with a fear appeal can show which strategy leads to higher message persuasiveness. The manipulations from this experiment were integrated in a fictional health website in order for it to be usable in practice, as most research on self-boost messages and fear appeals are lab experiments. It was expected that a strong fear appeal compared to a weak fear appeal and an efficacy or a self-affirmation message compared to a no self-boost message would lead to higher message persuasiveness. However, the results were non-significant. This study demonstrates the difficulty of increasing efficacy and self-affirmation through communication research.
Introduction
An author once wrote, “you can’t stop being afraid just by pretending everything that scares you isn’t there” (as cited in Lewis, 2013, pp. 87). Being afraid or fearful can work as a powerful incentive to perform a health recommendation, such as reducing alcohol
consumption. Thus, fear is often used as a persuasive tool, referred to as a fear appeal (Sutton & Hallet, 1988). Health interventions that use fear appeals tend to agree with the above quote, and hope that one will face fear and behave in a cautionary manner, according to a health recommendation (Rogers, 1975). A strong fear appeal enhances perceived threat towards one’s health. However, in order to behave according to the perceived threat, a persons’ self-efficacy (having confidence in performing the health recommendation), in addition to response efficacy (believing the health recommendation will decrease threat) must also be high (Witte & Allen, 2000). Increasing efficacy is therefore a common tactic when using fear appeals, so that one’s confidence in performing the health recommendation increases (Good & Abraham, 2011). In order to increase one’s efficacy, health interventions tend to provide a self-boost message before presenting the fear appeal (Fry & Prentice-Dunn, 2005).
Although fear appeals have been studied for more than 60 years, there is still no clear answer to the most effective technique of persuasion (Ruiter, Kessels, Peters & Kok, 2014). Research has shown the challenge of successfully increasing efficacy. It has been shown to be hard to provide personally relevant information and feasible recommendations (Ruiter et al., 2014). As a result, more research needs to be done on how to provide a motive for people to perform a health recommendation. Increasing one’s self-affirmation is one promising
solution, since people’s self-image can serve as an incentive to behave a certain way. People that are self-affirmed have greater motivation to act morally correct (Steele, 1988). When the self is threatened with a fear appeal, it acknowledges that a person’s current behavior is inadequate. As a result, a restoration process of the self is activated, as people want to be of
high integrity (Steele, 1988). Thus, self-affirmed people may feel obliged to perform healthy recommendations.
Unfortunately, there is little research done on self-affirmation and fear appeals (Napper, Harris & Epton, 2009). An increase in research on self-affirmation and fear appeals can result in more consistent data. The findings on self-affirmation manipulations and fear appeals are currently varied; while some result in message persuasiveness, others do not (Harris & Napper, 2005). Thus, more research needs to be done on self-affirmation in combination with fear appeals in order to know how to successfully manipulate
self-affirmation, in addition to expand the current findings and provide insight for future studies and health interventions.
The few studies on self-affirmation and fear appeals have mainly been lab experiments (McQueen & Klein, 2006). The advantage of lab experiments is that they are more reliable and more controlled (Falk & Heckman, 2009). However, the downside of lab experiments is that their results may not always be replicable in real life scenarios. Thus, there needs to be communication research done on this topic to discover how to increase self-affirmation in real health campaigns (McQueen & Klein, 2006). This is important since the manipulations from the current study can be integrated in mass-reach campaigns, which can persuade the public to change certain behaviors (Brown & West, 2014). When an intervention is successful in
changing behavior, it also decreases the chance of developing the health risk associated to the behavior (Brown & West, 2014).
Finally, although challenging to manipulate, efficacy has been widely studied in relation to fear appeals and is known to result in adaptive behavior (Floyd, Prentice-Dunn & Rogers, 2000). On the other hand, increasing one’s self-affirmation is also promising, as presented above. Thus, in order to examine what self-boost (efficacy vs. self-affirmation) in
combination with fear appeals results in the highest message persuasiveness, this communication research will compare a self-affirmation to an efficacy manipulation.
The health recommendation this research focuses on is decreasing alcohol consumption among female students. Research shows that alcohol consumption among women can lead to breast cancer (Longnecker, 1994). Breast cancer is the most common type of cancer for women (Fry & Prentice-Dunn, 2005). The data resulting from this study can be used when implementing health interventions on a mass scale or advertisements on
decreasing alcohol consumption among women.
The aim of this research is to investigate the relationship between fear appeals and self-boost messages and how they influence message persuasiveness regarding alcohol consumption and breast cancer. Thus this leads to the following research question: To what extent does an efficacy message vs. a self-affirmation message prior to an
advertisement containing a strong vs. a weak fear appeal, influence message persuasiveness (intention & message acceptance) among female students to reduce alcohol consumption?
The method of the study will be a 3 (self-boost: efficacy message vs. self-affirmation message vs. no self-boost) × 2 (Fear appeal: strong vs. weak) between subjects experimental design.
Theoretical Framework Theories and Models on Fear Appeals
One of the first models developed on fear appeals are drive models (Witte & Allen, 2000). This includes Janis’s Fear as Acquired Drive Model from 1967. Overall, this model proposes that arousing fear is necessary for a message to be successful in influencing one’s behavior (Maddux & Rogers, 1983). Drive models were one of the first models to discover that as fear is stimulated by external cues, as one is motivated to protect oneself in order to
reduce this sensation of fear (Leventhal, 1971). Throughout the years drive models have been rejected, as some details do not overlap with empirical evidence. For instance, drive models predicted that a moderate amount of fear would lead to the greatest amount of attitude change, yet this is not supported through research (Witte & Allen, 2000). However, the main point of drive models is that fear acts as a drive to stimulate behavior change, and this is reflected by a large percentage of experimental studies (Sutton & Hallet, 1988).
Another theory that points out the importance of evoking fear to stimulate behavior change is Sutton’s application of the Subjective Expected Utility Theory from 1982 (Witte, 1992). This theory takes three processes into consideration namely the perceived utility of the threat, the chance that the threat will occur if the health recommendation is accepted and the chance that the threat will occur when the health message is not accepted (Witte, 1992). Thus, according to Witte (1992), perceived threat is the main factor that influences whether a fear appeal’s recommendation is accepted. Yet, again due to the inconsistent empirical research on fear appeals, this theory is not fully supported (Witte & Allen, 2000). However, according to Witte and Allen (2000), the theory does support the findings of their meta-analysis, that the greater the perceived threat, the greater the intention to accept the health recommendation.
The more recent models on fear appeals like the Extended Parallel Process Model also describe the important role that perceived threat has on message persuasiveness (Witte, 1992). Primarily, it explains that when threat is low, the health message is not even processed (Witte, 1992). Thus, perceived threat is necessary for one to process the message.
The usage of fear appeals is a common communication strategy (Ruiter et al., 2014). Fear appeals can be used in various forms such as through text or images to catch one’s attention and trigger fear (Brown & West, 2014). In essence, fear appeals are a persuasive tactic to motivate people to follow a certain health recommendation and change one’s behavior (Witte, 1992).
When a fear appeal is successful it increases one’s perceived threat (Witte, 1992). The most influential manner of increasing perceived threat is increasing both the severity and susceptibility of the risk associated with the behavior the health message is trying to change (Witte, 1992). Severity is defined as how large the consequence of the risky behavior is, while susceptibility is described as how likely it is that one will develop the health risk (Ruiter et al., 2014). In this study for instance, it is important to portray the severity and susceptibility of developing breast cancer so the participant feels threatened.
