Introduction
Hey, I’m Bob. I teach labour stuff at Athabasca University. What I’m here to talk about today is how newspaper reporting about occupational injury is misleading and how that ultimately distorts public perceptions and public policy.
So let's start with some basic injury stats in Canada. The best data we have—but its not necessarily very good—comes from provincial and territorial workers’ compensation boards. These WCBs accept claims from injured workers who are seeking medical aid and wage-loss replacement.
Anyone want to guess the number of fatalities in Canada in 2016? Officially 904, of which 542 were from disease and 312 were from physical injury.
The problem with workers’ compensation data is that tends to under-report injuries. About 15% of folks are not in the ambit of the system so their injuries are not reported. Many of claims get rejected—like the 250 cancer claims at the GE plant in Peterborough.
Some fatalities also fall outside the narrow rules of WCBs. For example, if you are traveling to work and get killed, you don’t get counted, even though the injury would not have happened but for your need to go to work.
People who kill themselves due to work stress are mostly excluded. People who are killed due to occupational incidents but who are not workers (such at the residents of Lac Megantic when the tanker train derailed and exploded) are excluded.
Most numerically concerning are occupational illnesses that are not reported. Cancer and chronic obstructive pulmonary disease are two of the biggest sources of these deaths. They tend not to be reported because of long latency periods, worker ignorance of relevant occupational exposures, and murky causality.
All in, the annual death toll from work-related causes in Canada is 10 to 13 times what the official statistics show. So 10,000 deaths a year is a good solid number. That is a small city or the whole north shore of PEI killed every year by work.
Interestingly, fatalities are actually fairly uncommon. Other forms of injury are much more common. We have Canada-wide data on lost-time claims—basically an injury where a worker couldn’t go to work the next day so they “lost time”. Anyone want to guess how many lost-time claims there were in 2016? 240,682. Those claims represent about 60% of all lost-time claims—so the real number is over 400,000 serious injuries per year. So that’s like everyone in Saskatoon and Regina—kids, parents, grandparents—getting seriously injured each year.
Do those stats surprise you? Yeah, so what those stats tell us is that work is very, very dangerous. So let’s turn to newspaper reporting of workplace injuries and see what we find.
How accurate is reporting of work-related injury?
My colleague Jason Foster and I have been examining newspaper reports of injury. We’re going to look at national data in this next set of slides, but the results are pretty consistent nationally, by province, and between urban dailies and rural weeklies.
The most obvious finding is that newspapers report a minority of all injuries and fatalities. We can’t confidently estimate how much under-reporting there is but it’s fair to say that fewer than 1% are reported. And that’s understandable: not every injury is newsworthy. But an interesting question is, which injuries are newsworthy? Any guesses?
So we started by recording the ratio of fatalities to lost-time injuries in newspaper stories. We then compared that ratio to official statistics to get a sense of whether newspaper reports over- or under-emphasized injuries or fatalities.
Approximately 61.2% of newspaper reports (on the left) were about fatalities (which is the green part of the pie). When you look at official stats (on the right), you can see that fatalities actually comprise 0.4% of all serious injuries. So, basically, newspapers vastly over-report fatalities. Or, put another way, fatalities are much more newsworthy than injuries.
We then decided to look at the gender distribution of newspaper injury reports. Anyone want to guess what percentage of reported occupational injuries are actually experienced by women? Nationally, it was 37.1%. Anyone want to guess what percentage of newspaper injury reports were about women?
About 4.4% of newspaper reports (on the left) featured women (who are yellow). Official stats on the right show the true percentage is much higher. Basically, newspapers vastly over-report injuries to men. Or, put another way, injuries to men are apparently more newsworthy than injuries to women.
We then looked at what kind of injuries were reported. Newspapers are again green and official stats are yellow. Key take-aways are newspapers over-report fractures and traumatic injuries while entirely ignoring the most common kind of injury—a strain or sprain. Basically, if you got blowed up, or crushed, or burned to death, you made the paper. Otherwise, your injury wasn’t newsworthy.
We then looked at which industries received the most coverage for injuries and compared that to official stats. On the left we have the industries with the most newspaper report. On the right, we have the most injurious industries. You can see that that newspapers basically mis-represent which industries are most injurious. This pattern is likely a knock-on effect from the kinds of injuries that get reported, which are fatalities to men caused by a catastrophic event.
