Article summaries

Effects of Suicide Awareness Material on Implicit Suicide Cognition: A Laboratory Experiment Florian Arendta, Benedikt Tillb and Thomas Niederkrotenthalerb

aDepartment of Communication Science and Media Research, University of Munich; bSuicide Research Unit, Institute of Social Medicine, Center for Public Health, Medical University of Vienna

ABSTRACT In spite of the increasing adoption of suicide awareness campaigns to prevent suicide, little is known about the effective construction of awareness messages used and on their impact on suicidal cognition. We hypothesized that media reporting on an individual overcoming a suicidal crisis increases the automatic association between “life” and self. University students (N = 112) were randomly allocated to one of three groups in a laboratory experiment. Participants allocated to treatment group 1 or group 2 read awareness material about a person coping with suicidal ideation by getting professional help. The only difference between the two groups was the amount of social similarity (low vs. high) between the protagonist and the participants. The control group read an article unrelated to suicide. Awareness material increased implicit cognition in terms of a strengthening of self-life associations. This effect was restricted to participants scoring low on wishful identification with the suicidal protagonist. This finding suggests that only individuals who do not wishfully identify with a protagonist going through difficult life circumstances benefit from the awareness material in terms of suicidal cognition. These findings provide a rich basis for further research and have potentially high relevance to the construction of suicide-awareness messages.

Suicide is a considerable public health problem. There are estimates that approximately 1 million people worldwide die from suicide each year. This is in spite of the fact that suicide is preventable (World Health Organization, 2012). Within a public health framework of suicide prevention, the mass media are considered to be a key factor due to their potentials for increasing awareness and educate the public on suicide prevention. However, there are also risks for public health when suicide is reported in the mass media.

An increasing amount of literature suggests that sensation- alist portrayals of suicide, particularly celebrity suicide reports, are associated with subsequent increases in suicides (Fu & Yip, 2009; Niederkrotenthaler et al., 2009, 2010, 2012; Phillips, 1974; Pirkis & Blood, 2001a; Stack, 2005). This phe- nomenon of a spike of suicides after publicized suicide stories has been termed the “Werther effect” by Phillips (1974) and subsequent researchers (see Niederkrotenthaler et al., 2010). Most theoretical models focus on social learning theory (Bandura, 1977) in order to explain this effect. Researchers typically argue that individuals in adverse psychosocial cir- cumstances may learn from these stories that suicide was an appropriate way to solve problems and may think that suicide would be a solution for their problems, as for the suicide case portrayed (see Stack, 2003). Based on these considerations, national agencies and the World Health Organization (WHO) have developed media recommendations to assist journalists in the reporting on suicides (Pirkis, Blood, Beautrais, Burgess, & Skehans, 2006; World Health Organization, 2008). Studies

show that these recommendations have resulted in improve- ments in the quality of suicide-related reporting in several countries (Etzersdorfer, 1998; Michel, Frey, Wyss, & Valach, 2000; Pirkis et al., 2009), and may even have contributed to a decline in suicide rates (Etzersdorfer & Sonneck, 1998; Niederkrotenthaler & Sonneck, 2007).

However, media effects may not be restricted to harmful effects. A recent study showed an association of print-media articles on positive coping with adverse circumstances with decreases in suicide rates subsequently. This potential suicide-protective effect was termed the “Papageno effect” (Niederkrotenthaler et al., 2010). This study, together with previous literature (Fu & Chan, 2013; Stack, 2005), highlights that not all suicide-related media reporting is associated with subsequent increases in suicides. A need for a better under- standing of processes that may be involved in the specific impact of suicide-related media portrayals has repeatedly been noted in the literature (see, e.g., Pirkis, 2001b; Till, Niederkrotenthaler, Herberth, Sonneck,& Vitouch, 2010).

Beside the prevention of suicide-related effects via news media and film, the mass media play another important role in suicide prevention as well. In particular, suicide awareness and education campaigns, which aim at educating the public about suicidality and mental health issues, are increasingly used and recommended in National Suicide Prevention Plans (Niederkrotenthaler, Reidenberg, Till, & Gould, 2014; World Health Organization, 2012). Evaluations of such cam- paigns are scarce and indicate mixed results. For example, a

CONTACT Florian Arendt, PhD [email protected] Department of Communication Science and Media Research, University of Munich (LMU), Oettingenstr. 67, 80538 Munich, Germany.