The success of a fear appeal can be measured through message persuasiveness, or in other words, how influential the health recommendation is (Harris & Napper, 2005). Message persuasiveness can be measured through message acceptance. In addition, intention to
perform the recommended behavior is an indicator of message acceptance (Harris & Napper, 2005). Although the intention behavior gap demonstrates that intention does not always lead to behavior, it is an accepted predictor of behavior. Furthermore, actual behavior is very difficult to measure (Peter, Ruiter & Kok, 2013). Thus, overall, as one accepts the health message or recommendation and intends to perform the recommended behavior, message persuasiveness is achieved. As a result this leads to the first hypothesis of this experiment:
H1a-The presence of strong fear appeals will result in greater intention to decrease alcohol consumption than the presence of weak fear appeals
H1b-The presence of strong fear appeals will result in greater message acceptance of messages on decreasing alcohol consumption than the presence of weak fear appeals
Efficacy
However, threat alone will not suffice according to the EPPM (Extended Parallel Process Model). Perceived efficacy, in addition, is a crucial element in the later theories on
fear appeals such as the Protection Motivation Theory (PMT) and the EPPM. Perceived efficacy is comprised of two components namely response efficacy and self-efficacy (Ruiter et al., 2014). Response efficacy is defined as how effective the health recommendation
appears to be in order to reduce the health risk (Witte, 1992). For example this was influenced with a message in previous research by, explaining that using sunscreen reduces the chance of developing skin cancer (Good & Abraham, 2011). In the case of this study, response-efficacy could possibly be increased, by explaining that decreasing alcohol consumption can reduce the chance of developing breast cancer. Self-efficacy on the other hand, refers to whether one is confident in having the ability to follow the health recommendation (Witte, 1992). For instance, this can be influenced with a message by, providing the participant with tips on how to decrease their alcohol consumption (Brown & West, 2014).
The Theory of Self-Efficacy is an important component of Bandura’s Social Cognitive Theory, as the level of one’s self-efficacy plays a major role in whether a behavior is actually performed (Bandura, 2001). The Social Cognitive Theory describes personal, behavioral and environmental determinants, which influence how humans think, feel and behave (Bandura, 2001). Increasing self-efficacy can influence several processes that affect message
persuasiveness. Self-efficacy can influence cognitive processes, as higher self-efficacy results in greater goals and a stronger commitment to accomplish them (Bandura, 1989). One’s motivational process can also be influenced as one’s effort for behavior change increases when self-efficacy is high. Furthermore, the affective process is reflected as self-efficacy increases, positive emotions also increase toward the health recommendation. Lastly, the selection process is affected as when people have low self-efficacy, people usually avoid situations where they feel to have a lack of ability (Bandura, 1989). Thus, greater self-efficacy results in greater behavior change. Research has also shown that interventions aimed at
increasing one’s self-efficacy tend to result in higher message acceptance and intention to perform the recommended behavior (Good & Abraham, 2011).
The first theory on fear appeals to include efficacy was the Protection Motivation Theory from 1983, by Rogers (Snipes, LaTour & Bliss, 1999). This theory describes the crucial need to increase efficacy when perceived threat is increased in order to stimulate message persuasiveness (Maddux & Rogers, 1983). Efficacy is labeled as being the most important component in the presence of fear appeals (Snipes et al., 1999). As a result, this leads to the following hypothesis:
H2a- Presenting an efficacy message will result in greater intention to decrease alcohol consumption than presenting a no self-boost message
H2b- Presenting an efficacy message will result in greater message acceptance of messages on in decreasing alcohol consumption than presenting a no self-boost message
Self-Affirmation
As described above, efficacy and fear appeals have extensively been researched and have been proven to be an effective tactic to result in adaptive behavior. However, the downside of utilizing efficacy as a coping strategy is the difficulty of creating such a
personally relevant and feasible recommendation (Ruiter et al., 2014). A different technique when it comes to fear appeals is focusing on the self. The self is described as wanting to be morally correct and in control of important outcomes (Steele, 1988). People are constantly struggling to protect the self (Sherman & Cohen, 2002). According to Reed and Aspinwall (1988), one factor that may challenge the self is being presented with information on health risks. When the self is threatened it may result in short-term losses such as negative emotions and a decrease in self-value (Reed & Aspinwall, 1998). In order to protect one’s integrity,
people may react to the threatening information in a defensive manner. For example, a person can interpret the information in a way that reflects their preexisting beliefs, or even avoid the health message completely (Sherman & Cohen, 2006).
Self-affirmation is increased when a person reflects on values that one possesses which are irrelevant from the health message (Sherman & Cohen, 2006). As a result, when self-affirmation increases, defensive avoidance can be decreased (Steele, 1988). Steele (1988) explains in The Self-Affirmation Theory that when a person is reaffirmed of being a good person, his or her confidence increases. Thus, less energy needs to be spent on protecting one’s self and can instead focus on reducing the threat by performing the adaptive behavior (Reed & Aspinwall, 1998). Consequently he or she has the incentive of performing healthy behavior (McQueen & Klein, 2006).
Self-affirmation manipulations work best when they concern traits that are not related to the health recommendation, as it reduces threat (Good & Abraham, 2011). In addition, people can best be self-affirmed for health messages that are less familiar, like alcohol and breast cancer as described by Good and Abraham (2011). Less familiar health messages work best, as there is a smaller chance that the participant has already developed a defensive reaction for that particular behavior.
Previous research on self-affirmation manipulations is extremely varied, as there is little knowledge on the best way to manipulate self-affirmation (Napper et al., 2009). The following studies demonstrate different ways self-affirmation has been manipulated. Stone and Cooper (2003) successfully manipulated self-affirmation through a sentence scrabble where the participant would find the following hidden phrase: I am a compassionate person. Good and Abraham (2011) found a main effect for self-affirmed vs. non-self-affirmed
participants on message acceptance for sunscreen use on photo ageing. Their self-affirmation manipulation consisted of asking participants to rate certain values from most important to
least important. Fry and Prentice-Dunn (2005) used the same self-affirmation manipulation technique and found no significant difference among self-affirmed and non-self-affirmed participants on intention to perform breast self-examinations to prevent breast cancer. Another study on breast cancer is of Harris and Napper (2005) who successfully increased message acceptance on alcohol consumption and breast cancer for self-affirmed participants. They asked participants to complete a writing assignment on their most important trait. As a result, previous research has found a variety of methods to increase self-affirmation. However, an issue with the above studies is that the self-affirmation manipulation was presented separately from the fear appeal, having little connection between the two. As a result, these studies are difficult to replicate for real life health campaigns that try to include both fear appeal and self-affirmation strategies.
The first and what appears to be the only study to successfully integrate the self-affirmation manipulation and the fear appeal was by Jessop, Simmonds and Sparks (2009) on sunscreen and skin cancer. They compared three forms of self-affirmation manipulations namely values affirmation, kindness affirmation and a positive traits affirmation. Of the three manipulations, only the positive traits affirmation was integrated with the fear appeal on a health leaflet. It comprised of questions on possession of different traits where the participant answered the question with a yes or no (Jessop et al., 2009). The participants in the positive traits affirmation condition resulted in the greatest behavior change, which was measured by those who got a free sample of sunscreen. More research needs to be done regarding self-affirmation manipulations, as most studies have been lab experiments rather than experiments that can be applied in the real world (McQueen & Klein, 2006). This study aims to integrate a self-affirmation manipulation with a fear appeal.