Finally, we looked at who reporters quoted in their stories as a way to try and quantify who were the key sources. Reporters talked to government representatives and first responders most often, followed by the employer. They almost never talk to workers or their union representatives.
So, if you relied on newspapers for your information about workplace injuries, you’d likely think injuries are fatalities caused by violent physical events that happen to blue-collar men. This is, of course, profoundly misleading.
How is work-related injury framed in newspapers?
Now while aggregate analysis of newspaper coverage tells us some interesting things, we can also learn a lot by looking at specific stories. The thing that jumps out when you start to read dozens and dozens of these stories at one time is that they all start to sounds the same. In fact there are three basic story templates (or frames) that reporters use when writing about workplace injury.
The first is what we call under investigation and it almost always reads like this example. The reporter notes a worker has been killed and the authorities are investigating. About 55% of all newspaper reports of injuries and fatalities look almost exactly like this—literally you can just swap out the proper nouns and, voila, you are a PostMedia journalist—your layoff notice is in the mail!
The key message in this frame is that the authorities have the situation under control. The passive voice (“the employee has been killed”) focuses attention on the victim rather than on what killed the worker or who was responsible for the death. Rarely do we see follow-up stories—instead the even, like the worker, just passes on.
This report is atypical in that it also names the injured worker and where he lived. Normally, the worker is described in generic terms such as “a 33-year-old carpenter”. Such sparse descriptions tend to dehumanize the victim by framing them as a nameless job-holder rather than in a more relatable way.
The second media frame is very similar and we call it before the courts. It comprises articles reporting charges filed or resolved under provincial OHS laws. They usually look a lot like this one and recount the facts of the case (similar to other court reporting) and the penalty(s) imposed. The reports use technical and passive language.
While the issue of cause and blame can’t be avoided in these articles, they narrow the focus to legal culpability, thereby downplaying how the employer’s action (or inaction) killed or maimed a worker—in this case, a preventable electrocution due to employer inattention. Typically, the matter is discussed as simply a regulatory violation that’s addressed via a fine—a violation little different than being fined for speeding or jaywalking. What really happened here was involuntary manslaughter.
Finally, we have the human-tragedy frame. These articles typically recount the life story of a dead worker and most often appear around the National Day of Mourning in April. They often include an abbreviated summary of the incident followed by reminiscences about the worker’s interests, character, life history or social roles as told by a family member, friend or co-worker. These reports are highly idiosyncratic but the broad message is that the worker’s injury or death was a tragedy.
There are two key aspects of this frame. The first is that the “tragedy” is often accompanied by the word “accident” and implies the incident was unforeseeable and unavoidable. In this way, the human-tragedy narrative elides any discussion of wrongdoing, cause, or culpability.
Second, the tragedy in these articles is the personal loss and emotional suffering of the families. Focus is taken away from the workplace and put on the workers’ loved ones. In this way, the human-tragedy frame encourages readers to think of the worker, not as a worker, but as a father or son with interests and families--drawing out sympathy but, in doing so, removing the workplace context. Consequently, this frame discourages the reader from linking the human elements of the incident to economic, political, and structural factors giving rise to whatever caused the “tragedy” in the first place.
While each of the three media frames construct a different understanding of workplace injuries, taken together these frames create a meta-frame that guides readers’ understanding of workplace injuries. The four elements to the meta-frame are that injuries and fatalities are (1) isolated events (not part of a pattern) that (2) happen to “others” (generally blue-collar men) for which (3) no one is responsible (except maybe the worker), and thus (4) we ought not be concerned about them.
These conclusions are not true. Injuries are not one-off events. As we saw at the beginning of this talk, they are commonplace. A recent study Jason Foster and I completed found that 1 in 5 workers in Alberta had some kind of workplace injury in 2016 (~400,000 major or minor injuries). And 1 in 11 workers had a serious injury—where they need time off or modified duties.
Injuries don’t just happen to blue-collar men, although they do. They happen to all kinds of workers. We just tend to ignore injuries that happen to white-collar workers and to women because they are less likely to be newsworthy event. Think about the kinds of injuries that that serving staff get—burns, lacerations, chemical reactions, stress, and for women sexual harassment and problems caused by inappropriate footwear. That kind of injury almost never makes the paper.
