Fear appeals are still being used in today’s social marketing world to persuade the public to change dangerous habits. The use of fear appeals has proved effective in changing behaviour with moderate to high levels being most successful at persuasion. Other variables need to be taken into account however to ensure that fear appeals are not focussing just on the optimum levels of fear but on recipients themselves. A literature review of fear appeals which assesses the effectiveness of fear appeals in health promotion campaigns by looking at four themes; Level of fear, self-efficacy, social threat and source credibility. A further theme detailing the frameworks in the literature will look at models introduced such as the Parallel Response and Protection Motivation Theory. The predominant methodologies used in the literature are also discussed; the review output takes the form of a timeline which details key events and papers found in the literature.
Of Fear Appeals in Health Promotion Campaigns
Over the years there has been much written about the effectiveness of fear appeals as a social marketing exercise. A number of theories and subsequent models have been constructed by various experts in the field of both marketing and psychology to try to piece together the process. Various reviews of the literature have been carried out over the last fifty years emphasising the different stages of growth there has been in the subject. Preliminary research showed fragmentation, like all new subject areas to research however in the last decade the field has reached some stability and consensual understanding based around a set of models and ideas.
The effectiveness of fear appeals in health promotion campaigns is looked at due to the wealth of fear appeal literature relating to this area of study, there are exciting views and conflicting ideas which can be explored. A number of health promotion topics have been looked at in the literature from the earliest of dental hygiene and automobile safety, to cigarette smoking and HIV/AIDS. However with the broad scope of different areas, subjects researched and different time periods the literature has become fragmented with lots of different views of the effectiveness of the fear arousing messages.
This literature review will look at some of the literature surrounding fear appeals and attempt to suggest differing opinions on the effectiveness of using fear appeal as a social marketing campaign. By looking at a number of themes which are present in the literature one can begin to understand the different dimensions within the literature with regard to key theories and models. Fear appeals are used primarily as a persuasion technique to persuade recipients into changing their undesirable behaviour in favour of that suggested in the appeal.
Levels of Fear
The initial literature of fear appeals emerged in 1953 (Janis and Feschbach, 1953) when research focussed on drive theory models used in other academic fields to understand the reactions from fear arousal and whether it was effective. The drive model theory assumes that when the communication arouses fear the recipient will become highly motivated to try various responses to change their unpleasant state. . This pointed to a suggested inverted U-shaped model of fear arousal in relation to attitude change, with the highest point of the model in the centre being the middle (moderate) and most desired position when communicating to the source through fear (McGuire, 1968). Optimal fear level is reached when the interfering effects (the critical evaluation by the recipient and recommended action) start to increase at a faster rate than facilitating effects (the recipients’ motivation to find a means to averting danger).
Insko et al (1965) found that for smoking cessation a high level of arousal was seen to be most effective at persuasion of the message which was a finding shared by Leventhal & Watts (1966) who initially found that a moderate level of fear was effective at persuading the subjects to take X-Rays immediately after exposure, however in the findings suggested that more subjects exposed to a high fear arousal message had reduced or given up altogether.
Further research has suggested that an optimal fear level is not the case and that different subjects will react to fear appeal with different levels of fear, whether it be high, moderate or low (Higbee,1969).. A fear level which is too low will brand an appeal ineffective and not able to communicate the message however a fear level which is too high may lead to denial of the fear appeal with thoughts which suggest other reasons for the undesired effect.
One study by Keller (1996) also looks at anti-smoking however links the level of fear to an elaboration response to either the problem or solution. The suggestion of that the effectiveness of a fear appeal is not just related to the level of fear, but how well both the problem (i.e. Lung cancer) and solution are elaborated upon in the message. Keller (1996) suggest low fear levels may not be effective because there is insufficient elaboration on the harmful consequences and also that high level fear appeals may also be ineffective because there is too much elaboration. Janis and Feshbach (1953) hinted at the importance of elaboration in persuasion and this study supports this research suggesting that to have a successful appeal it is not necessarily down to the level of fear employed but the amount of additional information is given about the problem and solution.