HEALTH COMMUNICATION 2016, VOL. 31, NO. 6, 718–726 http://dx.doi.org/10.1080/10410236.2014.993495

© 2016 Taylor & Francis

review by Dumesnil and Verger (2009) found that awareness campaigns related to depression and suicide generally showed only modest evidence for improvements in knowledge on suicide. Furthermore, most studies on awareness campaigns did not assess any impact on actual help-seeking. There are also examples of unintended effects of awareness campaigns. For example, a billboard study conducted by Klimes-Dougan et al. (2009, 2010) indicated that when exposed to the public awareness message “Prevent Suicide. Treat Depression. See Your Doctor,” adolescents and young adults most vulnerable to suicide had an increase in maladaptive coping behaviors. These findings suggest that caution is warranted when awareness campaigns are used to educate the public about suicidality, and an urgent need for more evaluation work has been noted in the literature (Dumesnil & Verger, 2009; Mann et al., 2005; National Action Alliance for Suicide Prevention, 2014; Niederkrotenthaler et al., 2014).

In the present study, we used a new approach to the study of preventive effects of printed media awareness material by testing effects on implicit suicide cognition. We tested whether awareness material featuring an individual coping with suicidal ideation can influence relevant automatic asso- ciations in memory and thus may have a suicide-preventive effect.

Implicit suicide cognition

A recent study of 76 patients who died by suicide while in hospital, or immediately after discharge, revealed that 78% of patients did not disclose any suicidal ideation at their last communication (Busch, Fawcett, & Jacobs, 2003). Even if individuals are motivated to express their thoughts, there may be a lack of introspective access to precursors of suicide (Nisbett & Wilson, 1977; Nock & Banaji, 2007b). To over- come this limitation, techniques measuring implicit suicide cognition are increasingly used to assess suicidal ideation. This allows researchers to measure the “suicidal mind” with- out relying solely on the individual’s motivation or ability to explicitly disclose suicidal thoughts.

Implicit cognition is conceptualized as the strength of the automatic association between suicide-related concepts such as death and self in memory (Nock et al., 2010). The strength of the automatic association can be understood as the poten- tial for one concept to activate another. For example, cues that activate me in memory are likely to also activate death in individuals with a strong automatic association between both (see Greenwald et al., 2002).

Previous research suggests that implicit cognition can predict suicide-related behavior. Nock and Banaji (2007a) found that the strength of association between self and self- injury significantly improved the statistical prediction of non- suicidal self-injury after accounting for demographic and psy- chiatric variables, including clinical assessment of suicide risk. Nock and Banaji (2007b) further showed that implicit cogni- tion can differentiate groups of nonsuicidal adolescents, sui- cide ideators, and suicide attempters, and implicit cognition predicted subsequent suicidal ideation at 6-month follow-up. A recently published study showed that the automatic asso- ciation between self and death predicted posthospital episodes

of suicidal self-harm (Randall, Rowe, Dong, Nock, & Colman, 2013).

Within patients presenting to a psychiatric emergency department, implicit cognition even predicted the future suicide attempts when controlling for depressive disorders, prior suicide attempts, self-reported suicide ideation, clinical assessment, and patient prediction (Nock et al., 2010). The authors argued that the strength of the automatic association between self and death may guide which behavior a person chooses to cope in extreme situations. It may “represent one of the final steps in the pathway to suicide that is activated when a person is deciding how to respond to extreme distress” (Nock et al., 2010, p. 515).

Media effects research has revealed that media exposure can influence the strength of automatic associations in mem- ory (Arendt, 2013). Studies have so far investigated effects on associations between me and smoking after watching an action film with a smoking hero (Dal Cin, Gibson, Zanna, Shumate, & Fong, 2007), between me and aggressive (Uhlmann & Swanson, 2004) after playing a violent video game, and between me and attractive after exposure to unrealistic standards of female beauty (Gurari, Hetts, & Strube, 2006).

In the current study, we hypothesized that media aware- ness material about an individual coping with a suicidal crisis increases the automatic association between “life” and self, suggesting a protective effect on suicide-related cognition, in accordance with the findings on a potential suicide-protective “Papageno effect” of these types of media portrayals (Niederkrotenthaler et al., 2010).

H1: Exposure to press material on coping with a severe suicidal crisis can increase the strength of the automatic association between self and life (implicit cognition).

The role of identification in media effects

Previous research showed that identification with a portrayed hero moderated the media effect on smoking–self associations (Dal Cin et al., 2007). Greater identification with the hero predicted stronger automatic associations between self and smoking, but only when the protagonist smoked. Identification has also been hypothesized repeatedly as an influential factor in media effects on suicidal behavior. Stack (2005) noted that “the impact of a suicide story on copycat suicide is conditioned by the degree of identification between the model and the observer” (p. 123). In accordance with the concept of “vertical identification,” which suggests that indi- viduals tend to identify themselves with someone who is perceived as socially superior, such as, for example, a celeb- rity, media coverage of celebrity suicides has been frequently reported to be associated with postreport increases in suicide (Fu & Yip, 2009; Niederkrotenthaler et al., 2009, 2012; Stack, 1987, 1990). Social similarity between model and observer seems to increase the potential of identification with the portrayed model as well. This is often called “horizontal identification,” and there is some evidence that increases of suicides following media reporting on suicide are indeed more

HEALTH COMMUNICATION 719

likely when the model and the observer are similar on impor- tant characteristics such as age and sex (Fu & Yip, 2009; Niederkrotenthaler et al., 2009, 2012).