Injuries all have causes. Typically, an employer structured work in a way that resulted in the injury. These kinds of root causes are often difficult to see. For example, Kelly may have slipped on the wet floor but the real question is why was the floor wet? Did Kelly’s boss lay off the person who maintained the hoses or cheap out on the fittings? That’s the root cause of the injury.
The only time we talk about the cause of injury is to blame the worker, often implicitly: For example, consider this article.
Basically, the newspaper parrots the employer’s assertion that, while the cause of a worker’s death is technically unknown, neither equipment nor training was at fault. The implication here is clearly that the worker made a mistake.
Finally, the meta-framing says we should probably not be worried about work-related injuries—the government has injuries under control. Even though that another injury happened is clear evidence that the state doesn't havinjuries under control.
So what factors explain these patterns of reporting? Well, we asked some reporters and editors just that.
What factors contribute to reporting of work-related injury?
Reporters use frames for four main reasons: three practical and one political.
The first reason is accessibility. A frame helps readers understand the story by presenting it in a familiar way. “A worker died on 50th street today and OHS is investigating” is a well known frame and we hardly have to read the rest of the story to get the gist. It saves the reporter and the newspaper time and space that a more complicated or nuanced report would require. That contributes to the template-style of reporting.
The second is limited knowledge. There are, essentially, no labour reporters left in Canada (the Toronto Star would be the sole exception). Generalists have little knowledge of labour issues and few contacts. So they are highly reliant upon government press releases for notification and information. The government tends to issues releases only when there is a fatality or an injury that dramatic (e.g., explosion or fire). That contributes to the distortion we see in what kinds of injuries are reported at all as well as who the reporters rely upon.
The third is time pressure. Reporters are pushed to get an initial story up online as quickly as possible. Rewriting a press release is the easiest way to do that. They say they have no time to investigate or go to the scene and talk to people. Exceptions are when there are multiple fatalities, such as when two saw mills exploded within a month of each other in BC. Instead, reporters crank out a story and then move onto the next. Only when there is some really unusual event—like the mil explosion—do stories persist through multiple news cycles.
These three factors reflect the work intensification in modern news room. Basically, newspapers are driven by the profit imperative (although you wouldn’t know it if you were a Postmedia shareholder!) so the managers are trying to increase productivity to reduce labour costs to increase profitability.
This leads to the fourth factor that explains the use of templates: newspapers are employers and they an have an interest in maintaining the stability of capitalist social formation. Specifically, they have no desire problematizing workplace injury because it reveals the effects of the profound power imbalance between employers and workers and how employers use that to maximize profitability by externalizing cost on workers in the form of unsafe workplaces and injuries.
Now, occasionally you’ll see an expose that reflects reporter’s personal interest in workplace injury—like we did here in Alberta in the summer of 2015. But day to day, there is rarely condemnation or even note of the injury epidemic in the workplace. Rather, newspapers use stories injuries to sell papers but make sure that there is no link between the stories or delving into the what all of these injuries say about employment.
In the end, though, does any of this matter? Does it affect the public discourse about injuries? The answer is yes.
What effects does reporting have on public discourse?
Before we go into this last section about research we’ve just completed, we need to briefly sort out the idea that the world as socially constructed. The basic nub of social construction is that there is an infinite amount of stimulus in the social world. We select some of it to pay attention to and then we interpret what it means. And what we pay attention to and how we interpret it depends upon our beliefs, values, and experiences.
For example, you’re a female student, you go visit you male prof in his office to ask a question, and he closes the door after you enter. How do you interpret that? Is it a bit a creepy? Or is he just trying to protect your privacy? There’s not really an objective interpretation based on the facts I gave you. How you interpret closing the door depends on your beliefs, values, and experiences.
We typically have two main sources of information about the world. We have our experienced reality—the things we and our friends and families experience personally. And we have symbolic reality—things we know about from the press, government, and other institutions (lobby groups, Wikipedia). We use symbolic reality to supplement our experienced reality to create a shared reality.