As Rotfield (1999) synthesises in the literature, fear is an emotional response and therefore all individuals are different and will harbour different levels of individual fear than others may to the same appeal. The “optimal level of fear” is hindering the job of social marketers by suggesting that one level of fear is best for the whole audience of a fear appeal. Each audience will react differently to fear and so social marketing may need to look at other variables rather than simply suggesting a level of fear is best.
Fear Appeal Frameworks
The first sign of a model based on how we view fear appeals today was developed in the 1970s, with Leventhal’ parallel response model which was an alternative to the fear drive model and assumed two different variables Danger control and Fear control cognitive response work in parallel to each other rather than a sequential model (Leventhal, 1971). Danger control looks at the recipients’ ability to control the threat through behaviour and coping methods in the environment. Fear Control looks at the internal situation and cues of the recipients’ ability to handle fear through their emotions and behaviour. This model is seen to separate the emotions, fear control, from cognitive responses, danger control. The danger control aspect suggests that the message is used for problem solving and for guiding the adaptive behaviour of the recipient. Fear control may encourage behaviours of controlling the fear such as eating, drinking or resting to cope with an emotional response however it has no relation to evoke action toward the threat ( Sternthal,1974). The two processes are not related however one can influence the other, thus a high fear appeal will evoke more fear control which will therefore affect the behaviour of a danger control response.
Although not fully highlighting exactly when fear and danger control responses are evoked the model does show more robustness as a model in relationship to the simpler drive theory.
Even though the parallel response model changed thoughts and introduced a distinction of emotion cognition to fear appeals, due to the lack of evidence by Leventhal (1971) to support his theory an extended parallel process model (EPPM) emerged introduced by Witte (1992) which extended the initial parallel response model and made it more accessible. Drawing similarity to the original model the EPPM suggests that perceived threat contributes to the extent of a response to a fear appeal (i.e., how strong the danger or fear control responses are) whereas perceived efficacy (or lack thereof) contributes to the nature of the response (i.e., whether danger or fear control responses are elicited). If no information with regard to the efficacy of the recommended response is provided, individuals will rely on past experiences and prior beliefs to determine perceived efficacy.
Rogers (1975) protection motivation theory (PMT) can be viewed as a model which is still used in research to date (Tanner, 1991). The model assumes that emotional arousal is less important that cognitive appraisal of the threat and people will more likely accept advice when they can be convinced of the threats seriousness and susceptibility of the message (****).The theory initially involved only three variables , magnitude of a threat, probability of the threat and the effectiveness of coping. The PMT was further developed by Rogers (1983) to four factors; Threat appraisal, which incorporates the severity of the threat and the probability of it occurring (susceptibility). As well as the coping appraisal, this is the cognitive process of an individual. Response Efficacy is the individuals’ ability to cope with or control the threat message and Self-efficacy, which will be looked at later in this review, is the expectation of the individual to perform the coping action successfully. It is suggested the PMT will result in one of three responses to the fear appeal; no response, acceptance or rejection.
Mewborn and Rogers (1979) detested the theory and suggested that there was no multiplicative relationship between the variables outlined in the PMT, however did find that an increase in efficacy of a recommended coping response would increase intentions to comply with the suggested behaviour. Therefore, suggesting that it is not to do with the occurrence of perceived severity but the level of recommendation in the message.
There are limitations to these frameworks and although some are still used in studies today, they still need to be further developed for adaptation of present day.
