Abstract
Background:
Public perceptions of workplace injuries are shaped by media reports, but the accuracy of such reports is unknown.
Objective:
This study identifies differences between workers' compensation claims data and newspaper reports of workplace injuries in Canadian newspapers and media sources.
Methods:
This study applies quantitative content analysis to 245 Canadian English-language newspaper articles from 2009 to 2014. Workers' compensation claims data is drawn from the Association of Workers' Compensation Boards of Canada.
Results:
Newspapers dramatically overreport fatalities, injuries to men, injuries in the construction and mining/quarrying/oil industries, injuries stemming from contact with objects/equipment and fires/explosions, and acute physical injuries such as burns, fractures, intracranial injuries, and traumatic injuries. Newspaper reporters tend to rely upon government, police/firefighter, and employer accounts, rarely recounting the perspectives of workers.
Conclusion:
Newspapers overreported fatalities, injuries to men, and injuries in the construction and mining/quarrying/oil industries. This results in a misleading picture of occupational injuries in Canada.
Keywords: Workplace injury, Canada, Media, Content analysis
Introduction
Workplace injuries represent specific harms experienced by workers. Yet workplace injuries are also a social construction. Social constructions are understandings that emerge from social relations. For example, the behaviors expected of women in the workplace are socially, not biologically, determined. In this study, which injuries experienced by workers are deemed to be “workplace injuries” reflects a shared (albeit not universal) understanding of what this term means. This understanding is informed by personal experience, the experiences of others, and information provided by societal institutions.
Media reports play an important role in the development of shared understandings of many issues, yet media representations of workplace injury are rarely studied. Research on media representations of crime suggests that there is significant misrepresentation and privileging of the views of powerful groups, such as the state and the police. This study uses quantitative content analysis to examine Canadian English-language newspaper reports of workplace injuries and fatalities from 2009 to 2014 in order to answer three questions:
What types of injuries and fatalities are reported in Canadian newspapers, and how do these reports compare with workers' compensation injury and fatality statistics?
What demographic and occupational information is presented in newspapers reports and how do these data compare to workers' compensation claimant data?
Where do journalists seek information about workplace injuries and fatalities?
This study is the first in a three-part project applying quantitative and qualitative content analyses to media representations of workplace injury.
The social construction of workplace injury
In 2012, there were 245 365 successful workers' compensation claims filed for injuries requiring time away from work and 977 claims for workplace fatalities.1 These statistics underreport the true level of workplace injury in Canada by excluding injuries not requiring time away from work (even if they required medical attention), injuries to workers without workers' compensation coverage, and unreported injuries – which some studies put at 40% of all injuries.2–5
These systematic exclusions of injuries from official injury counts reveal the dual nature of workplace injuries. On the one hand, workplace injuries are specific harms experienced by workers. On the other hand, workplace injuries are social constructions in that they are the byproduct of choices made by various actors that contribute to a shared understanding of what is and is not a workplace injury. For example, the “arises and occurs” test (requiring workers to show a direct link between an injury and a work-related cause) applied by Canadian workers' compensation boards to determine if a workplace injury is compensable legitimizes some injuries, such as physical injuries with clear injury mechanisms. These same tests delegitimize other forms of injury – such as occupational diseases, which typically have long latency periods and murky causality – by declaring them to be non-compensable and excluding them from official statistics of workplace injuries.6
Over time, this bias toward acute physical injuries is consistent with and may reinforce long-held skepticism about the work-relatedness of certain injuries. For example, the recognition of asbestos-related disease occupational diseases in Canada took a long time and the characterization of asbestos as not intrinsically harmful continues to maintain lucrative export markets in the developing world.7,8 There has been similar resistance to the recognition of repetitive strain injuries, particularly among women.9,10 Such skepticism has denied injured workers access to workers' compensation benefits they might otherwise have been due. More broadly, constructing such injuries as not work related has limited the public policy response (e.g. limiting or prohibiting use of hazardous materials or work processes), which in turn has allowed employers to continue trading workers' health for profits.11
This view of workplace injuries as social constructions is based upon the theory that individuals “create” reality by what stimuli they pay attention to and how they interpret that stimuli – choices driven by individuals' values, beliefs, and experiences.12 Personal experience (or “experienced reality”) is a powerful but limited source of knowledge.13 Knowledge gleaned from interacting with other people in our social groups, institutions, and the media (“symbolic reality”) – knowledge that comprises much of what we know about the world – results in the development of shared reality.14,15 Thus, to use the above example, the under-acceptance of certain types of injury and illness in the workers' compensation system feeds into and entrenches a social acceptance that those occurrences are not work related and therefore do not require employer, society, or government action to remediate them.