H3a- Presenting a self-affirmation task will result in greater intention to decrease alcohol consumption than presenting a no self-boost message
H3b- Presenting a self-affirmation task will result in greater message acceptance of messages in decreasing alcohol consumption than presenting a no self-boost message
When it comes to fear appeals, efficacy has been described as being an essential aspect (Maddux & Rogers, 1983). However, due to the positive results from previous self-affirmation manipulations, it is interesting to test whether increasing self-self-affirmation works equally well as increasing efficacy when testing message persuasiveness. Thus, this leads to the following question:
RQ1. Does an efficacy message or a self-affirmation message together with a fear appeal increase message persuasiveness most?
Interaction Effect
Self-boosts such as efficacy messages or self-affirmation tasks performed prior to a fear appeal rather than after the fear appeal has been shown to increase one’s confidence to perform a health recommendation (Brown & West, 2014). According to Brown and West (2014), the amount of fear the health message elicits is important for the order in which the self-boost must be presented. For example, when the fear appeal produces a high-perceived threat, the self-boost must come before the fear appeal. This is because the participant is then prepared to be confronted with the threat and will not act defensively. On the other hand, when the fear appeal results in low threat, the self-boost may be presented after the fear appeal because the threat is not high enough to result in defensive avoidance.
Combining a self-boost with a fear appeal was first introduced in the Protection Motivation Theory (PMT) (Snipes, et al., 1999). This theory includes four aspects, namely perceived susceptibility, perceived severity, response efficacy and self-efficacy (Maddux & Rogers, 1983). When the four components of the PMT are high, it is said to lead to the greatest intention to perform a health recommendation and as such result in message
acceptance (Witte, 1992). However, according to Witte and Allen (2000), the PMT lacks in explaining exactly how and when a fear appeal can backfire.
The Extended Parallel Process Model (EPPM) on the other hand, includes what the Protection Motivation Theory lacks. The EPPM is the most prevalent theory nowadays on fear appeals and efficacy (Ruiter et al., 2014). The EPPM explains that when susceptibility and severity is high, it leads to perceived threat, and when self-efficacy and response efficacy is high it leads to perceived efficacy or one’s perceived confidence (Witte, 1992). Since self-affirmation has not been as extensively studied as efficacy, the expected interaction effect on fear appeals and self-boosts for this study will be according to the EPPM. Thus in this study, showing that alcohol consumption can lead to breast cancer, is expected to increase perceived threat, while the self-boost conditions are expected to increase confidence to decrease alcohol consumption. The interaction of perceived threat, stimulated with a fear appeal and one’s confidence which is stimulated via a self-boost, can either lead to danger control or fear control (Witte, 1992). Danger control occurs when one’s confidence is greater than the perceived threat, thus intention to perform the health recommendation and the message is accepted (Witte, 1992). This is for instance when someone sees a health message that alcohol consumption can lead to breast cancer and thus decides to limit his or her alcohol
consumption. On the other hand, fear control or defensive avoidance occurs when perceived threat is greater than one’s confidence (Witte, 1992). This occurs when someone is confronted
with the health message and decides to ignore it and continue consuming the same amount of alcohol or even more as he or she did than before seeing the health message.
Overall, it is important for perceived threat to be high, as threat initiates message processing which leads to one being persuaded by the message (Witte, 1992). If there is no perceived threat, then there is no message processing, thus the message will automatically be ignored (Witte, 1992). Yet, one’s confidence is also important to result in a persuasive message (Ruiter et al., 2014).
Thus, this research hypothesizes a moderation effect of self-boosts on fear appeals and message persuasiveness as presented in Figure 1.
Fig. 1. Conceptual model on message persuasiveness
The expected interaction effect is described in the following hypothesis:
H4a: When people see an efficacy message it results in greater intention to decrease alcohol consumption with a strong compared to a weak fear appeal
H4b: When people see a self-affirmation message it results in greater intention to decrease alcohol consumption with a strong compared to a weak fear appeal
Self-boost: No self-boost vs. Efficacy vs. Self-affirmation Message persuasiveness Intention & Message
acceptance Fear Appeal
Weak vs. Strong
(+)
H4c: When people see a no self-boost message it results in smaller intention to decrease alcohol consumption with a strong compared to a weak fear appeal
H5a: When people see an efficacy message it results in greater message acceptance with a strong compared to a weak fear appeal
H5b: When people see a self-affirmation message it results in greater message acceptance with a strong compared to a weak fear appeal
H5c: When people see a no self-boost message it results in smaller message acceptance with a strong compared to a weak fear appeal
Method Pretest Study Design
A pre-test was conducted in order to optimally select materials for the main
experiment. The pre-test investigated whether the manipulations of the self-boost messages and the fear appeals were successful.
Respondents
The target group of both the pre-test and the main experiment are female students who consume at least 1 glass of alcohol per week. Research shows that students are drinking large amounts of alcohol (Murphy, McDevitt-Murphy, Barnett, 2005). A total of one glass of alcohol per day could already increase one’s risk of developing breast cancer by 11% (Longnecker, 1994). In addition, women in relation to men have a higher risk in developing breast cancer (Fry & Prentice-Dunn, 2005). As a result, the control and demographic variables for both the pre-test and the main experiment include alcohol consumption, age (18-35), gender (female), education (current student) and nationality. Since, the researcher’s network is international, and participants were recruited through convenience sampling on platforms
such as Facebook, it is important for nationality to be equal among different conditions. A total of 80 people started the online pre-test; yet 18 respondents were removed due to incompletion of the questionnaire, which resulted in a total of 62 female students (mostly from a Bachelor Degree level). In addition, the age range of participants was from 19 to 25 (M = 22.21, SD = 1.62), and the average amount of alcohol they consume per week is 6 glasses (M = 6.32, SD = 6.15). Lastly, this sample includes participants from 20 different nationalities, with the Netherlands (N = 27) being the largest group.
Procedure and Measures
The overarching factor presented in the pre-test was to generate insight into the behavior of female students on alcohol consumption. The pre-test consisted of one efficacy manipulation and two self-affirmation manipulations (the affirmation Quiz and Self-affirmation Puzzle). The most successful self-Self-affirmation manipulation will be chosen for the main experiment. The participants were randomly and equally distributed among the three conditions.
Self-boost Conditions Self-affirmation quiz
The Self-affirmation Quiz condition consisted of a personality quiz. This quiz firstly asked participants to write a brief text about an important trait they posses and how it affects their daily life. This followed with 6 statements on personality traits that participants were asked to either agree or disagree to, such as ‘I am kind.’ This task was inspired by previous research conducted by Jessop, Simmonds and Sparks (2009) in which they asked participants to write about their most important value, in addition to a positive traits affirmation where participants could affirm or disaffirm to a list of traits. It was made into a quiz here because this can be used on a real health website, which can be integrated with questions regarding tailored or personalized information.