In spite of the widespread acceptance of the role of identification in suicidal processes related to media stories, studies in this area have so far exclusively focused on the impact of identification with media protagonists for negative outcomes such as completed suicide, or homicide (Till, Niederkrotenthaler, Herberth, Vitouch, & Sonneck, 2010). Most of these studies used aggregate ecologic study designs that did not allow for a measurement of individual identifica- tion with the protagonist. There is currently no study available that tested the opposite direction, that is, whether identification with someone coping with suicidal behavior or thoughts may also decrease the risk of suicidal behavior sub- sequently. In the present study, we therefore tested the impact of identification with awareness material featuring an individual who was in adverse psychosocial circumstances, experienced suicidal ideation, but managed to actively cope with his crisis.

RQ1: Does identification moderate the media effect on implicit cognition specified in H1?

Method

Participants

In total, 112 students enrolled in an introductory course on communication science at the University of Vienna, Austria, participated in the study. Of these, n = 95 (84.8%) were female, and n = 17 (15.2%) were male. The age of participants ranged from 18 to 46 years (M = 20.49, SD = 3.43). Most participants indicated their nationality as Austrian (80.5%), followed by German (11.5%) and other nationalities (8%). Among the participants, n = 41 (35.5%) indicated that they knew someone who died of suicide, and n = 18 (16.4%) said that they sometimes had suicide ideas. This percentage is consistent with other studies indicating a relatively high prevalence of suicidal thoughts among students (see, e.g., Garlow et al., 2008). Data collection was on December 3 and 5, 2013.

Conceptualization of identification

In previous literature, identification has been conceptualized as a content-related factor (e.g., Stack, 2005) and recipient- related factor (e.g., Dal Cin et al., 2007). While social simi- larity of the suicidal protagonist with the audience in terms of variables such as age group or profession is a content- related factor, it is important to note that a content-related factor has the potential to elicit recipient’s identification with the protagonist but only if the recipient accepts this “offer.” As a recipient-related factor, identification has been conceptualized as recipient’s desire to be like or act like the protagonist (Hoffner & Buchanan, 2005). Because content- and recipient-related factors may both play a role in media effects, we included both constructs in analysis. In the

following section of the article, we use the term social similarity for content-related identification potential as described by Stack (2005) and identification for recipient- related identification as described by Dal Cin and colleagues (2007).

Experimental manipulation

An experiment with three groups was conducted. Participants allocated to the control group read an article completely unrelated to suicide. Both other intervention groups read printed awareness material featuring a protago- nist who copes with his suicidal crisis by contacting the crisis intervention center. The material presented to these groups differed solely in terms of social similarity and was comple- tely identical in all other aspects (see the appendix). The original material adapted for this study had been published by an Austrian daily newspaper in 2005. The material was titled “Avoidance of suicide by a mere hair’s breath. Martin is one of the anonymous callers who found help at the crisis intervention phone service.” The story was about the prota- gonist named Martin who experienced a severe suicidal crisis when he received the overwhelming news that his best friend’s wife was pregnant by him. At the peak of his crises, Martin prepared everything for his suicide by carbon monoxide poisoning. In the last second before his potential suicide, Martin reconsiders his plan and calls the crisis help line. The subsequent portion of the article is dedicated to how he received help from the help line, and the procedures involved. A picture illustrates a counselor at work. The article finishes by citing Martin who is deeply grateful for the help received at that time, highlighting that he is glad about his decision to get help and not to die by suicide. The contact information to the crisis hotline is placed at the end of the text.

We manipulated this story, resulting in two versions that were different on social similarity. In one version, Martin was described as a 54-year-old unskilled worker, indicating low social similarity with the student participants (word count: n = 578). In a second version, Martin was described as a 24-year-old student of communication science in order to simulate high social similarity in terms of age and occupation status with the study participants (word count: n = 576).

For the control group, we used a text published by the same daily newspaper, written by the same journalist with a similar journalistic style and a matched text length (word count: n = 531). Suicide-related concepts were not mentioned in this text. In this text, the protagonist Martin wants to rebuild a railway track on the shoreline of a lake. The control text also uses the same pictures but it is framed in a comple- tely different way.

To ensure that participants read the stories and thought about the content, we asked them to write down a short summary after the reading (for this procedure, see Wittenbrink, Judd, & Park, 2001). All participants provided a short and accurate summary.