The media plays an important role in social construction processes because it is a source of symbolic reality and a channel used by claims makers to shape issue perceptions.16 More specifically, what the media choose to report as “news,” and how they frame the news helps create and reproduce societal narratives about what is newsworthy and how people should respond (i.e. what is the problem, who is to blame).
One outcome of combining our experienced and symbolic realities to socially construct our world is the development of frames – clusters of factual and interpretative claims – that help us organize our understandings of events and prescribe actions and policies.17 For example, if we adopt the frame that workplace injuries are caused by worker carelessness, we will likely identify a different public policy response to workplace fatalities compared to someone who adopts the frame that workplace injuries are the result of employer choices under the profit imperative of capitalism.18 Frames can be supplemented by narratives: narrow preestablished social constructions that are similar to rhetorical idioms (i.e. recurring rhetorical elements and speech that culturally anchor social problems).19 For example, the widespread view of young workers (especially, young males) as prone to risk-taking is a narrative used to quickly (and perhaps, inaccurately) characterize the cause of injury events that, in turn, supports a broader framing of injuries as caused by worker carelessness.20 The existing cultural stock of narratives and frames provides storylines that assist the media in making events understandable to readers, efforts that also reinforce these cultural stocks.
There are numerous actors who shape the symbolic reality of workplace injuries in Canada, including workers, employers, policymakers, workers' compensation boards, and trade unionists. For example, conservative politicians in the Canadian province of Alberta have advanced the view that injuries are caused by ignorance and inattention, workplaces are safe and getting safer, risk is inevitable, and hazard mitigation is too expensive.21,22 Some actors – by virtue of their knowledge and position – are better able to convey their message authoritatively. For example, provincial government health-and-safety spokespeople often inform the media about the occurrence of workplace injury events. As the first (and often only) source of information about these events, these spokespeople shape how injuries are reported.
While the media is an important source of symbolic reality about workplace injuries, what the media says about workplace injuries and the accuracy of these reports is largely unstudied.23 The importance of media framing of issues can be seen in other areas. For example, research into media portrayals of criminal activity suggests that coverage overrepresents the occurrence of violent crime, skewing public perceptions of crime.24–29 It is unclear what explains these discrepancies. Among the possible explanations are story templates (i.e. pre-existing frames used to structure reports) that assist reporters facing time and resource pressure.30 A related issue is the tendency of the media to rely upon authority figures (e.g. police, government) for information, which may constrain who participates in the social construction of issues.31–33
Methodology
The objective of this study was to answer the following questions:
What types of injuries and fatalities are reported in Canadian newspapers, and how do these reports compare with workers' compensation injury and fatality statistics?
What demographic and occupational information is presented in newspapers reports, and how does these data compare to workers' compensation claimant data?
Where do journalists seek information about workplace injuries and fatalities?
Quantitative content analysis reveals patterns in large bodies of textual materials by systematically coding and counting the manifest (i.e. literal, rather than latent) content of a communication.34,35 This technique makes it possible to see patterns of inclusion, exclusion, and description in a dataset that are hard to identify when simply reading the materials.36 The dataset for this study contains 245 newspaper articles that appeared in 31 Canadian English-language daily newspapers between 1 July 2009 and 30 June 2014. Articles were retrieved from the FPinfomart database using search terms that included “workplace fatality,” “injured worker,” “workplace accident,” and “workplace death.” Our search resulted in the identification of 856 articles. Review of the dataset eliminated 611 articles because they were duplicate reports (i.e. wire stories that appeared in multiple newspapers) or did not report on workplace injuries (i.e. used the search terms in an unrelated fashion). Coding was completed in three stages. We developed an initial coding scheme containing 18 measures relevant to the research questions and consistent with the Association of Workers' Compensation Boards of Canada (AWCBC) classification scheme. This coding scheme was refined after a pilot with 50 articles. Both authors coded the full dataset (245 articles) manually, following four small amendments. Results were then screened to identify and resolve a small number ( < 15) coding discrepancies.