Self-affirmation puzzle
The second self-affirmation manipulation is a word puzzle, where participants were asked to find the 6 personality traits. Afterwards, they were asked to rank the 6 words from what they considered to be most important to least important. Stone and Copper (2003) were the inspiration for this manipulation, who successfully increased self-affirmation through sentence scrabble. It was made into a word puzzle here because this could be presented on a real health website without the reader knowing its’ purpose.
Efficacy
The efficacy manipulation aimed to increase both self-efficacy and response-efficacy with the presentation of a self-boost message. This message included 7 tips to decrease alcohol consumption. These tips were taken from a website on alcohol consumption: www.drinkaware.co.uk, and from previous research by Brown and West (2014). Response efficacy was tackled by explaining that decreasing alcohol consumption does lead to a decrease in the risk of developing breast cancer. This was based on Good and Abraham (2011) who increased response efficacy by stating that using sunscreen blocks radiation and decreases the chance of developing skin cancer.
Self-boost measures
Self-affirmation measures. In order to be able to compare the conditions on the success of the affirmation manipulations, affirmation was measured after the self-affirmation manipulations, as well as before the efficacy condition. This latter measure was used to serve as a baseline. All the items from the self-affirmation manipulation check were measured on a 9-point Likert-scale. The self-affirmation manipulation check consisted of 5 items. Item 1, “How do you currently feel about yourself?” was based on research from Sherman, Nelson and Steele (2000). The remaining 4 items were based on research from
order to extend the scale as they are measuring the same construct. See Appendix A for a full overview of the self-affirmation items. A principal component analysis (PCA) was conducted which showed that only one component had an eigenvalue above 1 (E.V. = 3.46), which explained 69% of the total variance of the scale. Furthermore, the reliability for the scale of self-affirmation was good, Cronbach’s α = .88 (M = 6.93, SD = 1.28).
Efficacy measures. Moreover, the participant’s efficacy was measured after the efficacy manipulation in addition to before the self-affirmation manipulation was presented, to serve as a baseline.
The efficacy manipulation check consisted of measuring response-efficacy and self-efficacy. Both the response-efficacy scale (see Appendix B) and the self-efficacy scale (see Appendix C) comprised of two items, which are based on an article by Ruiter, Verplanken, Kok and Werrij (2003). Through performing the PCA, the response-efficacy scale proved to be uni-dimensional, with only one component having an eigenvalue above 1 (E.V. = 1.73), which explained 87% of total variance of the scale. In addition, the reliability of this scale is good, Cronbach’s α = .85 (M = 4.69, SD = 1.87). Furthermore, the self-efficacy scale also showed to be uni-dimensional, and had one eigenvalue above 1 (E.V. = 1.77), which explained 89% of the total variance of the scale. In addition, the reliability of the scale was good, Cronbach’s α = .87 (M = 7.78, SD = 1.96).
Fear Appeal Conditions
Afterwards, participants were presented with a fear appeal. Before the fear appeal was shown, however, a filler task was included with simple mathematical problems in order to diminish the possible effect that the self-boost has on the fear appeal. The idea of using a filler task to recover from possible effects was based on research conducted by Holmes and
Mathews (2005). A total of four advertisements were created presenting the risk that alcohol consumption has on breast cancer. These 4 advertisements consisted of 2 sets containing one
strong fear appeal and a weak fear appeal (see Appendix D). In order to increase validity, the pairs of advertisements resembled each other through color, layout, text, and style. The pre-test randomly assigned participants to two conditions (fear appeal: weak vs. strong). The weak fear appeal condition included the two weak fear appeal advertisements from the Fear Appeal Person set and the Fear Appeal Collage set while the strong fear appeal condition contained the two strong fear appeal advertisements from Fear Appeal Person set and Fear Appeal Collage set. The pair that has the greatest difference in measured fear between the weak fear appeal advertisement and the strong fear appeal advertisement is chosen for the main experiment.
Fear-appeal measures.
A total of 5 items based on Ruiter and colleagues (2003) measured fear (see Appendix E). The PCA showed that the scale was uni-dimensional (E.V. = 3.13) that explained 78% of the total variance of the scale. Furthermore, the reliability of the scale is very good,
Cronbach’s α = .90 (M = 4.39, SD = 1.97). Personality traits
In order to make an inventory on personality traits, the pre-test also examined which personality traits participants identified with most. Additionally, personality traits used in self-affirmation tasks must be unrelated to the behavior one is trying to change for it to be successful (Good & Abraham, 2011). Thus, the pre-test also studied which personality traits participants found to be closely related to alcohol consumption. The personality traits from the inventory will be used in the self-affirmation manipulation in the main experiment. A total of 20 personality traits were included in the pre-test, which were inspired by the Big 5
Results Pretest Manipulation Check
The goal of this pre-test was to check whether the fear appeal conditions and the self-boost conditions were appropriately manipulated. Through several one-way ANOVA’s, manipulation checks were analysed. Firstly the self-boost manipulation checks are presented, and then the fear appeal manipulation checks are presented.
Self-boost manipulation checks
Self-affirmation manipulation check. Self-affirmation was measured via a one-way ANOVA with the self-boost conditions as the independent variable on self-affirmation. The analysis had an unequal variance as explained in the Levene’s test and thus was corrected for with the Welch test. This resulted in non-significant data among the three conditions, F (2, 59) = 1.58, p = .216, η2
= .051. Yet, although it was non-significant, self-affirmation had a slightly higher mean in the Self-affirmation Quiz condition (M = 7.01, SD = 1.02) than in Self-affirmation Puzzle condition (M = 6.59, SD = 1.63), as presented in Table 1. As a result, the Self-affirmation Quiz condition is chosen for the main experiment and will be made more extensive in the main experiment.
Efficacy manipulation check. It is assumed that the efficacy condition has a higher
score in response-efficacy than in the self-affirmation conditions. However, a one-way ANOVA with these self-boost conditions as the independent variable on response efficacy resulted in no significant differences for the three conditions, F (2, 59) = 0.03, p = .967, η2 = .001, see Table 1.
Self-efficacy was measured via a one-way ANOVA with the self-boost conditions as the independent variable on self-efficacy but the manipulation check resulted in no significant differences as presented in Table 1, F (2, 59) = 0.40, p = .670, η2
self-efficacy and response efficacy manipulations must be more extensive in the main experiment.
