High Quality
Open the downloaded document, and select print from the file menu (PDF reader required).
................................................................................................
The gap between suicide characteristics in theprint media and in the population
Thomas Niederkrotenthaler
1
, Benedikt Till
2
, Arno Herberth
3
, Martin Voracek
4
,Nestor D. Kapusta
5
, Elmar Etzersdorfer
6
, Markus Strauss
1
, Gernot Sonneck
1,7
Background:
Programmes to educate media professionals about suicide are increasingly established,but information about which suicide cases are most likely to be reported in the mass media is sparse.
Methods:
We applied binomial tests to compare frequencies of social characteristics of all domesticsuicides in the 13 largest Austrian print media in 2005 with frequencies of suicide characteristics in thepopulation. Additionally, each reported suicide case was linked to its respective entry in the suicidedatabase. We performed a logistic regression analysis, with presence of an article as outcome, and sexof the suicide case, age, religious affiliation, family status, conduction of an autopsy and location ofthe suicide as explaining variables. Time of the year and federal state where the suicide happenedwas controlled for.
Results:
Binomial tests showed that suicides involving murder or murder attemptwere over-represented in the media. Reporting on mental disorders was under-represented. In theregression analysis, the likelihood of a report was negatively associated with the age of suicide cases.Foreign citizenship was a further predictor of a suicide report. The methods of drowning, jumping,shooting and rare methods were more likely to be reported than hanging, which is the most frequentsuicide method in Austria.
Conclusions:
Suicide characteristics in the media are not representative ofthe population. The identified discrepancies provide a basis for tailor-made education of massmedia professionals.
Keywords:
mass media, suicide, education, health promotion, Austria.
................................................................................................
Introduction
T
he World Health Organization and national agenciesrecommend collaborations of public health specialistswith the media to educate the public about suicidality
1,2
andto prevent copycat effects after media reports.
3–5
In orderto offer tailor-made educative programmes it is important toknow which suicide characteristics are most likely to bereported by the mass media.
Methods
We obtained all print media reports from the 13 largestnationwide Austrian newspapers that included the word‘suicide’ or ‘self-murder’ in 2005. The search yielded 2016reports. These reports were scanned for accounts of individualdomestic suicides. Reports on suicides in foreign countries,fictional suicides, suicides in another time period and reportsthat did not contain a completed suicide were excluded fromthe analysis. Reports on the same suicide were clusteredtogether to avoid redundant counting of the same case.Reports on 61 different suicides were identified. From thesereports, we extracted the sex of the suicide case, age group(adolescents up to 19, 20–64 or elderly from 65 years),nationality (Austrian or foreign citizenship), severe criminality [involvement in murder (attempt) or not], urbanity (suicidehappened in the capital of a state or in more rural regions),suicide method (hanging, jumping, shooting, drowning,poisoning or other) and reported illness (mental disorderreported or not). The extraction of data was done indepen-dently by two trained coders, and inter-coder-reliability wasmeasured for qualitative variables. Krippendorff’s agreementvalue
a
reached values of 0.89–1, corresponding toconcordance rates above 0.95.
6
Data on suicide characteristics in the population includingdate of death, sex, age, nationality, county, conduction of anautopsy, religious affiliation, family status, location of deathand suicide method according to ICD-10 classification wereobtained from Statistics Austria. Prevalence of mentaldisorders and involvement in murder and murder attempt insuicide cases was obtained from police files in the largestfederal state, Styria.