720 F. ARENDT ET AL.

Measures

Implicit cognition The strength of the automatic association between self and life was measured with the Implicit Association Test (IAT), originally developed by Greenwald, McGhee, and Schwartz (1998) and modified by Nock and colleagues (2010). This is a reaction-time-based task, and the tool measures response times in milliseconds needed to sort stimulus words asso- ciated with life or death into categories. Individuals were asked to classify stimulus words that appeared in the mid- dle of the screen into four categories (me, not me, life, death). Two of those categories were suicide-related con- cepts (death, life) and two categories were self-related con- cepts (me, not me). The participants were required to sort 20 stimulus words (e.g., die, alive, I, they) into one of the four categories. We used the same stimulus words as Nock and colleagues (2010). A team of suicide experts (BT, FA, TN) translated them into German. In the IAT, the classifi- cation of words is assumed to be faster when the pairing of two concepts reflects a stronger association in memory. Therefore, a person who has a strong automatic association between life and self would more quickly categorize stimu- lus words when “life” and “me” are paired together as categories as opposed to when “death” and “me” are paired. A validated scoring algorithm was used (Greenwald, Nosek, & Banaji, 2003).

Positive IAT scores indicate a stronger automatic asso- ciation between life and self, whereas negative values signal a stronger automatic association between death and self. Absolute values of around 0.2 are generally interpreted as small, around 0.5 as medium, and scores from 0.8 to infinite as large (Greenwald et al., 2003). As expected, our sample showed a stronger association between self and life compared to self and death (M = 0.49, SD = 0.36). Figure 1 visualizes the frequency distribution. We used the two IAT

scores of the first and the second combined block to construct the latent variable implicit suicide cognition (for a detailed description of all IAT blocks see Greenwald et al., 1998).

Identification Recipient-related identification was measured using the wish- ful identification scale (Hoffner & Buchanan, 2005). The original scale has five items and measures the desire to be like or act like the protagonist. A factor analysis indicated that one item did not load high on the underlying factor. Without this item, one factor explained 76% of the variance. The items used included “Martin is the sort of person I want to be like myself,” “Sometimes I wish I could be more like him,” “I’d like to do the kinds of things he does in the story,” and ”He is someone I would like to emulate.” We used the factor scores of this factor for statistical analysis.

Procedure

Participants in groups of up to eight persons were invited to the University of Vienna and were placed in individual research cubicles. All participants were informed about the option to quit the study at any time during the experiment without any negative consequences. Participants with increased suicidality scores, that is, those with a score of <46 on the Brief Reasons for Living Inventory (Ivanoff, Jang, Smyth, & Linehan, 1994), were identified immediately after the experiment and all of them were informed about their scoring and offered counseling in a separate room by a trained psychologist (BT) at that point in time. Counseling was also offered to all other participants. All participants were further provided with a contact to the local crisis intervention cases in order to help them cope with any distress they might experi- ence at a later point in time.

Statistical analysis

We used structural equation modeling, which allows the use of latent variables to analyze the data. This approach was taken because implicit cognition cannot be measured directly but only via observed (reaction-time based) vari- ables with measurement error. Structural equation model- ing explicitly considers measurement error and may therefore provide a more exact estimate of media’s impact on implicit cognition (for an example of structural equation modeling in implicit cognition research, see Cunningham, Nezlek, & Banaji, 2004).

We dummy-coded the experimental condition (reference group = control group), which resulted in two dummy vari- ables. Dummy variable 1 (D1) was indicative for the effect of intervention group 1 (awareness material with low social similarity vs. control), and dummy variable 2 (D2) repre- sented the effect of intervention group 2 (awareness material with high social similarity vs. control). We calculated the interaction terms by multiplying each dummy with the mod- erator (i.e., identification). Then we specified a structural

Figure 1. Frequency distribution of implicit cognition. Values above zero indi- cate a stronger automatic association between self and life than between self and death. Most participants had a moderate (value ≈ 0.5) or strong (value ≈ 0.8) automatic association between self and life.

HEALTH COMMUNICATION 721

equation model to predict implicit cognition as a latent vari- able using both dummies, identification, and both interaction terms in one model (see Hayes, 2013, pp. 399–401).1 In order to probe the interaction effect, we tested the effect of the awareness material separately for those scoring low (n = 58) or high (n = 52) on recipient-related identification based on a median split.

Ethics statement

Ethical approval for this study was obtained from the Ethics Committee of the Medical University of Vienna (EKR- Number 1446/2013, dated June 18, 2013). All participants provided written consent to participate in the study.

Results

We hypothesized that reading a suicide-awareness text featur- ing an individual who managed to cope with suicidal ideation increased implicit cognition toward life and self (H1). However, this effect may be moderated by identifica- tion (RQ1).

The model showed a good fit to the data (N = 110; χ2 = 2.51; df = 4; p = .64; RMSEA < .001; CFI = 1.00). The interaction term of D1*identification (with D1) was signifi- cant, Coeff = −0.183, SE = 0.093, p = .049. This result indi- cated that the size of the effect of reading the text with low social similarity was significantly dependent on identification. The second interaction term of D2*identification (with D2) was not significant, Coeff = −0.181, SE = 0.115, p = .117.