This variable added to the protection motivation theory in more detail by Rogers (1983) engaged in more cognitive thinking of how an individual will react to a fear appeal message. Self efficacy of an individual is involved with their competence and belief that they have the power to help themselves in completing a task of activities to prevent a threat, it is a terminology originally discussed by Bandura (1977). Self efficacy in the fear appeal literature was introduced with Rogers (1983) protection motivation theory. The fourth variable in his theory, the individuals’ perceived ability to coping behaviour, was directed towards self-efficacy (Rogers, 1983). High levels of self efficacy will promote an adaptive response whilst low levels of self efficacy with promote a maladaptive response. High self efficacy responses harbour thoughts of being able to give an effective response to the threat, thoughts of avoidance and not being able to deal with the threat shows a low self-efficacy response.
Self-efficacy is an individual behaviour of the recipient and although similar to self esteem and self image is a judgement of personal capability to change behaviour (Mailbach, 1995) and not how individuals see themselves. Those who are not taking part in an undesirable behaviour are more likely to be high in self-efficacy as the appeal will be preventative to them; however those who partake in the behaviour are more likely to have lower levels of self-efficacy because
Ripptoe and Rogers (1987) study of breast examination screening showed a relationship between responses to a fear appeal and self efficacy. High self efficacy appeals (those appeals which promote self-awareness to the adaptive behaviour) showed a positive feeling towards coping with the message and being able to deal with it whereas the low self-efficacy appeal evoked thoughts of hopelessness to breast cancer in the recipient. Self efficacy has been closely correlated with response efficacy through the literature (Cauberghe et al, 2009) this study focuses on speeding drivers and shows that high level of threat in the fear appeal results in lower levels of perceived self-efficacy for this subject, this contrasts with Witte and Allen (2000) who conclude in their meta-analysis that high levels of fear are best to promote self efficacy as long as the fear appeal includes skills and suggestions of recommended action of how best to deal with the threat which complements the work of Snipes (1999).
It is difficult to simply suggest a person has high self-efficacy or has low self-efficacy, its determinants are fragmented and all individuals behave differently in different situations (Maddux & Rogers, 1983). Someone who may show high levels of self-efficacy in one situation may not in another. Thompson (2009) hints at the deeper issues of self-efficacy, with social society and how certain demographics may have no alternative but to have low self-efficacy because they have a feeling that nobody is around to help them, helpline numbers are just not enough for some ‘high risk’ subjects. Hastings (2004) adds that the process of fear appeals may be more damaging for those most vulnerable i.e. those who partake in undesired behaviours and have low self-efficacy about adopting an adaptive response. This is where the ethics of fear appeals come into play and whether it is ethically sound to encourage fear in those who are already fearful.
Snipes (1999), looks at the ethicality of using fear appeals and high and low self-efficacy. He suggests that a strong fear appeal can be ethical if information is given on how to change behaviour and the audience are self-efficaous in their belief system, rather than a threat message being delivered without any real recommendation or information on how to change the undesirable behaviour (Snipes, 2009)
In some studies (Keller, 1999, Prentice-Dunn et al 2001) the positioning of the protection motivation theory is questioned and whether it is more effective to have the efficacy variables (response efficacy & self-efficacy), which recommend action, before the threat message in an appeal rather than the opposite. In the study of students and contraception it showed that the conventional structure of threat followed by efficacy was effective for the unconverted group but not for those targeted by the appeal. Prentice-Dunn (2001) opposes these findings and suggests that the sequential flow of the message makes no difference. However these two studies did not use control groups in their tests and so lack credibility. There is further evidence that the conventional threat before efficacy message is more effective than the reversed format (Hall, 2006) in her study of women smokers and cervical cancer. The later two studies both dealt with female subjects therefore suggesting that different groups will react differently to the model. It is clear that self-efficacy needs to be looked at further as a variable on its own rather than just within the protection motivation theory.
A look at the literature reveals that fear appeal messages work most effectively when they are given to an audience who have high levels of self-efficacy and are able to respond to the appeal positively rather than those who have low levels of self-efficacy and chose to ignore the appeal. This should be understood by social marketers when they are pursuing health promotion campaigns and who to target the appeal towards for maximum success.