Canada-wide workers' compensation time-loss injuries and fatalities claims data for 2012 as published by the AWCBC were used to answer the first and second research questions. Workers' compensation data provide an imperfect proxy for actual workplace injuries. These data focus only on injuries resulting in a fatality or where the worker required time away from work due to injury. Injuries requiring modified work and/or medical aid that did not result in a time-loss and injuries requiring only first aid or no treatment are excluded from these data. Further, accepted workers' compensation claims differ from overall injuries due to the absence of coverage of some sectors and types of employment, differing rates of claim acceptance by injury type, and low claim-filing rates by some workers.3,4 While these issues are concerning, AWCBC data provide useful information about the overall distribution of “serious” workplace injuries by age, gender, and industry. Further, medical aid only claims and other non-reported injuries are unlikely to trigger media reports; therefore, it is reasonable to consider the AWCBC data relatively comparable to the dataset being analyzed.
Results
When compared to national workers' compensation claims data, newspapers were found to dramatically over-report fatalities, injuries to men, and injuries in the construction and mining/quarrying/oil industries. Newspapers also overreported injuries stemming from contact with objects/equipment and fires/explosions as well as acute physical injuries such as burns, fractures, intracranial injuries, and traumatic injuries. By skewing public perceptions of workplace injury, this pattern of bias may affect the public policy response. Newspaper reporters also tended to rely upon government, police/firefighters, and employer accounts, rarely recounting the perspectives of workers. As set out in Tables 1–6, the differences between media reports and AWCBC data (except those in Table 5 related to age of victim) were statistically significant at the P < 0·01 level.
Table 1. Reported and actual workplace injuries and fatalities1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Fatalities | 61·2% | 0·4% |
| Injuries | 31·0% | 99·6% |
| Combined reports | 0·8% | n/a |
χ2 = 95·855, P < 0·01.
Table 5. Reported and actual workplace injuries and fatalities by injury age1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Under 20 | 4·1% | 3·3% |
| 20–24 | 12·2% | 9·5% |
| 25–29 | 15·7% | 10·5% |
| 30–34 | 13·6% | 10·5% |
| 35–39 | 15·7% | 10·8% |
| 40–44 | 9·9% | 11·5% |
| 45–49 | 9·9% | 13·3% |
| 50–54 | 5·8% | 13·5% |
| 55–59 | 7·0% | 9·8% |
| 60–64 | 3·4% | 5·3% |
| 65+ | 2·3% | 2·3% |
χ2 = 13·690, n.s.
The most striking result of the analysis was the overrepresentation of workplace fatalities in newspaper reports. As set out in Table 1, occupational fatalities account for 61·2% of newspaper reports, yet represent only 0·4% of all time-loss and fatality claims in Canada. In considering this finding, it is important to remember that the AWCBC statistics already overrepresent the proportion of workers' compensation claims that fatalities comprise because the AWCBC data exclude injuries that result in modified work and medical aid claims as well as minor injuries for which no claims are filed.
Newspapers also overrepresented injuries and fatality caused by contact with objects/equipment and fires/explosions. The more common bodily reactions and exertions are not mentioned in newspaper coverage (Table 2). Other notable differences include underreporting of assault/violent acts, exposures to harmful substances, and overreporting of transportation incidents. Similarly, acute physical injuries such as burns, fractures, intracranial injuries, and traumatic injuries were overrepresented in newspaper reports (Table 3). By contrast, the most common type of injury (i.e. sprain/strain) is vastly underreported along with bruises and contusions. Given that AWCBC data only include time-loss injuries and fatalities (i.e. severe injuries), it is likely that the true proportion of all sprain-and-strain injuries is much higher than the comparative data suggest, and thus, the degree of newspaper underreporting is much greater.
