Abstract
This study analyzed violent crime reports in three large circulation newspapers and tested by survey the relationship between newspaper reporting of crime and fear of crime. As was expected, there was a nexus between more sensationalist newspaper reporting and the public’s fear of crime. Unexpected findings were an inverse relationship between newspaper format and awareness of violent crime, and also between awareness of crime and fear of crime.
Crime is inherently newsworthy, thanks to its stark deviance from deeply engrained social norms. 1 Does the sensationalistic presentation of crime with oversized headlines and emotion-laden photographs and graphics by newspaper influence readers’ fear of crime in the real world? This article reports findings of a research project investigating the relationship between news media coverage of violent crime in Hong Kong and readers’ knowledge of crime and fear of personal risk from crime. 2 In this study, physical assaults on individuals, including homicides, beatings and sexual assaults, are treated as violent crimes.
Hong Kong’s crime rate has fallen steadily over the past decade 3 and it has become one of the world’s safest jurisdictions. 4 Public perceptions are very different, however. Following identification of appropriate target newspapers and an appropriate survey cohort, an empirical content analysis of crime news reports in three of the largest circulation Chinese-language newspapers in Hong Kong was performed. A group of 274 newspaper readers was surveyed by means of a questionnaire and the results were then analyzed to determine the impact of different press reports on the readers’ knowledge of and concern over crime.
Sensational Newspaper Reports and Fear of Crime
When direct experience is lacking, perception of issues is often shaped by images in the media, 5 particularly those related to crime. 6 U.S. and U.K. studies in particular link readers’ fear of crime to sensationalist style reporting. 7 Sensationalism supplies emotionally charged content, particularly violent crime reports, 8 and attracts an audience seeking emotional arousal in its media consumption. 9 Emotional arousal has been linked to enhanced memory and information retention, 10 with the result that readers are most likely to recall sensationalist reporting exemplified by large headlines, colorful photographs, “short, pithy, highly stereotyped prose” 11 and dramatic interpretations of events. Higher recollection of crimes reported in articles adopting these techniques in turn increases perceptions that similar crimes are more likely to occur in the future. 12 U.S. and U.K. studies show fear of crime correlates directly with sensationalism 13 and readers consuming sensational newspapers have a greater fear of crime than their counterparts reading newspapers that present crime in a less sensational manner. 14
The current study set out to test the findings of conventional U.S. and U.K. studies in a new setting, looking at newspapers in Hong Kong. Newspapers play an unparalleled cultural role in Hong Kong, where a population of 7.3 million persons supports 47 commercial newspapers operating with minimal administrative control. 15 The study looked to see whether hypotheses based on U.S. and U.K. studies would also hold true in Hong Kong. It also investigated in more detail the nature of newspaper readers’ perceptions, in particular whether sensational press presentation techniques that generated fear of crime also yielded enhanced awareness of crime by readers.
Study Objects and Methodology
The study is based on the distribution of newspaper reports to a selected readership group and a survey of the readers to determine reactions to the materials read. Three hypotheses were developed to test findings of previous U.S. and U.K. media studies on the impact of sensationalism and news media format and style on perceptions of crime in the context of a fundamentally different cultural environment. The study looks at Chinese-language newspapers in the unique environment of a largely free press spanning a spectrum of support for or against a non-democratic government. The three hypotheses were as follows:
Identification of Target Press and Classification of Target Articles
As noted, Hong Kong is well provided for in terms of access to the press. Twelve major newspapers predominate the Chinese-language press with five most widely recognized. Three newspapers were selected based on their level of support for or against government actions. The government is regarded by the public as the primary guardian of public safety, and crime reporting can thus be used by newspapers as a way of influencing indirectly readers’ faith in the competence of government as the protector of society. The newspapers’ stances vis-à-vis the government were determined using data assembled by the School of Journalism and Communication of the Chinese University of Hong Kong, 16 based on coverage of one seminal political event in modern Hong Kong, the government’s proposal in 2002 to introduce an “anti-subversion” law. Of the five highest profile newspapers, one adopts a neutral to supportive of the government stance (Sing Tao Daily), one clearly opposed (Apple Daily) and three had more ambiguous positions. The newspaper with the highest public profile and circulation of these three, Oriental Daily, was used in the study as representative of the intermediate category.
In contrast to the more limited English language press, which ranges across a spectrum of styles, Chinese-language newspapers in Hong Kong tend to resemble what would be termed “tabloid” style newspapers were they English language publications. The three Chinese-language newspapers reviewed for this study all adopt a broadsheet format, but Apple Daily and Oriental Daily feature tabloid-style sensationalist reporting while the Sing Tao Daily uses a more conventional broadsheet non-sensationalist style. The study did not distinguish between the views of readers who consumed newspapers in hard copy and those who read online. Print remains the most trustworthy media format for young generation in Hong Kong. 17 Also, the content and style are similar in the online and print versions of articles. To simplify content analysis, only the physical newspapers were examined as it was easier to measure features such as article and graphic sizes in this format.