We calculated a further structural equation model using multigroup analysis to probe the interaction effect. We pre- dicted implicit cognition by both dummies and estimated the effects separately for those scoring low or high on recipient- related identification (N = 110; χ2 = 0.86; df = 2; p = .65; RMSEA < .001; CFI = 1.00). This analysis revealed the direc- tion of the interaction effect. As visualized in Figure 2, expo- sure to awareness material increased implicit cognition only in those scoring low on identification: Intervention group 1, Coeff = 0.642, SE = 0.186, p < .001, as well as intervention group 2, Coeff = 0.511, SE = 0.185, p = .006, showed signifi- cant effects on implicit cognition. In contrast, in individuals scoring high on identification, neither intervention group 1, Coeff = −0.212, SE = 0.156, p = .174, nor intervention group 2, Coeff = −0.080, SE = 0.120, p = .509, showed an effect on implicit cognition.2

Discussion

This study indicates that suicide awareness material featuring an individual coping with suicidal ideation material may have a protective effect on suicidal cognition in those exposed to the material. This effect seemed to be restricted to participants who did not wishfully identify with the protagonist.

Considering the results of previous research showing that implicit cognition predicts actual suicide behavior (Nock et al., 2010), we cautiously argue that awareness material featuring an individual mastering his crisis may have (indir- ect) suicide-protective effects by its influence on implicit cognition.

The present study is probably the first that tested the role of identification with a suicidal media protagonist with regard to a potentially suicide-protective outcome, and indicated that, opposite to the literature on harmful outcomes, only individuals who did not wishfully identify with the protago- nist showed an increase in life–self associations. A possible explanation for this divergence may derive from social com- parison theory (Corcoran, Crusius, & Mussweiler, 2011; Festinger, 1954). According to this theory, individuals tend to evaluate their thoughts and actions by comparing them- selves with other people. The exposure to the severe crisis situation of the protagonist, who managed to deal with his adverse circumstances but was in serious trouble over a relatively long time, may have produced some type of “con- trast effect” (Till, Niederkrotenthaler, Herberth, Sonneck, &Vitouch, 2010) that resulted in a positive self-evaluation and a stronger beneficial effect in the group that did not wishfully identify with the protagonist.

Different from sensationalist media stories on completed suicide or violence, which are deemed to have the highest potential to result in further suicide (World Health Organization, 2008) or violence, the story of the protagonist in the awareness material used was neither portrayed in a

Figure 2. Identification as an effect moderator. Influence of exposure to printed suicide awareness material featuring an individual coping with a suicidal crisis on the strength of the automatic association between self and life (implicit cognition). Effects are expressed as unstandardized coefficients with standard errors in parentheses. Bold arrows indicate significant effects. The upper part of the figure shows the model for participants scoring low on identification, and the lower part shows the effects for those scoring high on identification.

1Two participants had missing values on the wishful identification scale and were excluded from the analysis, resulting in a sample size of N = 110 for structural equation modeling.

2Of note, individuals who scored low versus high on identification did not differ with regard to age, t(108) = 1.16, p = .25, gender, χ2 = 0.72, p = .40, or prevalence of suicidal ideas, χ2 = 0.06, p = .80.

722 F. ARENDT ET AL.

sensationalist nor a heroic way. The crisis or the burdensome act of working through the crisis was portrayed as a difficult but ultimately successful process that took more than 3 years. This conceptualization of the story may have influenced the quality of social comparison of participants with the protago- nist. Specifically, participants who scored low on wishful identification may have perceived themselves as stronger and less vulnerable as compared to the featured protagonist, thereby strengthening their self–life association.

It must be noted that readers may interpret the content of suicide stories in different ways. For example, it might be important whether readers identify more with the protago- nist’s personality and suicidality or rather with the process of coping and his success in doing so. When readers identify with the process of coping, the object of identification is clearly more positively connotated. The quality of identifica- tion may moderate media effects on implicit suicide cognition. Further studies are necessary to scrutinize aspects of identification that may play a role in the identified associations.

In terms of current suicide awareness campaigns (see Andreasen, 1994; Hassan, Walsh, Shiu, Hastings, & Harris, 2007), which often target broad population samples, the pre- sent finding has potentially high relevance. In these cam- paigns, a frequent approach is to feature portrayals that are likely to increase identification with the portrayal in order to achieve the desired effect. However, it has been noted repeat- edly that broad suicide awareness campaigns failed to elicit intended effects in their target populations (Dumesnil & Verger, 2009; Niederkrotenthaler et al., 2014). For example, one study showed that a suicide awareness campaign promot- ing a crisis intervention center resulted in an increased utili- zation of these services, but there was a decrease in the proportion of suicidal individuals among the clients (Sonneck, Kapusta, Tomandl, & Voracek, 2012). Similarly, a suicide awareness campaign that featured family issues as a source of interpersonal conflicts associated with suicidal idea- tion resulted in an overall increase of individuals utilizing the service, but there was also a decrease in individuals presenting with family problems (Till, Sonneck, Baldauf, Steiner, & Niederkrotenthaler, 2013). Based on the current study, a con- tributing factor to these unintended and unexplained findings may be that the awareness material used in these campaigns resulted in changes in help-seeking behavior primarily in individuals who did not identify themselves with the por- trayed situation. Further studies are necessary in both indivi- duals who identify and individuals who do not identify with the story contents to investigate this hypothesis further. These studies should analyze effects both in specific general popula- tion samples and in individuals with increased vulnerability to suicidal behavior, and should carefully investigate the object of identification.