Table 2. Reported and actual workplace injuries and fatalities by injury event1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Assaults/violent acts | 0·8% | 2·8% |
| Bodily reactions/exertion | 0·0% | 42·1% |
| Contact with objects/equipment | 53·0% | 23·8% |
| Fires/explosions | 5·9% | 0·2% |
| Falls | 17·8% | 18·1% |
| Exposure to harmful substances | 2·4% | 5·7% |
| Transportation | 5·9% | 3·3% |
| Other | 5·9% | 3·3% |
| Unknown | 8·3% | 0·7% |
χ2 = 68·874, P < 0·01.
Table 3. Reported and actual workplace injuries and fatalities by injury type1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Burn | 4·5% | 1·9% |
| Bruise/contusion | 3·8% | 10·2% |
| Fracture | 29·5% | 8·0% |
| Intracranial injury | 6·8% | 2·1% |
| Cuts/laceration/puncture | 3·0% | 8·0% |
| Sprain/strain | 0·3% | 47·6% |
| Traumatic injury/disorder/complication | 22·7 | 18·2% |
| Other | 6·1% | 3·6% |
| Unknown | 23·1% | 0·5% |
χ2 = 90·552, P < 0·01.
We also found gender-bias in newspaper reports. As shown in Table 4, men account for 62·9% of accepted time-loss and fatality claims, but are featured in 95·6% of newspaper reports about workplace injuries and fatalities. Newspaper reports also tend to over-report injuries to workers under age 40 (Table 5), particularly to workers aged 25–39. Injuries in the construction and mining/quarrying/oil industries were significantly over-reported by newspaper reports, while injuries in the health/social services and retail industries are significantly underreported (Table 6).
Table 4. Reported and actual workplace injuries and fatalities by gender1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Male | 95·6% | 62·9% |
| Female | 4·4% | 37·1% |
χ2 = 32·512, P < 0·01.
Table 6. Reported and actual workplace injuries and fatalities by industry1.
| Newspaper reports, n = 245 | Official statistics, n = 246 342 | |
| Agriculture | 3·2% | 1·3% |
| Business services | 4·4% | 2·7% |
| Construction | 30·3% | 11·3% |
| Education | 0·3% | 2·7% |
| Fishing/trapping | 3·2% | 0·2% |
| Government | 1·6% | 6·6% |
| Health/social services | 0·0% | 16·8% |
| Finance/insurance/real estate | 0·0% | 0·8% |
| Logging/forestry | 3·6% | 0·6% |
| Manufacturing | 8·4% | 15·6% |
| Other services | 3·6% | 3·5% |
| Mining/quarrying/oil | 14·3% | 1·0% |
| Retail | 4·0% | 11·0% |
| Accommodation/food/beverage | 0·8% | 5·9% |
| Transport | 7·6% | 7·0% |
| Communication/utilities | 7·6% | 2·0% |
| Wholesale | 0·3% | 4·7% |
| Unknown | 6·8% | 6·4% |
χ2 = 328·270, P < 0·01.
While most newspaper reports indicated the injured or killed worker's gender and a majority indicated the injured or killed worker's age, Table 7 shows that only a minority of reports indicate the injured or killed worker's name, family status, experience, training, or performance. Overall, relatively little information about the victims of workplace injuries and fatalities is reported.
Table 7. Reported personal characteristics of injured and killed workers (n = 245).
| Worker gender | 93·1% |
| Worker age | 64·5% |
| Worker name | 39·2% |
| Worker family status | 21·2% |
| Worker experience | 13·1% |
| Worker training | 5·3% |
| Worker performance | 4·1% |
Finally, the data reported in Table 8 demonstrate that journalists are most likely to rely upon reports from government spokespeople, police officers/firefighters, and the employer for information about workplace injuries. The perspectives of workers, workers' family and friends, and union representatives are rarely included in newspaper reports.