Table 1 - Statistics on self-boost manipulation: Pre-test Dependent variables Efficacy Condition M (SD) Self-affirmation Quiz Condition M (SD) Self-affirmation Puzzle Condition M (SD) F value (2, 59) p η2 Response- efficacy 4.70 (1.94) 4.78 (1.64) 4.62 (2.01) 0.03 .967 .001 Self- efficacy 7.93 (1.69) 8.00 (1.78) 7.50 (2.33) 0.40 .670 .013 Self Affirmation 7.25 (0.92) 7.01 (1.02) 6.59 (1.63) 1.58 .216 .051
Fear appeal manipulation check
Fear appeal person manipulation check. The fear appeal manipulations consisted of
two sets of two advertisements, namely one strong and one weak fear appeal. A manipulation check conducted via a one-way ANOVA with the fear appeal conditions as the independent variable on the Fear Appeal Person set had unequal variances as explained in the Levene’s test, and thus was corrected for with the Welch test. Participants found the strong fear appeal more frightening than the weak fear appeal in the Fear Appeal Person manipulation, F (1, 59) = 22.88, p < .001, η2
Table 2 - Statistics on fear appeal manipulation: Pre-test Dependent variables Strong Fear Condition M (SD) Weak Fear Condition M (SD) F value (1, 59) p η2 Person set: Fear Appeal 5.98 (1.43) 3.86 (1.96) 22.88 < .001 .279 Collage set: Fear Appeal 5.47 (1.80) 2.49 (1.40) 52.99 < .001 .473
|Fear appeal collage manipulation check. The second set, namely, Fear Appeal
Collage demonstrated a statistically significant effect of fear in the fear appeal conditions as well, F (1, 59) = 52.99, p < .001, η2= .473, as described in the one-way ANOVA with the fear appeal conditions as the independent variable and Fear Appeal Collage as the dependent variable. Thus, participants also found the strong fear appeal more frightening than the weak fear appeal in the Fear Appeal Collage manipulation. Although both manipulations were statistically significant, the Fear Appeal Collage had a slightly greater difference in means (2.98) than the Fear Appeal Person manipulation (2.12). In addition, Fear Appeal Collage has a greater effect size than Fear Appeal Person as demonstrated in Table 2. Hence, the Fear Appeal Collage manipulation had a greater effect on the amount of fear participants felt than the Fear Appeal Person manipulation. As a result, the Fear Appeal Collage manipulation is chosen for the main experiment.
Personality Traits
Through a frequency test, the personality traits that participants identified with most and the personality traits they closely identified to alcohol consumption were analyzed. A total of 10 personality traits will be used rather than the intended 6 in order to enhance the
Self-affirmation Quiz manipulation since it was non-significant in the pretest. These personality traits include sympathetic, efficient, imaginative, insightful, kind, trusting, organized, reliable, responsible, and forgiving as they had the least overlap with traits participants connected to alcohol consumption. Personality traits including active, assertive, energetic, enthusiastic, outgoing, generous, curious, original and artistic were excluded because they were more closely related to alcohol consumption as described through the high percentages demonstrated in Appendix F.
Method Main Experiment Study Design
A 3 (self-boost: efficacy message vs. self-affirmation message vs. no self-boost) × 2 (Fear appeal: strong vs. weak) between subjects experiment was conducted via an online questionnaire. Thus, the main experiment consisted of six conditions. The participants were either presented with a self-affirmation, an efficacy, or a control condition (no self-boost). Afterwards the participants either saw a strong or a weak fear appeal on alcohol consumption and breast cancer.
Respondents
For an explanation of the sample choice see the Method Pretest section. A total of 241 people started the online questionnaire for the main experiment. However, several
respondents were deleted as they were either male, under the age of 18, consumed no alcohol, or had participated in the pre-test. As a result 191 participants were included in the final sample. Most of the 191 female students selected high school as their highest degree of completed education. Furthermore, their age ranged from 18 to 32 (M = 21.81, SD = 2.37), and they consumed an average of 5 glasses of alcohol per week (M = 5.52, SD = 7.40). Lastly,
the nationalities of these participants came from 34 different countries, with the Netherlands (N = 92), Brazil (N = 20) and Germany (N = 14) being the largest groups.
Procedure and Measures
The materials for this experiment were selected based on the pre-test from which a health website was created solely for the purpose of this research. In addition, the
manipulations were integrated into the health website in order to reflect a real life health campaign. The cover story for the participants was to ask their opinion on the health website. The efficacy condition included tips on decreasing alcohol consumption which were
integrated nicely in the health website. In the self-affirmation manipulation, participants were told the website would provide personalized information, and thus they had to fill in a
personality test.
The website consisted of three pages, namely a Home page (see Appendix G), an Information page (see Appendix H) and a Breast Cancer page (See Appendix I). The self-boost conditions were represented in the Home page and the Information page. The Home page included a brief explanation of the correlation between alcohol consumption and breast cancer, while the Introduction page included either the efficacy manipulation or the self-affirmation manipulation. Finally, the Breast Cancer page consisted of the fear appeal conditions. Participants first saw the Home page, then the Information page, and lastly the Breast Cancer page. Yet, participants in the control condition only saw the Home page and then the Breast cancer page, as they were not presented with any self-boost manipulation. The participants were randomly and equally distributed among the conditions.
Self-boost Conditions
Self-affirmation condition
The self-affirmation condition was presented with the Home page and then an Information page. The self-affirmation manipulation presented on the Information page
reflects the Self-affirmation Quiz condition explained in the pre-test. However, in order to improve the manipulation, since the self-affirmation manipulation check in the pre-test showed non-significant differences among the conditions, 10 rather than 6 statements on personality traits were included. These personality traits included being kind, sympathetic, imaginative, insightful, trusting, efficient, organized, reliable, responsible and forgiving.
Efficacy condition
The efficacy condition also included both the Home page and an Information page. The Information page in this condition was presented with a manipulation that is similar to the efficacy condition explained in the pre-test, however, since it was not significant, it was extended by providing 12 tips on decreasing alcohol consumption rather than 7.
Self-boost measures
Self-affirmation measures. All manipulation checks were composed of a 9-point Likert-scale and presented at the end of the survey in order to prevent the influence of responses. The manipulation check for the self-affirmation condition consisted of the same 5 items as described in the pre-test. The PCA demonstrated that all the items measuring self-affirmation formed one cohesive scale, since only one component had an eigenvalue above 1 (E.V. = 3.90), which explained the total variance of the scale with 78%. In addition, the self-affirmation scale showed to be very reliable, Cronbach’s α = .93 (M = 6.95, SD =1.41).
Efficacy measures. The efficacy condition aimed to increase both response-efficacy as well as self-efficacy. Both the response-efficacy scale as well as the self-efficacy scale
consisted of 2 items as explained in the pre test. The PCA demonstrated that the scale for measuring response-efficacy was uni-dimensional as only one component had an eigenvalue above 1 (E.V. = 1.74), which explained 87% of the total variance of the scale. Furthermore, the scale also showed to be reliable, Cronbach’s α = .85 (M = 4.86, SD = 1.92).
The 2 items measuring self-efficacy presented one cohesive scale, demonstrated through a PCA. The scale only had one eigenvalue above 1 (E.V. = 1.59) that explained 79% of the total variance of the scale. Moreover, the reliability of the scale was acceptable, Cronbach’s α = .73 (M = 7.07, SD = 1.99).
Fear Appeal Conditions
The fear appeal was presented on the Breast Cancer page. Each condition was exposed to either a weak fear appeal or a strong fear appeal. The weak fear appeal condition consisted of the weak fear collage advertisement and the strong fear appeal condition
consisted of the strong fear collage advertisement as the pre-test demonstrated statistically significant differences between the two. In addition to the advertisement, the conditions were also presented with a text. Both conditions included similar information in the text however, the text in the strong fear appeal condition was written in a more serious manner. This tactic is influenced by research conducted by Ruiter, Verplanken, Kok and Werrij (2003).