7
We applied binomial tests to calculate if frequencies of suicide characteristics in the media were equal to theirfrequencies in the population. Subsequently, we assignedeach suicide report to its respective entry in the suicidedatabase of 2005 (
n
=1392). For 54 (88.5%) of the reportedsuicides, we identified exactly one compatible entry. For eachof the remaining seven suicide cases (11.5%), at least one of thereported suicide characteristics was contradictory to itspossible entry in the suicide database. These suicide reportswere excluded from further analysis. We conducted a multiplelogistic regression analysis, with presence of a report asoutcome, and sex of suicide cases, age group, religious affili-ation, family status, conduction of an autopsy and location of the suicide as independents. Because the choice to report asuicide may vary with the season of the year and with localeditorial staffs of the Austrian Press Agency who provide the
Correspondence:
Thomas Niederkrotenthaler, Medical University of Vienna, Center for Public Health, Institute for Medical Psychology,Severingasse 9, 1090 Vienna, Austria, tel: +43 1 4277 65626, fax: +43 14277 9656, e-mail: thomas.niederkrotenthaler@meduniwien.ac.at1 Institute for Medical Psychology, Center for Public Health, MedicalUniversity of Vienna, Austria2 Department of Communication, University of Vienna, Austria3 Department of German Language and Literature, University of Vienna, Austria4 Department of Basic Psychological Research, School of Psychology,University of Vienna, Austria5 Department of Psychoanalysis and Psychotherapy, MedicalUniversity of Vienna, Spitalgasse 23, A-1090 Vienna, Austria6 Furtbach Hospital, Furtbachstraße 6, 70178 Stuttgart, Germany 7 Ludwig Boltzmann Institute for Social Psychiatry, Vienna, Austria
European Journal of Public Health
, 1–4
ß
The Author 2009. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.doi:10.1093/eurpub/ckp034
The European Journal of Public Health Advance Access published March 19, 2009
news for the different media, we adjusted the model forseasons and federal states. We performed data analysis withSAS (SAS, 9.1).
Results
Binomial tests showed that suicides of children and adoles-cents were more frequent in the media than in the population(table 1). Suicides of people over 64 years were under-represented. There were considerably more reports aboutsuicide cases with foreign citizenship than expected. Suicidesby hanging were under-represented and there were morereports about suicides by jumping and drowning thanexpected. In articles that discussed the causality of suicide,reporting of mental disorders was under-represented. Suicidesinvolving murder or murder attempt were over-represented.In the regression analysis, the likelihood of a reportdecreased with increasing age of suicide cases (table 2).Suicides by foreign citizens were more likely to be reportedthan suicides by Austrian citizens. Suicides by drowning, jumping, shooting and rare methods were associated witha higher likelihood of reporting than suicides by the mostfrequent suicide method, hanging. The likelihood-ratio teststatistic
8
indicated that the total model was significant(
2
=85.17, df=26,
P
<0.001). The Hosmer–Lemeshow good-ness-of-fit test
8
was not significant (
2
=9.38,
P
=0.31),implying an acceptable fit of the model. The Nagelkerke’sPseudo-
R
2
statistic
8
of the regression model was 0.21, which isconsidered satisfactory.
Discussion
The present study identified huge discrepancies betweensuicide characteristics in the population and the media.Adolescent suicides seem to be newsworthy for the media,which is in sharp contrast to suicides by elderly, who have
Table 1
Binomial tests compared frequencies of suicide characteristics in the mass media and the population
Characteristics of suicide Suicides in the Population Suicides in the Media Binomial test
n
(Proportion)
n
(Proportion) 95% CI
P
*
SexMale 1043 (0.75) 40 (0.67) 0.52–0.77 0.132Female 349 (0.25) 21 (0.33) 0.23–0.48 0.132Age group<20 48 (0.04) 9 (0.17) 0.09–0.31 <0.00120–64 897 (0.64) 37 (0.76) 0.60–0.85 0.200>64 447 (0.32) 4 (0.07) 0.02–0.19 <0.001NationalityAustrian 1343 (0.97) 43 (0.81) 0.68–0.91 <0.001Foreigner 49 (0.04) 10 (0.19) 0.09–0.32 <0.001Severe criminalityHomicide (attempt)
a
124 (0.02) 9 (0.13) 0.07–0.26 <0.001Illness
a