Historically findings of fear appeals research have ignored the emotional response of an individual in accordance with literature which suggests it is not relevant (Rogers, 1983).to the psychology of fear appeals, opting for more cognitive processes instead. Physical threats have been more widely used to evoke a response in research recipients. This is acceptable and much research has successfully proven that a physical threat (i.e. smoking will cause lung cancer, speeding will cause you to eventually crash) is effective in evoking an adaptive response.
The fear appeal literature has begun to explore the use of emotional responses and social implications with more detail and although research is scarce it has been growing with more emphasis on social threat of the recipients to undesirable behaviours. Evans et al (1970) was the first to compare physical and social threats with his study of college students although the messages of physical threat were most effective in the study, his variable of social approval suggested that actual behaviour change was more effective in the long term.
Burnett and Oliver (1979) discovered that the level of fear used is segment specific and levels of fear which work for one segment may not be effective for others. Complementing this, Tanner et al (1991) suggests that social implications of a message could be more important to those in the younger generation therefore playing a more important role in their behaviour change Adolescents live their lives more emotionally than adults and will therefore feel alienated from a physical threat message. Although Henley and Donovan(2003) found that once exposed to a threat of very high fear (of death) adolescents showed signs of reaction to the threat it is generally suggested that adolescents have a high feeling of immortality and do not have a high level of risk in general and so a long term message discussing what will happen in the future will have a limited affect in this target group.
The fear appeal literature has begun to explore the use of emotional responses and social implications with more detail, Schoelenbachler and Whittler (1996) explored social threat in their study of drug prevention appeals suggesting that a social threat messages (i.e. Smoking will make you socially undesirable) are used in other advertising for example personal hygiene but have been lacking in fear appeals. The study showed that their subjects were less likely to ever use drugs after seeing the social threat message therefore showing that a social threat message evoking the emotional response of recipients was more effective in persuading the audience to stop than a physical message was.
The aim of the fear appeal is to persuade the recipient to have an adaptive response (be persuaded by the message and change behaviour) to the fear appeal and change their behaviour. Dickinson (2008) suggests that physical threats are better at inducing a greater level of fear but are unable to trigger an adaptive response whereas a social threat is more successfully at this, even though studies have found it to produce the lowest level of emotional response.
Some studies (Smith & Stutts, 2003) highlight the gender difference in responses to fear appeal. This study looked at college students, their smoking behaviour and long term versus short term health messages. The findings suggest physical threats are more effective than social threats and that the use of long-term messages is still effective, but only when in relation to females. Short term cosmetic fear appeals featuring negative consequences (bad breath, yellow teeth) were more relevant to males and less effective with the females. This difference in gender suggests that social marketers could target different audiences with their fear appeal campaigns in order to get the most success in persuasion.
The international dimension of social threat is discussed by Laroche (2001) who conducted research of both Chinese and Canadian subjects towards smoking which proved that Canadian subjects reacted to the social threat appeal whilst the Chinese subjects did not. The differences between cultures shows that either a social or physical threat is not the same for every culture and that for effective fear appeals there must be full understanding of the audience and how they will react to the persuasive message;
The social threat variable in fear appeals is one which is becoming more prominent; from the literature we can see that in order for social marketers to be successful in their campaigns they need to take into account the actions of certain groups (adolescents) and their thoughts and feelings, including the most effective way to convey a message to them. A physical threat, although effective for some groups, may lead to lower levels of persuasion than a social threat message.
Along with themes such as level of fear and other suggested in the frameworks the effectiveness of a fear appeal of persuading subjects to change behaviour through persuasion can also be down to the source of the message and whether it is credible enough to communicate that message. The theme of source credibility has been widely researched, (Hovland and Weiss 1951; Dholakia and Sternthal 1977; Ohanian, 1991) in a number of areas and different variables (Haley, 1996).