Table 8. Journalist workplace injury and fatality information sources (n = 245).
| Politicians | 4·6% |
| Bystander | 2·6% |
| Employer | 17·5% |
| Family/friends | 8·6% |
| Government spokesperson | 23·8% |
| Lawyer/court/corner | 1·4% |
| Medical professional/paramedic | 0·3% |
| Other | 1·7% |
| Police officer/firefighter | 23·5% |
| Union representative | 5·2% |
| Victim | 2·0% |
| Other worker | 2·6% |
| Unknown | 6·3% |
Discussion
A typical English-language newspaper story about a workplace injury in Canada reports the death of a male construction worker, somewhere between 19 and 40 years old, who was crushed by contact with an object or equipment. The tendency of newspapers to overreport fatalities, injuries to men, injuries in the construction industry, and injuries caused by contact provides readers with a misleading picture of who is injured and how they are injured in Canadian workplaces. This finding is important because unrepresentative media reports, which play an important role in shaping our symbolic reality, pose a significant threat to the validity of the views of workers, employers, and policymakers about workplace injury. Of particular note is the significant gender-bias evident in media reports. Injuries to women are mentioned in only 4·4% of newspaper reports, even though women comprise 37·1% of workers' compensation claimants. The invisibility of women in newspaper reports may be partly explained by the media's bias toward fatalities (men comprise 95·6% of fatalities in the workers' compensation claims data) and injuries in the construction and mining/quarrying/oil industries (predominantly, male industries).1,37 That said, it is important to recognize the long history of dismissing occupational injuries and diseases reported by female workers.9,38 The non-reporting of female injuries and fatalities perpetuates this gender-bias by reinforcing the construction of workplace injuries as primarily a male concern.
Fatalities are also dramatically overrepresented in media reports. Despite comprising fewer than 0·4% of compensation claims, fatalities were mentioned in 62·2% of newspaper reports. Newspapers' focus on fatalities may reflect that, while workplace injuries are common, workplace fatalities are rare enough to be notable and even shocking. In addition to the shock value of fatalities, reporters and editors may emphasis fatalities because of properties intrinsic to fatalities: they have a high salience to local readers, they can be conveyed in relatively little space, and they can be written from a government press release and/or after speaking with only government and/or emergency responder sources. These aspects of fatalities may make them attractive to reporters and editors facing pressure to quickly report events while facing diminishing newsroom budgets.39 The overreporting of fatalities has two, possibly paradoxical, effects. On the one hand, it creates the perception that workplace injuries are fatal more often than is the case. This likely heightens readers' concern about the risk associated with such injuries. At the same time, focusing attention on the relatively few workplace fatalities (977 in 2012) masks the vastly larger issue of non-fatal injuries (245 365 in 2012).1 Emphasizing relatively rare injuries at the expense of reporting common injuries likely reduces readers' perception of the probability of experiencing a workplace injury. This, in turn, likely reduces the relative importance readers' accord to injury prevention efforts. In effect, overreporting catastrophic (but relatively rare) events and portraying those events as restricted to a small number of occupations contributes to the construction of workplace injuries and fatalities as tragic but uncommon events – a social construction that belies the hundred of thousands of workplace injuries that occur each year in Canada.
Similarly, while injuries caused by contact with objects and equipment comprise only 23·8% of injury events in the AWCBC data, such injuries were mentioned in 53·0% of newspaper articles. By contrast, bodily reactions/exertions are never mentioned in newspaper articles but comprise 42·1% of AWCBC injury events. A similar bias toward infrequent but dramatic physical injuries (e.g. burns, fractures, traumatic injuries) is also evident while more commonplace injuries (e.g. sprains and strains) are largely absent from newspaper reports. Again, this may reflect an editorial preference for shocking injuries. Whatever the reason for not reporting such injuries, their exclusion from media reports likely reinforces the understanding that low-level injuries are an unavoidable part of employment.40 That is to say, newspapers report notable (generally, exceptional) events, and the exclusion of common injuries essentially normalizes these events. In effect, emphasizing dramatic and physical injuries renders less visible the most common forms of workplace injury. Such injuries are also among the most easily prevented through engineering and administrative controls and are more likely to be experienced by women.