The survey was limited to news reports that focused on violent crime during two months, October 2011 and January 2012. The approach taken in newspaper articles to report crime was subject to content analysis, taking into account the presentation details as well as the actual text of articles. 18
The content analysis examined presentation of violent crime reports over the two full study periods, considering four factors:
The number of articles on violent crimes;
The physical size of violent crimes reports (measured as cm2);
The number of photographs and graphics accompanying violent crime reports; and
The sizes of photographs and graphics accompanying violent crime reports.
Identification of Survey Cohort and Survey Content
The survey subjects were 274 university students enrolled in the City University of Hong Kong. Most students lived at home and had access to newspapers through family subscriptions and on-campus distribution. A written survey was used with a high participation rate attributable to institutional support for the survey. No incentives were offered for participation.
Previous studies on the link between newspaper reporting and readers’ perceptions have spanned a spectrum from a single question about how frequently survey respondents read crime stories they encountered in the newspaper 19 to complex surveys with extensive content analyses of multiple newspapers documenting both style and extent of newspaper reports paired with multiple questions about readership patterns. 20 This study follows the latter precedents. The assumptions for the hypotheses are that, first, press coverage of crime heightens readers’ fear of crime and, second, sensationalist newspapers with more prominent coverage of crime would have a greater impact on perceptions than newspapers with more balanced coverage. To test the hypotheses, the questionnaire measured four variables:
Readership patterns (frequency of press reading and choice of newspaper);
Readers’ recall of reports on violent crimes;
Readers’ anxiety about crime risks to the readers and to others;
Readers’ demographic factors (age and gender).
The dependent variable measured in the survey was fear of crime. Measuring fear of crime is challenging given its wholly subjective and amorphous nature. At best there is considerable confusion among researchers as to what constitutes “fear of crime,” 21 with the multiplicity of divergent meanings in the literature obscuring any attempt to secure specificity. 22 It is uncertain whether what has been measured in some past research as “fear of crime” is actually fear of crime at all, 23 and researchers are clearly concerned about the difficulty in settling on a consistent and specific understanding of fear of crime. 24
An initial step in the measurement of fear of crime is to segregate fear as an emotional and physiological response to danger 25 or feeling of alarm caused by an awareness or expectation of danger 26 from judgments of risk or even general concerns about the crime problem. 27 This can be done by establishing a conceptual framework setting out steps of a reader’s perception of direct nexus to crime ranging from direct personal to general on a vertical axis and placing different types of perception of crime ranging from the cognitive to the affective on a horizontal axis, 28 and then using specific indicators of different perceived risks that effectively measure fear of crime to the respondent directly. 29 Two primary questions were used in this study, the first modified from Lee 30 and the second modified from a questionnaire used by Williams and Dickinson: 31
“How much do you worry about the safety of your loved ones from crime?”
“How much do you fear being a victim of crime?”
Both questions were placed on a seven-category Likert-type scale from 1 (least worry/least fear) to 7 (most worry/most fear) and, respectively, divided into six questions distinguishing fear of molestation, rape, personal assaults, theft, robbery and burglary. Testing the responses to the two groups of questions yielded alpha of .893 for the first group and .889 for the second.
Fear of Crimes: Content Analysis Results
A total of 453 crime stories were analyzed. The pro-government Sing Tao Daily had the lowest number of total crime stories, the lowest number of photographs or graphics in each story and the smallest average size for each photograph or graphic used. The critical Apple Daily held the middle position in terms of the overall number of crime stories, the average number of photographs or graphics in each story and the average size of photographs and graphics. It did, however, have the highest average size for crime articles. The intermediate Oriental Daily had the largest space per article devoted to photographs and graphics (Table 1).
Crime News
Fear of Crime and Awareness of Crime Survey Results
Variable Analysis of Fear of Crime Survey Results
The study sought to measure the impact of crime reporting on fear of crime among readers of three newspapers of different genres. The relative levels of fear of crime held by readers of the three newspapers were compared in terms of the mean score. The 70 respondents who read Oriental Daily frequently had the highest fear of crime mean of 57.50. The 85 respondents who read Apple Daily had a mean score of 51.75. The 52 respondents who read Sing Tao Daily had the lowest fear of crime mean of 40.73.
The one-way analysis of variance (ANOVA) test revealed a notable difference in the fear of crime among the three cohorts, F(2, 204) = 21.737, p < .001, and the “Bonferroni” test showed the results were statistically significant (Table 2). The most significant differences were between the readers of Sing Tao Daily and readers of the other two newspapers. The differences between readers of Apple Daily and Oriental Daily were less stark.
Bonferroni Method Multiple Comparisons of the Dependent Variable Fear of Crime
The mean difference is significant at the .05 level.
Variable Analysis of Awareness of Crime Survey Results
In addition to questions on fear of crime, respondents were asked questions to determine their awareness of crime by testing recollections of 13 specific incidents that were covered in the newspapers. The results were initially unexpected. Readers of the least sensationalist paper with the least fear of crime had the highest awareness of crime. However, more detailed analysis showed the correlations between the three groups of readers between fear of crime and awareness of crime were statistically weak (Pearson’s r = –.195, p < .005) (Table 3).