Fear appeal measures
A scale composed of 4 items measured the extent to which a person felt fear when seeing either a weak fear appeal or a strong fear appeal. Since the fear appeal manipulation was successful in the pre-test, the manipulation check for the fear appeal was presented last in the survey for the main experiment. The scale measuring fear showed to be uni-dimensional as presented in the PCA. Only one component had an eigenvalue above 1, (E.V. = 1.59) which explained 74% of the total variance of the scale. The reliability of the scale was good,
Cronbach’s α = .88 (M = 4.22, SD = 1.91). Dependent Variable Measures
Message persuasiveness was tested via two variables, namely message acceptance and intention towards decreasing alcohol consumption. They were asked directly after the
alcohol consumption comprised of 3 items, which were inspired from research done by Harris and Napper (2005) (see Appendix J). The PCA showed that these 3 items formed a cohesive scale as only one component had an eigenvalue above 1 (E.V. = 2.76), which explained 92% of the total variance of the scale. Furthermore, the scale proved to be very reliable,
Cronbach’s α = .96 (M = 4.06, SD = 2.46).
The scale for message acceptance was composed of 4 items, which was inspired from research done by Good and Abraham (2011) (see Appendix K). The scale proved to be cohesive as the PCA showed that only one component had an eigenvalue above 1 (E.V = 2.95), which explained 74% of the total variance of the scale. The reliability of the scale was good, Cronbach’s α = .88 (M = 5.12, SD = 1.58).
Results Main Experiment Randomization Check
A comparative analysis was conducted in order to check whether age, alcohol consumption, nationality and education were equally distributed among the fear appeal conditions and the self-boost conditions. Firstly, a correlation analyses with the variables presented above and the dependent variables (message acceptance and intention) was
conducted which showed that alcohol consumption had a significant negative correlation with message acceptance, r(191) = -.218, p = .002 and intention r(191) = -.218, p = .003. However, a one-way ANOVA with the fear appeal conditions as independent variable on alcohol
consumption resulted in non-significant differences between conditions, F (16, 174) = 1.31, p = .197. Additionally, the one-way ANOVA with the self-boost conditions as independent variable on alcohol consumption also showed non-significant differences between conditions, F (16, 174) = 0.83, p = .656. As a result, it was not necessary to enter covariates because the variables were not significantly different among the conditions.
Manipulation Check
Before the manipulations checks were conducted, the data was explored for outliers (following Pallant, 2007). However, the trimmed means (data when outliers are deleted) altered at most with 0.1 from the original mean, thus the outliers were not deleted.
Manipulation checks were conducted through several one-way ANOVA’s to test whether the manipulations in the main experiment were successful. Table 3 presents data on the self-boost manipulations.
Self-boost manipulation checks
Self-affirmation manipulation check. The self-affirmation manipulation check had unequal variances as explained in the Levene’s test, and thus was corrected for with the Welch test. The affirmation condition was assumed to have the highest scores in the self-affirmation, yet the one-way ANOVA with the self-boost conditions as the independent variable on self-affirmation showed the data was not significant, F (2, 186) = 1.55, p = .214, η2
= .016.
Efficacy manipulation check. The efficacy condition was assumed to have higher scores in response-efficacy, however, the results from the one-way with the self-boost
conditions as the independent variable on response efficacy was not statistically significant, F (2, 186) = 2.24, p = .109, η2
= .024. Furthermore, self-efficacy was also assumed to be highest in the one-way ANOVA with the boost conditions as the independent variable on self-efficacy, however, this too was not statistically significant, F (2, 186) = 0.25, p = .777, η2= .003.
Table 3 - Statistics on self-boost manipulation: Main experiment Dependent variables Efficacy Condition M (SD) Self-Affirmation Condition M (SD) No Self-boost Condition M (SD) F value (2, 186) p η2 Response efficacy 5.27 (1.97) 4.74 (1.75) 4.61 (1.96) 2.24 .109 .024 Self-efficacy 7.16 (1.88) 7.14 (1.90) 6.94 (2.14) 0.25 .777 .003 Self Affirmation 6.74 (1.77) 6.91 (1.17) 7.16 (1.20) 1.55 .214 .016
Fear appeal manipulation check
Table 4 presents the information from the fear appeal manipulation check. Contrarily to the pre-test, the results from the one-way ANOVA with the fear appeal conditions as the independent variable on fear shows that there was no significant difference between results in the strong fear condition vs. the weak fear condition, F (1, 186) = 0.17, p = .684, η2
= .001. This is surprising as the same manipulation yielded statistically significant in the pre-test.
Table 4- Statistics on fear appeal manipulation: Main experiment Dependent variables Strong Fear Condition M (SD) Weak Fear Condition M (SD) F value (1, 186) p η2 Fear Appeal 4.16 (1.88) 4.28 (1.94) 0.17 .684 .001
As a result, since the manipulations for the experiment were not statistically
significant, conclusions derived on the effect of the independent variable and the moderator on the dependent variable must be interpreted with caution. Reasons to why the manipulation checks may not have been successful are described in the discussion.
Main effects & interaction effect
Two two-way ANOVA’s were conducted to test for the main effects and the
interaction effect of this research; namely to explore the extent to which an efficacy message vs. a self-affirmation message prior to an advertisement containing a strong vs. a weak fear appeal influences message persuasiveness (intention & message acceptance) among female students to reduce alcohol consumption.
A two-way ANOVA with the self-boost conditions and fear appeal conditions as the independent variables on intention to decrease alcohol consumption was conducted. The main effect of the fear appeal conditions on intention was not significant, F (1, 184) = 2.54, p = .113, η2
= .014. This means that people in the strong fear condition (M = 3.82, SD = 2.35) did not score significantly higher than the weak fear condition (M = 4.29, SD = 2.55), thereby rejecting Hypothesis 1a.
Moreover, the main effect of the self-boost conditions on intention was not significant, F (2, 184) = 1.21, p = .301, η2
= .013. Showing that the people in the efficacy condition (M = 4.43, SD = 2.46) and the self-affirmation condition (M = 3.98, SD = 2.64), did not score significantly higher than people in the no self-boost condition (M = 3.80, SD = 2.31), thus rejecting Hypothesis 2a and 3a.
The interaction effect between the fear appeal conditions and the self-boost conditions also was not significant, F (2, 184) = 1.37, p = .257, η2
= .015. This means that people in the strong fear condition (M = 4.04, SD = 2.53) compared to the weak fear condition (M = 4.84, SD = 2.36) did not score significantly higher when presented with an efficacy message, thus
rejecting Hypothesis 4a. In addition, those in the self-affirmation condition also did not score significantly higher when presented with a strong fear appeal (M = 3.42, SD = 2.40) compared to a weak fear appeal (M = 4.56, SD = 2.80), hence rejecting Hypothesis 4b. Lastly, it shows that those in the no self-boost condition did not score significantly less when presented with a strong fear appeal (M = 3.91, SD = 2.18) compared to a weak fear appeal (M = 3.69, SD = 2.45), consequently rejecting Hypothesis 4c. Although the interaction was non-significant, Figure 2 shows that the efficacy and the self-affirmation condition, when presented with a weak fear appeal result in a slightly higher mean for intention to decrease alcohol
consumption than in the no self-boost condition.
Fig. 2. Interaction effect of self-boost conditions and fear appeal conditions on intention to decrease alcohol consumption
Additionally, a two-way ANOVA with the self-boost conditions and the fear appeal conditions, as independent variables on message acceptance was conducted, which had
1 2 3 4 5 6 7 8 9
No Self-boost Efficacy Self-affirmation
Strong Fear Condition Weak Fear Condition
with caution. The main effect of the fear appeal conditions on message acceptance was not significant, F (1, 185) = 0.34, p = .560, η2
= .002. This demonstrates that the people in the strong fear condition (M = 5.06, SD = 1.61) did not score significantly higher than the weak fear condition (M = 5.19, SD = 1.56), thus rejecting Hypothesis 1b.