Study limitations

This study has limitations. First, the sample used in the study was small. Due to the resulting lack of statistical power, we cannot rule out the presence of undetected associations. Furthermore, we did not assess long-term effects. It has

been shown that effects of media materials can stretch over a long time period, and may even influence suicidal behavior months after the actual media exposure (Fu & Yip, 2007). Further, the featured protagonist in our awareness material was male, although the majority of study participants were women. Male and female participants did not differ in terms of identification in our study, but future studies should use gender as a further variable to enhance or reduce social similarity. Finally, we tested only the effect on implicit cogni- tion. Other target outcomes of awareness campaigns such as help-seeking behavior were not considered.

Conclusion

This study suggests that awareness material featuring the positive coping with a suicidal crisis may strengthen indivi- duals in their association with life. However, this effect seemed to be restricted to individuals who did not wishfully identify with the featured protagonist. The explanation for this finding may be a stronger defense against the tragic circumstances based on social comparison, which may help individuals who do not perceive themselves as similar to the protagonist to strengthen their own self–life association. Further studies are necessary to investigate this hypothesising in order to prevent unintended backlash effects of awareness campaigns and to build effective messaging strategies that can ultimately save lives.

References

Andreasen, A. (1994). Social marketing: Its definition and domain. Journal of Public Policy and Marketing, 13, 108–114.

Arendt, F. (2013). Dose-dependent media priming effects of stereotypic newspaper articles on implicit and explicit stereotypes. Journal of Communication, 63, 830–851.

Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.

Busch, K., Fawcett, J., & Jacobs, D. (2003). Clinical correlates of inpatient suicide. The Journal of Clinical Psychiatry, 64, 14–19. doi:10.4088/JCP. v64n0105

Corcoran, K., Crusius, J., & Mussweiler, T. (2011). Social comparison: Motives, standards, and mechanisms. In D. Chadee (Ed.), Theories in social psychology (pp. 119–139). Oxford, UK: Wiley-Blackwell.

Cunningham, W., Nezlek, J., & Banaji, M. (2004). Implicit and explicit ethnocentrism: Revisiting the ideologies of prejudice. Personality and Social Psychology Bulletin, 30, 1332–1346. doi:10.1177/0146167204264654

Dal Cin, S., Gibson, B., Zanna, M., Shumate, R., & Fong, G. (2007). Smoking in movies, implicit associations of smoking with the self, and intentions to smoke. Psychological Science, 18, 559–563. doi:10.1111/ j.1467-9280.2007.01939.x

Dumesnil, H., & Verger, P. (2009). Public awareness campaigns about depression and suicide: A review. Psychiatric Services, 60, 1203–1213. doi:10.1176/ps.2009.60.9.1203

Etzersdorfer, E., & Sonneck, G. (1998). Preventing suicide by influencing mass-media reporting. The Viennese experience 1980–1996. Archives of Suicide Research, 4, 67–74. doi:10.1080/13811119808258290

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140. doi:10.1177/001872675400700202

Fu, K. W., & Chan, C. H. (2013). A study of the impact of thirteen celebrity suicides on subsequent suicide rates in South Korea from 2005 to 2009. PLoS ONE, 8(e53870). doi:10.1371/journal. pone.0053870

HEALTH COMMUNICATION 723

Fu, K. W., & Yip, P. S. (2007). Long-term impact of celebrity suicide on suicidal ideation: Results from a population-based study. Journal of Epidemiology and Community Health, 61, 540–546. doi:10.1136/ jech.2005.045005

Fu, K. W., & Yip, P. S. (2009). Estimating the risk for suicide following the suicide deaths of 3 Asian entertainment celebrities: A meta-analy- tic approach. Journal of Clinical Psychiatry, 70, 869–878. doi:10.4088/ JCP.08m04240

Garlow, S., Rosenberg, J., Moore, D., Haas, A., Koestner, B., Hendin, H., & Nemeroff, C. (2008). Depression, desperation, and suicidal ideation in college students: Results from the American Foundation for Suicide Prevention college screening project at Emory University. Depression and Anxiety, 25, 482–488. doi:10.1002/(ISSN)1520-6394