It is suggested that assuming that a high credibility source is more persuasive than a low credibility source should not be continued and that other variables such as source message and time should be taken into account (Sternthal et al 1977) Studies suggest that when an individual has a pre-disposed positive attitude toward a message then a low or moderate source credibility is acceptable. Similarly, when individuals have a negative pre-disposed attitude to the message high source credibility is more appropriate (Sternthal, 1978). This study looks at source credibility in terms of cognitive response and reactions to source credibility from the offset of a message i.e. Reaction to the message is more positive when the source is specified at the beginning of the message rather than at the end (Sternthal, 1978) When a message source is better recalled, the perceived credibility of the source is more likely to exert an impact on attitudes.
Petty & Cacioppo’ (1981) Elaboration likelihood model (ELM) is referred to in some of the literature around source credibility and that the effect of perceived source credibility on persuasion is of a direct nature and not involved with cognition. The ELM is a model produced to understand attitudes toward persuasion through elaboration via either a central route, unique positive cognitive responses determine the outcome; or peripheral route, involved with source credibility or attractiveness (Petty & Cacioppo, 1981). Jones et al (2003) studied the effect of source credibility in relation to the ELM and exercise promotion campaign and found that a positive message and credible source resulted in higher levels of elaboration with more positive exercise intentions over the two week period. The limitation of this study was its length of only a 2 week time period other studies have suggested that over time the source credibility may change.
Expertise, attractiveness and trustworthiness have found to be key variables in establishing the source credibility of a fear appeal (Sternthal, 1978). There has been some questioning of this and whether trustworthiness really is a variable which has a positive relationship with persuasion (Weiner and Mowen, 1986). Weiner and Mowen’ findings suggest that trustworthiness and expertise are important variables when discussing source credibility finding those who were trusted and with higher levels of expertise had more of an impact on the effectiveness of the persuasion. The study also highlighted the bias that may be present with the source; a highly credible source may be bias towards their message and therefore present it in more favourable way than it really is.
The literature search began with a simple search engine result of keywords including, “fear appeal” “Threat appeal” “health promotion” “health prevention” etc. this lead to the discovery of key papers and a timeline of the advancement of the literature. From key papers (Janis and Feshbach, 1953: Leventhal, 1971: Rogers, 1975: Witte, 1992: Rogers and Maddux,1983: Sternthal, 1978) search was then focussed on what these papers discussed and further search from these papers to look at meta-analysis (Witte and Allan, 2000) and other literature review’s (Higbee, 1969: Sternthal & Craig, 1974) of the topics.
The key papers allowed an understanding of the topic through both e-library and Social Sciences Citation Index the citing and cited references were found. Any journal articles which were relevant were added to the reading list. After reading the articles themes became apparent in the literature some of which had been highly discussed i.e. Self-Efficacy and others which had been looked at in only minimal detail (social threat). The literature search included articles from psychology and marketing related articles with an additional amount related to the health sciences. The search lead to other areas of discussion including the literature about source credibility which did not just focus on fear appeals but other subject topics.
The methodologies used by the authors of the articles changed over time. Qualitative data was used to gain understanding of the subjects i.e. through questioning. One can see that data is qualitative as the results of the findings would differ if done by others dependant on questions asked.. The subjects were gathered in focus groups for data collection, which included the viewing of the fear appeal messages. The subjects answered questions and viewed fear appeals of low, moderate and high fear arousal. In some cases a control group was used in the study of which only one variable being tested was different, used to add weight to the research. Quantative data was used also in research in response to analysising findings with manipulation checks and variables in a majority of research.
From the review a timeline of key events can be plotted to have a visual conceptual framework of summary. The timeline begins in the 1950’s and progresses to more recent findings of the fear appeal literature in the 2000’s. Although not directly linked to the question in hand it does provide a diagrammatic review of what has been occurring over the last fifty years in the field. This review output can be found in appendix 1 which follows from the conclusion of this literature review.