Furthermore, the main effect of the self-boost conditions on message acceptance was also not significant, F (2, 185) = 0.52, p = .595, η2
= .006. This shows that the people in the efficacy condition (M = 5.16, SD = 1.50) and the self-affirmation condition (M = 5.27, SD = 1.39), did not score significantly higher than people in the no self-boost condition (M = 4.99, SD = 1.76), thereby rejecting Hypothesis H2b and H3b.
The interaction effect between the fear appeal conditions and the self-boost conditions on message acceptance was not significant, F (2, 185) = 0.58, p = .560, η2
= .006. This shows that people in the strong fear condition (M = 4.92, SD = 1.29) compared to the weak fear condition (M = 5.41, SD = 1.68) did not score significantly higher when presented with an efficacy message, thereby rejecting Hypothesis 5a. In addition, those in the self-affirmation condition also did not score significantly higher when in the strong fear condition (M = 5.30, SD = 1.70) compared to the weak fear condition (M = 5.25, SD = 1.01), thus rejecting
Hypothesis 5b. Finally, it means that those in the no self-boost condition did not score significantly less when in the strong fear condition (M = 5.01, SD = 1.80) compared to the weak fear condition (M = 4.97, SD = 1.74), therefore rejecting Hypothesis 5c.
While the interaction effect for message acceptance was non significant, it is
interesting to still look at the overall trend presented in Figure 3. Figure 3 illustrates that the self-affirmation condition with both a weak fear condition resulted in slightly higher means for message acceptance than in the no self-boost condition.
Fig. 3. Interaction effect of self-boost conditions and fear appeal conditions on message acceptance
Conclusion and Discussion
The aim of this research was to examine the relationship between fear appeals and self-boost messages and how they together influence message persuasiveness regarding alcohol consumption and breast cancer. It was hypothesized that the presence of strong fear appeals will result in greater intention (H1a) and message acceptance (H1b) to decrease alcohol consumption than the presence of weak fear appeals. However, the main effect of fear appeals on message persuasiveness (intention and message acceptance) was not significant, hence, H1a and H1b are not confirmed.
In addition, it was hypothesized that presenting an efficacy message will result in greater intention (H2a) and message acceptance (H2b) in decreasing alcohol consumption than presenting a no self-boost message. Yet, this yielded non-significant results, and thus, H2a and H2b cannot be accepted. Furthermore, it was hypothesized that presenting a self-affirmation task will result in greater intention (H3a) and message acceptance (H3b) in
1 2 3 4 5 6 7 8 9
No Self-boost Efficacy Self-affirmation
Strong Fear Condition Weak Fear Condition
showed non-significant results, thus, H3a and H3b are not accepted. As a result, the main effect of a self-boost on message persuasiveness was non-significant. Therefore, this study cannot determine whether an efficacy message or a self-affirmation message leads to greater message persuasiveness (RQ1).
Lastly, it was hypothesized that when people see an efficacy message or a self-affirmation message it results in greater intention and message acceptance with a strong compared to a weak fear appeal. Yet when people see a no self-boost message it results in smaller intention and message acceptance with a strong compared to a weak fear appeal. However, since the interaction effect showed non-significant results, the hypotheses on the interaction effect (H4a, H4b, H4c, H5a, H5b, H5c) cannot be confirmed.
One reason why the main effects and interaction effects of this study were non-significant is because the manipulations did not non-significantly increase efficacy or
affirmation. As a result, this study demonstrates the difficulty of increasing efficacy and self-affirmation through communication research performed outside of the lab. Previous research by Ruiter and his colleagues (2014) describe the challenge of increasing a person’s efficacy, which is reflected in this study. However, this study adds to current research by showing that one’s self-affirmation is difficult to increase as well. Most of the previous research that significantly increased a persons’ self-affirmation was through lab experiments (McQueen & Klein, 2006). This demonstrates that theories on self-affirmation (The Self-Affirmation Theory) and efficacy (EPPM) together with fear appeals are difficult to apply in
communication research.
In addition, this study gives an insight on how to integrate efficacy and
self-affirmation manipulations within a fictional health website, which can be improved in future research. An idea to improve the health website is to include videos as previous research shows that videos are more effective than text when it comes to changing behavior (Stanczyk
et al., 2014). In addition, a possibility for a follow up study is to still integrate efficacy and self-affirmation manipulations within a health website, but to have participants complete the experiment in isolated rooms. Then they may be more focused and have fewer distractions than in their home environment.
Another reason to why this experiment may benefit from isolated rooms is the following. Previous research has shown that since students generally consume a lot of alcohol, they tend to respond defensively towards messages against alcohol consumption (Brown & West, 2014). Thus, this may be a reason to why participants only spent a few seconds viewing the manipulations. The average amount of seconds participants spent on the efficacy manipulation was 16, 175 on the self-affirmation manipulation, 33 on the strong fear manipulations and 46 on the weak fear manipulation. Potentially participants spent longer on the self-affirmation manipulation since they had to write a short text with a minimum amount of 20 characters. Future research can see whether more time is spent on each manipulation if it is conducted in an isolated room, and whether this leads to significant manipulations.
Another limitation for this study is regarding the personality traits in the
self-affirmation manipulation. Jessop and colleagues (2009) demonstrated in their research that the positive traits affirmation leads to greater self-affirmation. One reason to why this was not the case for this study was perhaps due to the personality traits that were chosen in the affirmation manipulation. Only 13 of the 51 participants who were exposed to the
self-affirmation manipulation affirmed to all the 10 positive traits listed. Future research could use even broader terms for personality traits so they could be applicable to more people.
Confidence or self-esteem may be one reason to why only 13 people affirmed to all of the positive traits presented. Confidence may be an issue with decreasing one’s alcohol consumption, since it can be an addiction, and those who do try to limit their alcohol
consumption may differ per person, a recommendation for future research is to measure this. Thus, confidence can be measured before the manipulations are shown to see whether people with for instance, a high level of confidence have different results than those with a low level of confidence.
Overall, it is important to consider that although the manipulations for this experiment were non-significant, they may still have influenced the participants in a manner that was not measured with the scales used in the main experiment. The only manipulation that was significant in this research was the fear appeal manipulation in the pre-test. One possible reason to why this differed in the main experiment is that, due to the successful manipulation of the fear appeals in the pre-test, fear was the last construct that was measured in the main experiment. As a result, it may have decreased the participants’ recall of the fear appeal.
In conclusion, this research demonstrates the challenge to increase efficacy and self-affirmation through communication research. However, with the ideas provided in the discussion, future research may be able to improve the manipulations and method from this study. We advise scientists to continue further research on self-affirmation and find ways for it to be implemented in real health campaigns. By having people affirm to positive personality traits, it increases a person’s incentive to act morally correct (Steele, 1988). As a result, people may perform a health recommendation, believing it is best for oneself. Increasing one’s self-affirmation can decrease one’s feelings of threat and increase confidence; as a result, this tactic could be very influential in real health campaigns.
References
Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44(9), 1175.