Greenwald, A. G., Banaji, M., Rudman, L., Farnham, S., Nosek, B., & Mellot, D. (2002). A unified theory of implicit attitudes, stereotypes, self-esteem, and self-concept. Psychological Review, 109, 3–25. doi:10.1037/0033-295X.109.1.3

Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74, 1464–1480. doi:10.1037/0022-3514.74.6.1464

Greenwald, A. G., Nosek, B., & Banaji, M. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197–216. doi:10.1037/ 0022-3514.85.2.197

Gurari, I., Hetts, J., & Strube, M. (2006). Beauty in the “I” of the beholder: Effects of idealized media portrayals on implicit self- image. Basic and Applied Social Psychology, 28, 273–282. doi:10.1207/s15324834basp2803_6

Hassan, L. M., Walsh, G., Shiu, E., Hastings, G., & Harris, F. (2007). Modeling persuasion in social advertising: A study of responsible thinking in antismoking promotion in eight eastern EU (European Union) member states. Journal of Advertising, 36, 15–31. doi:10.2753/ JOA0091-3367360201

Hayes, A. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press.

Hoffner, C., & Buchanan, M. (2005). Young adults’ wishful identification with television characters: The role of perceived similarity and char- acter attributes. Media Psychology, 7, 325–351. doi:10.1207/ S1532785XMEP0704_2

Ivanoff, A., Jang, S. J., Smyth, N. J., & Linehan, M. M. (1994). Fewer reasons for staying alive when you are thinking of killing yourself: The brief reasons for living inventory. Journal of Psychopathology and Behavioral Assessment, 16, 1–13. doi:10.1007/BF02229062

Klimes-Dougan, B., & Lee, C. Y. S., in collaboration with Suicide Awareness Voices of Education. (2010). Suicide prevention public service announcements. Perceptions of young adults. Crisis, 31, 247– 254. doi:10.1027/0227-5910/a000032

Klimes-Dougan, B., Yuan, C., Lee, S., & Houri, A. K. (2009). Suicide prevention with adolescents. Considering potential benefits and unto- ward effects of public service announcements. Crisis, 30, 128–135. doi:10.1027/0227-5910.30.3.128

Mann, J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., & Hendin, H. (2005). Suicide prevention strategies. A systematic review. Journal of the American Medical Association, 224, 2064–2074. doi:10.1001/jama.294.16.2064

Michel, K., Frey, C., Wyss, K., & Valach, L. (2000). An exercise in improving suicide reporting in print media. Crisis, 21, 71–79. doi:10.1027//0227-5910.21.2.71

National Action Alliance for Suicide Prevention. (2014). A prioritized research agenda for suicide prevention: An action plan to save lives. Retrieved from http://actionallianceforsuicideprevention.org/sites/ actionallianceforsuicideprevention.org/files/Agenda.pdf

Niederkrotenthaler, T., Fu, K. W., Yip, P., Fong, D. Y. T., Stack, S., Cheng, Q., & Pirkis, J. (2012). Changes in suicide rates following media reports on celebrity suicides: A meta-analysis. Journal of Epidemiology and Community Health, 66, 1037–1042. doi:10.1136/jech-2011-200707

Niederkrotenthaler, T., Reidenberg, D. J., Till, B., & Gould, M. (2014). Increasing help-seeking and referrals for individuals at risk for

suicide by decreasing stigma—The role of mass media. American Journal of Preventive Medicine, 47, SS235–43. doi:10.1016/j. amepre.2014.06.010

Niederkrotenthaler, T., & Sonneck, G. (2007). Assessing the impact of media guidelines for reporting on suicides in Austria: Interrupted time-series analysis. Australian and New Zealand Journal of Psychiatry, 41, 419–428. doi:10.1080/00048670701266680

Niederkrotenthaler, T., Till, B., Kapusta, N., Voracek, M., Dervic, K., & Sonneck, G. (2009). Copycat effects after media reports on suicide: A population-based ecologic study. Social Science & Medicine, 69, 1085– 1090. doi:10.1016/j.socscimed.2009.07.041

Niederkrotenthaler, T., Voracek, M., Herberth, A., Till, B., Strauss, M., Etzersdorfer, E., . . . Sonneck, G. (2010). Role of media reports in com- pleted and prevented suicide: Werther v. Papageno effects. The British Journal of Psychiatry, 197, 234–243. doi:10.1192/bjp.bp.109.074633

Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84, 231–259. doi:10.1037/0033-295X.84.3.231

Nock, M., & Banaji, M. (2007a). Assessment of self-injurious thoughts using a behavioral test. The American Journal of Psychiatry, 164, 820– 823. doi:10.1176/appi.ajp.164.5.820

Nock, M., & Banaji, M. (2007b). Prediction of suicide ideation and attempts among adolescents using a brief performance-based test. Journal of Consulting and Clinical Psychology, 75, 707–715. doi:10.1037/0022-006X.75.5.707