Interestingly, the victims of a workplace injury rarely figure in newspaper stories. While a victim's age and gender are often provided, their names usually are not. Referring to a victim by their role in the workplace (i.e. a worker) makes the victim an object, rather than a person, thereby dehumanizing the victim. It is also uncommon for newspapers to discuss injured workers' family status, training, experience, or performance. Personal information presented and omitted may reflect the limited data provided by official announcements combined with time and resource pressures. Overall, the effect is to construct the victims of injuries and fatalities in police-blotter terms (e.g. “male, 34, construction”) rather than as fully fleshed out people (with families and interests and histories) that readers can relate to more easily. When this depersonalization of injured workers is combined with the narrow range of occupations featured in such reports, the result may be a distancing – a sense among most readers that workplace injuries and fatalities are not something they need to be concerned by in their own lives, as they happen to “other” people. This distancing may be intensified by the absence of any follow-up stories. Workplace injuries and fatalities are presented as one-off events. There is little consideration that an injury is the beginning of process for the victim and his or her family. An exception to this depersonalization of “workers” tends to occur in longer, retrospective newspaper reports about injuries or fatalities that often occur in conjunction with the National Day of Mourning for injured workers. Such pieces tend to provide a much fuller picture of the victim and his or her life. Nevertheless, these stories, while putting a human face on workplace injury and death, also strip the incident from its work context by removing aspects of what happened and why. In this way, the deaths are reported as tragedies for which no one is really responsible and which cannot be easily prevented, once again reinforcing social frames regarding workplace injury.
Journalists most commonly used information provided by government spokespeople (23·8% of stories) and police or fire officials (23·5%). Journalists also used information provided by employers in 17·8% of stories. Less utilized were comments from less socially powerful actors, such as the victim (2·0%), other workers (2·6%), friends and family of the victim (8·6%), and union representatives (5·2%). Bystanders were also featured in only 2·6% of newspaper stories. This reliance by reporters on “official” sources is consistent with the notion that time and resource limitations may constrain reporting. The use of powerful actors' views may disproportionately shape the symbolic reality around workplace injuries and fatalities allowing employers and policymakers to influence individual reports in ways that collectively shape public perceptions of workplace injuries.
In summary, this study finds that newspaper reports of workplace injuries present a profoundly misleading picture of workplace injuries and fatalities in Canada. Given the importance of media reports in our symbolic reality, misleading reports potentially skew the social construction of occupational injury. Of particular concern is the virtual invisibility of injuries to women and non-acute injuries, such as repetitive strain injury and occupational diseases. To the degree that newspaper report results in the public misunderstanding the nature and frequency of workplace injury, misleading reporting has significant public policy implications. Such reporting can retard demands for safer workplaces by “hiding” the true level of risk faced by workers. It also can bolster employer and policymaker claims that workplaces are safe and getting safer.
This study also finds that the newspaper reporting tends to focus on the injury, rather than the worker. Injured workers are usually nameless victims, known only by their gender, age, and occupation. The non-work roles of injured workers – parents, spouses, community members – are rarely mentioned. This may reflect that journalists appear to draw most of their information about workplace injuries from official sources – government emergency responders and employers. There is little evidence in the dataset of first-hand reporting or follow-up reporting of injuries. This raises questions about the validity and reliability of newspaper reports. While such reports often become the official record of public events, the absence of investigation and scrutiny suggests a potential for powerful actors (such as policymakers and employers, who have electoral and economic interests in the social construction of workplace injuries) to systematically influence what is reported and how it is reported by “pre-digesting” events and providing this information to reporters.
This study has a number of limitations. Although the newspaper dataset is drawn from daily newspapers in all provinces and one territory, all the newspapers are English-language papers. It is possible that workplace injury is treated significantly differently in French-language papers (primarily published in the province of Quebec). There may also be provincial or regional differences in reporting. Owing to the small size of the dataset, we did not perform inter-provincial or regional comparisons. Further, the use of daily papers (predominantly urban-based) excludes assessing the portrayal of workplace injury in weekly newspapers (primarily published in rural areas). Rural–urban differences are the subject of a separate analysis.
While this study provides useful data about how newspapers portray workplace injuries, it is unclear to what degree these portrayals affect public perceptions. It is expected that media reports are an important source of information about workplace injuries. But the widespread nature of workplace injury in Canada suggests the public may also have first- or second-hand experiences with workplace injuries. Future research might profitably examine how readers interpret such newspaper stories and the degree to which media reports influence their overall perceptions of workplace injuries.