Mental illness 715 (0.51) 12 (0.19) 0.11–0.32 <0.001Suicide methodHanging 667 (0.44) 9 (0.15) 0.08–0.29 <0.001Jumping 79 (0.11) 18 (0.32) 0.21–0.48 <0.001Shooting 234 (0.17) 8 (0.15) 0.07–0.27 0.866Drowning 148 (0.06) 11 (0.20) 0.11–0.34 <0.001Poisoning 147 (0.11) 4 (0.07) 0.02–0.18 0.619Other method 167 (0.12) 6 (0.14) 0.04–0.84 0.999UrbanitySuicide in capital city 401 (0.29) 18 (0.37) 0.25–0.54 0.206In other area 991 (0.71) 29 (0.64) 0.46–0.76 0.206
*:
P
-values are two-tailed.a: The prevalence of murder, murder attempts (1995–2004) and mental disorders (2000–04) in suicide cases was derived frompolice files in the largest Austrian federal state, Styria
Table 2
ORs from multiple logistic regression analysis forthe presence of media reports
Variable
a
OR (95% CI) Chi-square
P
SexMale 0.52 (0.26–1.01) 3.69 0.054Female (Ref) 1.00Age
b
Age increase of 10 years 0.62 (0.50–0.77) 18.59 <0.001CitizenshipForeign citizen 5.49 (2.07–14.58) 11.67 <0.001Austrian citizen (Ref) 1.00Religious affiliationOther religion 1.09 (0.98–1.03) 0.06 0.789Christian
c
(Ref) 1.00Family statusSingle 0.70 (0.33–1.51) 0.81 0.367Divorced or widowed 0.50 (0.20–1.27) 2.12 0.146Married (Ref) 1.00Place of deathAsylum 1.23 (0.43–3.49) 0.15 0.697Other public institution 1.11 (0.13–9.74) 0.01 0.923Other place 1.01 (0.53–1.90) 0.00 0.983Home address (Ref) 1.00MethodDrowning 7.63 (2.64–22.02) 14.13 <0.001Shooting 4.81 (1.90–12.21) 10.96 <0.001Jumping 4.10 (1.52–11.03) 7.80 0.005Poisoning 2.58 (0.87–7.66) 2.91 0.088Other method 3.56 (1.41–8.98) 7.22 0.007Hanging (Ref) 1.00AutopsyAutopsy conducted 1.08 (0.50–2.32) 0.04 0.849No autopsy (Ref) 1.00
a: For facility of inspection, the variables for season (spring:March–May; summer: June–August; autumn: September–November; winter: December–February) and federal stateare not given in the tableb: The following age groups were used: <20, 20–29, 30–39,40–49, 50–59, 60–69, 70–79, 80+ yearsc: Roman Catholic, Protestant, Old Catholic
2 of 4
European Journal of Public Health
experienced high suicide rates in recent years in many European countries, including Austria.
9
Of note, a preferenceto report on child death has also been identified previously inreporting on suicide in Swiss print media.
10
Foreign citizenshipis a further characteristic of suicide cases that represents only a small minority of suicides in the population but predictedreporting. Furthermore, the suicide methods of jumping,drowning, shooting and rare methods were associated with anincreased likelihood of reporting compared with hanging,which is the most frequent suicide method in Austria and inmany other European countries.
11
Similar to previous findings on homicide reporting in theUS,
12
the present results suggest that public health with itsfocus on risk groups for suicide is at least partially at odds withthe mass media, which defines newsworthiness as being ‘out of the ordinary’.
13
The materials that are provided by the massmedia for ‘reality construction’ are selective and do not reflectobjective social reality.
14
Educational collaborations with themedia may not be able to change the basic media interest inextraordinary suicide cases. Rather, educational efforts need tobridge the gap between suicide characteristics in the massmedia and the population that result from selective mediainterest by giving a complementary voice to the social realitiesof suicide. Specifically, it is necessary to address the problem of increasing suicide rates over the life-span in Europe,
9
and theconsistency of suicide rates of first-generation immigrants withthe rates in the respective countries of origin, which has beenidentified for several European host countries.
15
The relation-ship between mental illness and suicide
16
and the fact thathomicide suicides are only a small fraction of all suicidesshould be further key messages in educative programmes.Given that copycat effects after media reporting may result inan increase of method-specific suicide rates
17
it is necessary topay special attention to media reporting about comparatively infrequent methods to prevent such effects.There are some limitations to the present study. Weexcluded international suicide reports to identify suicidecharacteristics in the population that predict suicide reporting.International reports are likely to be restricted to mostextraordinary cases like celebrity suicides. Another limitationwas that the prevalence of murder (attempts) and mentaldisorders in suicide cases were only obtainable for one federalstate.