Nock, M., Park, J., Finn, C., Deliberto, T., Dour, H., & Banaji, M. (2010). Measuring the suicidal mind: Implicit cognition predicts suicidal beha- vior. Psychological Science, 21, 511–517. doi:10.1177/0956797610364762

Phillips, D. (1974). The influence of suggestion on suicide: Substantive and theoretical implications of the Werther effect. American Sociological Review, 39, 340–354. doi:10.2307/2094294

Pirkis, J., & Blood, R. W. (2001a). Suicide and the media. Part I: Reportage in nonfictional media. Crisis, 22, 155–162. doi:10.1027// 0227-5910.22.4.155

Pirkis, J., & Blood, W. (2001b). Suicide and the media. Part III: Theoretical issues. Crisis, 22, 163–169. doi:10.1027//0227-5910.22.4.163

Pirkis, J., Blood, R. W., Beautrais, A., Burgess, P., & Skehans, J. (2006). Media guidelines on the reporting of suicide. Crisis, 27, 82–87. doi:10.1027/0227-5910.27.2.82

Pirkis, J., Dare, A. R., Blood, W., Rankin, B., Williamson, M., Burgess, B., & Jolley, D. (2009). Changes in media reporting of suicide in Australia between 2000/01 and 2006/07. Crisis, 30, 25–33. doi:10.1027/0227- 5910.30.1.25

Randall, J., Rowe, B., Dong, K., Nock, M., & Colman, I. (2013). Assessment of self-harm risk using implicit thoughts. Psychological Assessment, 25, 714–721. doi:10.1037/a0032391

Sonneck, G., Kapusta, N., Tomandl, G., & Voracek, M. (2012). Krisenintervention und Suizidverhütung [Crisis intervention and sui- cide prevention]. Vienna, Austria: UTB facultas.wuv.

Stack, S. (1987). Celebrities and suicide: A taxonomy and analysis, 1948–1983. American Sociological Review, 52, 401–412. doi:10.2307/ 2095359

Stack, S. (1990). A reanalysis of the impact of non-celebrity suicides: A research note. Social Psychiatry and Psychiatric Epidemiology, 25, 269– 273. doi:10.1007/BF00788648

Stack, S. (2003). Media coverage as a risk factor in suicide. Journal of Epidemiology & Community Health, 57, 238–240. doi:10.1136/ jech.57.4.238

Stack, S. (2005). Suicide in the media: A quantitative review of studies based on nonfictional stories. Suicide and Life-Threatening Behavior, 35, 121–133. doi:10.1521/suli.2005.35.issue-2

Till, B., Niederkrotenthaler, T., Herberth, A., Sonneck, G., & Vitouch, P. (2010). Coping and film reception: The influence of coping strategies on emotional modes of film reception and the resulting impact on viewers. Suicide and Life-Threatening Behavior, 40, 319–327. doi:10.1521/suli.2010.40.4.319

Till, B., Niederkrotenthaler, T., Herberth, A., Vitouch, P., & Sonneck, G. (2010). Suicide in films: The impact of suicide portrayals on non-suicidal viewers’ well-being and the effectiveness of censorship. Suicide & Life- Threatening Behavior, 40, 319–327. doi:10.1521/suli.2010.40.4.319

724 F. ARENDT ET AL.

Till, B., Sonneck, G., Baldauf, G., Steiner, E., & Niederkrotenthaler, T. (2013). Reasons to love life. Effects of a suicide-awareness campaign on the utilization of a telephone emergency line in Austria. Crisis, 34, 382–389. doi:10.1027/0227-5910/a000212

Uhlmann, E., & Swanson, J. (2004). Exposure to violent video games increases automatic aggressiveness. Journal of Adolescence, 27, 41–52. doi:10.1016/j.adolescence.2003.10.004

Wittenbrink, B., Judd, C., & Park, B. (2001). Spontaneous prejudice in context: Variability in automatically activated attitudes. Journal of

Personality and Social Psychology, 81, 815–827. doi:10.1037/0022- 3514.81.5.815

World Health Organization. (2008). Preventing suicide. A resource for media professionals. Retrieved from http://www.who.int/mental_ health/prevention/suicide/resource_media.pdf

World Health Organization. (2012). Public health action for the prevention of suicide: A framework. Retrieved from http://apps.who.int/iris/bitstream/10665/75166/1/9789241503570_ eng.pdf

Appendix

Suicide Awareness Material Used in the Present Study

HEALTH COMMUNICATION 725

English translation of the title page: High social similarity Series: This moves the Austrian public: How Martin managed to

cope with his suicide plans and now lives a fulfilled life. A student (24) tells his moving story. “Life is beautiful.”

Low Social Similarity Series: This moves the Austrian public: How Martin managed to

cope with his suicide plans and now lives a fulfilled life. A construction worker (54) tells his moving story. “Life is beautiful.”

Control text:

726 F. ARENDT ET AL.

Copyright of Health Communication is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.