Bandura, A. (2001). Social cognitive theory of mass communication. Media Psychology, 3(3), 265-299. doi:10.1207/S1532785XMEP0303_03
Brown, S. L., & West, C. (2014). Sequencing the threat and recommendation components of persuasive messages differentially improves the effectiveness of high‐and low‐
distressing imagery in an anti‐alcohol message in students. British Journal of Health Psychology, 20(2), 324-340. doi:10.1111/bjhp.12103
Falk, A., & Heckman, J. J. (2009). Lab experiments are a major source of knowledge in the social sciences. Science, 326(5952), 535-538. doi:10.1126/science.1168244
Floyd, D. L., Prentice‐Dunn, S., & Rogers, R. W. (2000). A meta‐analysis of research on protection motivation theory. Journal of Applied Social Psychology, 30(2), 407-429. doi:10.1111/j.1559-1816.2000.tb02323.x
Fry, R. B., & Prentice-Dunn, S. (2005). Effects of coping information and value affirmation on responses to a perceived health threat. Health Communication, 17(2), 133-147. doi:10.1207/s15327027hc1702_2
Good, A., & Abraham, C. (2011). Can the effectiveness of health promotion campaigns be improved using self-efficacy and self-affirmation interventions? An analysis of sun protection messages. Psychology & Health, 26(7), 799-818.
doi:10.1080/08870446.2010.495157
Harris, P. R., & Napper, L. (2005). Self-affirmation and the biased processing of threatening health-risk information. Personality and Social Psychology Bulletin, 31(9), 1250-1263. doi:10.1177/0146167205274694
Holmes, E. A., & Mathews, A. (2005). Mental imagery and emotion: A special relationship?
Emotion, 5(4), 489.
Jessop, D. C., Simmonds, L. V., & Sparks, P. (2009). Motivational and behavioural consequences of self-affirmation interventions: A study of sunscreen use among women. Psychology and Health, 24(5), 529-544. doi:10.1080/08870440801930320 Leventhal, H. (1971). Fear appeals and persuasion: the differentiation of a motivational
construct. American Journal of Public Health, 61(6), 1208-1224.
Lewis, L. (2013). Chapter 20. A Fool for Love (pp.87). North Carolina: Lulu Publishing Services.
Longnecker, M. P. (1994). Alcoholic beverage consumption in relation to risk of breast cancer: meta-analysis and review. Cancer Causes & Control, 5(1), 73-82.
Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19(5), 469-479.
McQueen, A., & Klein, W. M. (2006). Experimental manipulations of self-affirmation: A systematic review. Self and Identity, 5(4), 289-354. doi:10.1080/15298860600805325 Murphy, J. G., McDevitt-Murphy, M. E., & Barnett, N. P. (2005). Drink and be merry?
Gender, life satisfaction, and alcohol consumption among college students. Psychology of Addictive Behaviors, 19(2), 184. doi:10.1037/0893-164X.19.2.184
Napper, L., Harris, P., & Epton, T. (2009). Developing and testing a self-affirmation manipulation. Self and Identity, 8, 45–62. doi:10.1080/15298860802079786
Niaura, R. S., Rohsenow, D. J., Binkoff, J. A., Monti, P. M., Pedraza, M., & Abrams, D. B. (1988). Relevance of cue reactivity to understanding alcohol and smoking
O'Malley, P. M., & Bachman, J. G. (1983). Self-esteem: Change and stability between ages 13 and 23. Developmental Psychology, 19(2), 257. doi:10.1037/0012-1649.19.2.257
Pallant, J. (2007). SPSS survival manual: A step by step guide to data analysis using SPSS for
windows. New York: Open University Press.
Peters, G. J. Y., Ruiter, R. A., & Kok, G. (2013). Threatening communication: A critical re-analysis and a revised meta-analytic test of fear appeal theory. Health Psychology Review, 7(sup1), S8-S31. doi: 10.1080/17437199.2012.703527
Reed, M. B., & Aspinwall, L. G. (1998). Self-affirmation reduces biased processing of health-risk information. Motivation and Emotion, 22(2), 99-132.
Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. The Journal of Psychology, 91(1), 93-114.
Ruiter, R. A., Kessels, L. T., Peters, G. J. Y., & Kok, G. (2014). Sixty years of fear appeal research: Current state of the evidence. International Journal of Psychology, 49(2), 63-70. doi:10.1002/ijop.12042
Ruiter, R. A., Verplanken, B., Kok, G., & Werrij, M. Q. (2003). The role of coping appraisal in reactions to fear appeals: Do we need threat information?. Journal of Health
Psychology, 8(4), 465-474. doi:10.1177/13591053030084006
Sherman, D. A., Nelson, L. D., & Steele, C. M. (2000). Do messages about health risks threaten the self? Increasing the acceptance of threatening health messages via self-affirmation. Personality and Social Psychology Bulletin, 26(9), 1046-1058.
doi:10.1177/01461672002611003
Sherman, D. K., & Cohen, G. L. (2002). Accepting threatening information: Self–Affirmation and the reduction of defensive biases. Current Directions in Psychological
Sherman, D. K., & Cohen, G. L. (2006). The psychology of self-defense: Self-affirmation theory. Advances in Experimental Social Psychology, 38, 183. doi:10.1016/S0065-2601(06)38004-5
Snipes, R. L., LaTour, M. S., & Bliss, S. J. (1999). A model of the effects of self-efficacy on the perceived ethicality and performance of fear appeals in advertising. Journal of Business Ethics, 19(3), 273-285. doi:10.1023/A:1005822414588
Stanczyk, N., Bolman, C., van Adrichem, M., Candel, M., Muris, J., & de Vries, H. (2014). Comparison of text and video computer-tailored interventions for smoking cessation: Randomized controlled trial. Journal of Medical Internet Research, 16(3), e69.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the self. Advances in Experimental Social Psychology, 21(2), 261-302.
Stone, J., & Cooper, J. (2003). The effect of self-attribute relevance on how self-esteem moderates attitude change in dissonance processes. Journal of Experimental Social Psychology, 39(5), 508-515. doi:10.1016/S0022-1031(03)00018-0
Sutton, S., & Hallett, R. (1988). Understanding the effects of fear-arousing communications: the role of cognitive factors and amount of fear aroused. Journal of Behavioral
Medicine, 11(4), 353-360.
Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health, Education & Behavior, 27(5), 591-615.
doi:10.1177/109019810002700506
Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communications Monographs, 59(4), 329-349.
Appendix A Self-affirmation Items
Variable Items Min/Max score Source: changed
IV/ DV accordingly
Self-affirmation 1. How do you currently feel about yourself? 2. I take a positive attitude toward myself 3. I feel I am a person of worth 4. I am able to do things as well as most other people 5. On the whole, I am satisfied with myself 1= Extremely negative 9= Extremely positive 1= Completely disagree 9= Completely agree (Sherman et al., 2000) (O'Malley & Bachman, 1983).
Appendix B Response-efficacy Items
Variable Items Min/Max score Source: changed
IV/ DV accordingly Response-efficacy 1. Decreasing the amount of alcohol I consume will decrease my chances of getting breast cancer 2. Decreasing the amount of alcohol I consume is an effective way to reduce the risk of developing breast cancer 1= Totally disagree 9= Totally agree (Ruiter et al., 2003)