Further, while quantitative content analysis provides a useful overview of what newspapers do and do not report about workplace injuries, qualitative analysis of reporting offers a way to draw out narratives (or the “frames”) present in newspaper coverage of workplace injuries and fatalities, particularly around the responsibility for and inevitability of workplace injuries. This framing analysis of workplace injury coverage (for which there are no antecedents) will draw upon techniques used to analyze the representation of female injuries and fatalities in domestic abuse cases.
Disclaimer Statements
Contributors The authors contributed equally.
Funding None.
Conflicts of interest None.
Ethics approval Not required.
References
- 1.Association of Workers'Compensation Boards of Canada. National work injury, disease and fatality statistics 2010–2012. Ottawa: Association of Workers; 2014. [Google Scholar]
- 2.Arthurs H. Funding fairness: a report on Ontario's workplace safety and insurance system. 2014. [document on the Internet]. Toronto: WSIB Funding Review, [cited 2014 Jul 10]. Available from: http://www.wsib.on.ca/files/Content/FundingReviewFundingFairnessReport/FundingFairnessReport.pdf. [Google Scholar]
- 3.Barnetson B. The validity of Alberta safety statistics. Just Lab. 2012;19:1–21. [Google Scholar]
- 4.Cox R, Lippel K. Falling through the legal cracks: the pitfalls of using workers' compensation data as indicators of work-related injuries and illnesses. Pol Prac Health Safety. 2008;6(2):9–30. [Google Scholar]
- 5.Shannon H, Lowe G. How many injured workers do not file claims for workers' compensation benefits? Amer J Ind Med. 2002;42(6):467–73. doi: 10.1002/ajim.10142. [DOI] [PubMed] [Google Scholar]
- 6.Barnetson B. The political economy of workplace injury in Canada. Edmonton: Athabasca University Press; 2010. [Google Scholar]
- 7.Brophy J, Keith M, Schieman J. Canada's asbestos legacy at home and abroad. Int J Occ Env Health. 2007;13(2):236–43. doi: 10.1179/oeh.2007.13.2.236. [DOI] [PubMed] [Google Scholar]
- 8.Storey R, Lewchuk W. From dust to dust: asbestos and the struggle for worker health and safety at Bendix Automotive. Labour. 2000;45:103–40. [PubMed] [Google Scholar]
- 9.Dembe A. Occupational and disease: How social factors affect the conception of work-related disorders. New Haven: Yale University Press; 1996. [Google Scholar]
- 10.Reid J, Ewan C, Lowy E. Pilgrimage of pain: the illness experiences of women with repetition strain injury and the search for credibility. Soc Sci Medi. 1991;32(5):601–12. doi: 10.1016/0277-9536(91)90295-n. [DOI] [PubMed] [Google Scholar]
- 11.Kotelchuck D. Asbestos: the funeral dress of kings – and others. In: Rosner D, Markowitz G, editors. Dying for a living: workers' safety and health in twentieth-century America. Bloomington: Indiana University Press; 1989. pp. 192–207. [Google Scholar]
- 12.Lindlof T. Media audiences as interpretive communities. In: In: Anderson J, editor. Communications yearbook. 11th edn. Newbury Park: Sage; 1988. pp. 87–107. [Google Scholar]
- 13.Maxson C, Hennigan K, Sloane D. Factors that influence public opinion of the police. Washington: National Institute of Justice; 2003. [Google Scholar]
- 14.Adoni H, Mane S. Media and the social construction of reality. Communication Res. 1984;11(3):323–40. [Google Scholar]
- 15.Spector M, Kitsuse J. Constructing social problems. Hawthorne: de Gruyter; 1987. [Google Scholar]
- 16.Best J. Images of issues: typifying contemporary social problems. New York: de Gruyter; 1999. [Google Scholar]
- 17.Best J. The diffusion of social problems. In: Best J, editor. How claims spread: national diffusion of social problems. New York: de Gruyter; 2001. pp. 1–18. [Google Scholar]
- 18.Graebner W. Doing the world's unhealthy work: the fiction of free choice. Hastings Cent Rep. 1984;14(4):28–37. [PubMed] [Google Scholar]