7
Furthermore, the present analysis was restricted to the year 2005. Our extensive search strategy for articles probably resulted in the identification of most suicide reports of this year, but limited the resources to analyse a longer time period.Future studies should analyse a longer time frame to replicateour findings. Additionally, it is necessary to identify patterns of suicide reporting in other countries. Cultural patterns andattitudes may influence the choice to report a suicide case.Finally, the present study restricted its focus on print mediareporting. The high coverage rates of print media in Austriaprovided a basis for this strategy.
18
However, future studiesshould also consider TV and internet. Television reports onsuicide were found to have a larger impact on the audiencethan print media reports in an Australian study.
19
The mass media is a proximate resource of information onsuicide for the public. Deviant portrayal of suicide may resultin social and political responses to the problem that are notbased on realities in the population.
20
The present study demonstrated that only a small proportion of suicides (3.9%)is reported in the print media, and indicated which suicides aremost likely to get attention by the media. The findings make iteasier for public health professionals to address suicidecharacteristics in the population that fall short of reporting.Thereby, the study contributes to more tailor-made educativeprogrammes for media professionals.
Funding
Austrian Academy of Sciences (70036 to T.N., 70034 toB. T. and 70035 to A.H.).
Conflicts of interest
: None declared.
Key points
A huge majority of suicide cases are not reported inthe print media.
Young age of suicide victims, foreign citizenship andrare suicide methods predict media reporting onsuicide.
Mental illness is under-represented in media accountsof suicide, homicide–suicide is over-represented.
Programmes to educate media professionals aboutsuicide have to focus on the identified discrepanciesbetween suicide characteristics in the media and thepopulation.
References
1 Knox KL, Conwell Y, Caine ED. If suicide is a public health problem,what are we doing to prevent it?
Am J Public Health
2004;94:37–46.2 Hegerl U, Althaus D, Schmidtke A, Niklewski G. The alliance againstdepression: 2-year evaluation of a community based intervention to reducesuicidality.
Psychol Med
2006;36:1225–33.3 World Health Organization. Preventing suicide. A resource for mediaprofessionals. Available at:http://www.who.int/mental_health/prevention/suicide/resource_media.pdf (5 December 2008, date accessed).4 Niederkrotenthaler T, Sonneck G. Assessing the impact of mediaguidelines for reporting on suicides in Austria: interrupted time seriesanalysis.
Aust NZ J Psychiatry
2007;41:419–28.5 Grey C, Michel K, Valach L. Suicide reporting in Swiss print media:Responsible or irresponsible?
Eur J Public Health
1995;7:15–19.6 Krippendorff K.
Content analysis. An Introduction to its Methodology
.London: Sage, 2003.7 Watzka C.
Sozialstruktur und Suizid in Austria
(Social structure of suicides inAustria) (in German). Wiesbaden: Verlag fu¨r Sozialwissenschaften, 2008,299–310.8 Hazard Munro B.
Statistical Methods for Health Care Research
. Philadelphia:Lippincott, Williams & Wilkins, 2005, 301–20.9 Kapusta ND, Etzersdorfer E, Sonneck G. Trends in suicide rates of the elderly in Austria, 1970-2004: an analysis of changes in terms of age groups, suicide methods and gender.
Int J Geriatr Psychiatry
2007;22:438–44.10 Michel K, Frey C, Schlaepfer E, Valach L. Suicide reporting in the Swiss printmedia.
Eur J Public Health
1995;5:199–203.11 Gunnell D, Bennewith O, Hawton K, et al. The epidemiology andprevention of suicide by hanging: a systematic review.
Int J Epidemiol
2005;34:433–42.12 Sorenson SB, Peterson Manz JG, Berk RA. News media coverage and theepidemiology of homicide.
Am J Public Health
1998;88:1510–14.13 Browne K.
An Introduction to Sociology
. Cambridge, UK: Polity Press,2005, 173.14 McQuail.
Mass Communication Theory
. London, UK: Sage, 2008, 102.15 Voracek M, Loibl LM. Consistency of immigrant and country-of-birthsuicide rates: a meta-analysis.
Acta Psychiatr Scand
2008;118:259–71.16 Mortensen PB, Agerbo E, Erikson T, et al. Psychiatric illness and risk factorsfor suicide in Denmark.
Lancet
2000;355:9–12.17 D Gunnell. Reporting suicide. The effect of media coverage on patterns of self harm.
Br Med J
1994;308:1446–7.
Media and suicide
3 of 4
Add a Comment