Multiple Comparisons Dependent Variable: Crime Awareness
The mean difference is significant at the .05 level.
The reliability of the results was tested using one-way ANOVA, which confirmed an overall significant difference on crime awareness among three groups of readers, with an F(2, 181) = 4.372, p < .05. However, the difference was just statistically significant between the readers of Apple Daily and Sing Tao Daily in which the Sig. value was .011, p < .5.
Regression analysis revealed relationships between the three independent variables (gender, crime awareness and choice of newspaper) and the dependent variable (fear of crime). While each of the three independent variables correlates with the dependent variable fear of crime, readers’ gender and choice of paper have more significant relationships with the dependent variable (the Sig. value for each of these is less than .001) than does the third independent variable, crime awareness (p < .005). Sensationalist press readers have a greater fear of crime than their counterparts reading less sensationalist press and females have a greater fear of crime than do males.
Evaluating the Hypotheses
The results of the study were mostly consistent with
Surprisingly, however, and contrary to
The answer to both questions might be found in the results of content analysis. The academic literature suggests sensationalist presentation with graphics and text is more likely to incite emotion and be recalled. It appears, however, that what is being recalled is the emotional response (fear), not the actual details of the crime. Readers of less sensational press are more likely to be aware of the “whole picture” of crimes such as details of the crime, some background to the perpetrator’s possible motives and police investigative work. Consequently, those readers may recall more detail of the crime and, starting with information for a more objective analysis, harbor less fear.
Limitations
The study did not consider other personal attributes of readers apart from gender. As noted, there is a correlation between gender and fear of crime with female readers reporting higher levels of fear than their male counterparts. This may be attributable in part to the fact that almost all perpetrators are males, heightening female vulnerability fears, and in particular, male readers are unlikely to fear sex crimes (almost all sex crimes target females; Figure 1).
Gender Correlation with Crime Perpetrators and Victims of Sexual Assault
There may be similar correlations with other individual or environmental factors not surveyed.
A further limitation is the focus of the study on one form of news media. Readers no doubt have access to a range of other types of news media including television and the internet that may have played roles in molding respondents’ views to a greater extent than did the newspapers. This possibility would be enhanced if readers of particular newspapers gravitated to particular television or internet news sources. The role of the internet, in particular, may be significant. The rise of the internet and the consequent boom of mobile media has created a world of “ambient news” 32 that derives from professional and citizen journalists using blogs, social media and other websites to produce news. 33 The news found in peer-to-peer communications clearly operates within different frames from traditional media. 34
Fourth, the study did not consider the extent to which the news media presentations are based on filtered information. The ultimate source of almost all crime stories reported in the press is the police and this initial filter affecting the information that newspapers can potentially report may also have an impact on the way in which the stories are presented.
Another caveat concerns participants’ personal experience of crime. In pre-survey design interviews, no participants reported any experience with crime, an outcome that may reflect actual experience or may reflect a concern of Chinese respondents that crime and victims of crime are associated with lower socio-economic backgrounds.
Penultimately, the study does not dissect the impact of frequency of and overall use of newspapers on knowledge or fear of crime. While the initial survey tracked newspaper exposure, to exclude the impact other exogenous factors might have had on crime perceptions, the final data set used to evaluate fear of crime was limited to the group that read newspapers often.
Finally, the findings may only be valid in respect of the respondent cohort because of the convenience sample drawn from a discrete subset of the general population. The findings may not represent wider views in Hong Kong.
Conclusion and Future Research
This study investigated the extent to which different reporting styles in three representative major Chinese-language newspapers in Hong Kong had an impact on public fear and awareness of crime. Hypotheses derived from U.S. and U.K. studies proved equally valid in Hong Kong: readers’ fear of crime was related to newspapers’ sensationalistic presentation style and readers of the sensational type of newspaper had higher levels of fear of crime than those of newspapers with less sensational crime news reporting. A wholly unexpected finding was the apparent inverse relationship between awareness of crime and fear of crime. The results suggest that the public’s level of fear of crime is related to the effects of news media “framings.” Sensationalist framing of crime news may have an inverse impact on awareness of crime and this in turn may have an inverse relationship with fear of crime.
The findings in this study suggest three paths for future research. First, random sampling of interviewees with different demographic characteristics should be adopted to overcome the limitation of a convenience sample. Second, mixed methods including both quantitative and qualitative analysis can be used to provide more nuanced and richer results concerning readers’ decoding processes when reading sensational newspaper reports of crime. For example, the study could investigate how readers select particular articles to read in detail and which to scan or skip. Finally, the Hong Kong study could be conducted in mainland Chinese cities. A number of commercial newspapers have sprung up in large urban centers in China since the country began its shift to a market economy, and while the Communist Party retains strict control over political content of newspapers in China, there are no restrictions on crime reporting. It remains to be seen whether U.S. and U.K. studies conducted in democratic full market economies can hold true in a Communist state with significant political censorship and pervading attempts to influence the attitudes of citizens.