- 19.Ibarra P, Kitsuse J. Vernacuar constituents of moral discourse: an interactionist proposal for the study of social problems. In: Holstein J, Miller G, editors. Reconsidering social construction. New York: de Gruyter; 1993. pp. 25–58. [Google Scholar]
- 20.Barnetson B. Politically, how do you make it relevant? Kill more young people.' The prospects for greater enforcement of teen employment laws in Alberta, Canada. Edmonton: Athabasca University; 2014. [Unpublished manuscript] [Google Scholar]
- 21.Barnetson B. Framing and blaming: construction of workplace injuries by legislators in Alberta. Canada. Int J Occ Env Health. 2013;19(4):332–43. doi: 10.1179/2049396713Y.0000000040. [DOI] [PubMed] [Google Scholar]
- 22.Barnetson B, Foster J. Bloody Lucky: the careless worker myth in Alberta, Canada. Int J Occ Env Health. 2012;18(2):135–46. doi: 10.1179/1077352512Z.00000000020. [DOI] [PubMed] [Google Scholar]
- 23.Deneberg R, Deneberg T. Workplace violence and the media: the myth of the disgruntled employee. Work. 2012;42(1):5–7. doi: 10.3233/WOR-2012-1321. [DOI] [PubMed] [Google Scholar]
- 24.Altheide D. Creating fear: news and the construction of crisis. Hawthore: deGruyter; 2002. [Google Scholar]
- 25.Sprott J. Understanding public views of youth crime and the youth justice system. Can J Crim. 1996;38(3):271–90. [Google Scholar]
- 26.Hera-Rapp A. The social construction of local school violence threats by news media and professional organizations. Soc Inq. 2003;73:545–74. [Google Scholar]
- 27.Lowry D, Nio T, Leitner D. Setting the public fear agenda: a longitudinal analysis of network TV crime reporting, public perceptions of crime, and FBI crime statistics. J Comm. 2003;53(1):61–73. [Google Scholar]
- 28.Weitzer R, Kubrin C. Breaking news: how local TV news and real-world conditions affect fear of crime. Just Quart. 2004;21(3):497–520. [Google Scholar]
- 29.Pollack J, Kubrin C. Crime in the news: how crimes, offenders and victims are portrayed in the media. J Crim Just Pop Cul. 2007;14(1):59–83. [Google Scholar]
- 30.Kitzinger J. Media templates; patterns of association and the (re)construction of meaning over time. Med Cult Soc. 2000;22:61–84. [Google Scholar]
- 31.Schneider B, Chamberlain K, Hodgetts D. Representations of homelessness in four Canadian newspapers: regulation, control and the social order. J Soci Soc Welf. 2010;37(4):147–72. [Google Scholar]
- 32.Rauhala A, Albanese P, Ferns C, Law D, Haniff A, Macdonald L. Who says what: election coverage and sourcing of child care in four Canadian dailies. J Child Fam Stud. 2012;21(1):95–105. [Google Scholar]
- 33.Schlessinger P, Tumber H. Reporting crime: the media politics of criminal justice. Oxford: Claredon Press; 1994. [Google Scholar]
- 34.Berelson B. Content analysis in communications research. New York: Free Press; 1952. [Google Scholar]
- 35.Bryman A. Social research methods. Toronto: Oxford University Press; 2001. [Google Scholar]
- 36.Krippendorff K. Content analysis: an introduction to its methodology. 2nd edn. Oaks: Sage Thousand; 2004. [Google Scholar]
- 37.Canada Employment by industry and sex. CANSIM Table 282-0008. 2014. [document on the Internet]. Ottawa: Statistics Canada; 2014 [cited 2014 Jul 10]. Available from: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/labor10a-eng.htm. [Google Scholar]
- 38.Messing K. One-eyed science: occupational health and women workers. Philadelphia: Temple University Press; 1998. [Google Scholar]
- 39.O'Neil D, Harcup T. News values and selectivity. In: In: Wahl-Jorgensen K, Hanitsch T, editors. Handbook of journalism studies. New York: Routledge; 2009. pp. 161–74. [Google Scholar]
- 40.Breslin C, Polzer J, MacEachen E, Morrongiello B, Shannon H. Workplace injury of “part of the job”? Towards a gendered understanding of injuries an complaints among young workers. Soc Sci Med. 2007;64(4):782–93. doi: 10.1016/j.socscimed.2006.10.024. [DOI] [PubMed] [Google Scholar]