If the symbolic reality we are exposed to is biased, it may bias our shared and thus skew what stimulus we select and how we interpret it. For example, most of us have no personal experience with the macro-economic effects of increasing the minimum wage. But we hear from employer lobby groups and conservative politicians via the media that wage increases kill jobs. So we probably view increase in the minimum wage with at least some trepidation.
The reality is that historically there has been virtually no relationship between minimum-wage increases and job losses. But you’d never know that unless you looked very carefully at the academic research (which no one ever does). The point being, media distortions of facts have the potential to skew public perception and impact what public policy options are considered appropriate and acceptable.
So lets come back to injury. Newspapers provide demonstrably distorted coverage of workplace injuries. So it is possible they are skewing public perceptions. That said, the widespread nature of injury also means that most of us have at least some personal experience with injuries. So, that may offer a corrective to offset the distortion. We decided to test this with a survey of 2000 Alberta worker sin 2016. We asked them four questions about the level of injury in Alberta. We then compared their results with government statistics and we compared the answers of those who were injured with the answers of those who weren’t.
We started by asking respondents to estimate the number of serious injuries in Alberta—disabling injuries in WCB terminology. We asked this because most work-related injuries are not reported by newspapers. What we found was that 97.6% of respondents under-estimated the number of serious injuries. Most vastly under-estimated the number of serious injuries. On average, respondents estimated there were 5,545 serious injuries annually (which is about 11 or 12% of the official number) and almost no one estimated higher than 20,000 injuries.
Personal experience of injury had no significant impact on worker estimates so doesn’t appear to be acting as a corrective.
We also asked respondents to estimate the ratio of fatalities to serious injuries. We asked about this because newspapers tend to over-report fatalities. The official ratio of was 1 fatality for every 384 serious injuries in Alberta. By contrast, newspapers report three fatalities for every serious injury.
What we found was that the vast majority of respondents over-estimated the ratio of fatalities to serious injuries. Respondents’ average estimate was 1 fatality for every 44 serious injuries. This estimate is not as extreme as newspaper coverage but it is still way out of whack with reality. And, again, personal experience of injury had no significant impact on worker estimates.
Our third question asked respondents to select from a list of 9 industry groupings the 3 most injurious industries. We asked this because newspaper coverage centres on industries with relatively low injury rates.
There was significant agreement among respondents (right-hand column) about the most dangerous industries: 91% selected construction, 72% selected mining and petroleum development, and 65% selected agriculture and forestry.
These were the exact same industries that newspapers reported on the most and almost in the same order.
However, the industries with the highest disabling injury rates in Alberta (left-hand column) were entirely different from those selected by respondents and reported on by newspapers. Again, there was no significant difference in the responses between workers who were injured and those who weren’t.
Overall, what these three questions suggest is that workers’ views of injury tended to align with newspaper reports and diverge from the realities of workplace injury. And there was also no evidence that workers’ personal experiences with injury served as any sort of corrective.
The only exception we found to this pattern was in the fourth question, wheh we asked respondents what proportion of all serious injuries were experienced by women. The correct answer in Alberta was 32.7%.
Overall, the mean answer given by respondents was pretty much bang on the money (33.8%) and men and women were about equally accurate in their estimates.
When you disaggregate the respondents’ answers a bit more nuance appears. There was quite a spread in responses. Only about 17% of respondents estimated the correct percentage (+/-5%). Inaccurate estimates were split evenly between estimating too high and too low.
While respondents weren’t particularly good at estimating the correct percentage, they were more accurate than newspaper reports. Workers’ more accurate estimates may simply reflect that Alberta newspaper reports were so extremely skewed towards injuries to men (91.7%) that workers pretty much couldn’t help but be more accurate.
The upshot of this research is that workers tend to view injury in ways that are consistent with distorted media reports. We can’t infer causality from correlation, but given the theory around social constructivism, the results are strongly suggestive of causality.
Conclusions
So what are the take-aways?
First, media reports of occupational injury paint a distorted picture of how gets injured and how. Second, the frames that are used serve to downplay that this issue is not being effectively regulated by the stat. Third, it is likely that these distortions are skewing the public’s view of injury.
More conjecturally, the absence of accurate information is likely undermining the political verve this issue should have. This takes the heat of policy makers to take meaningful action and allows employers to continue to organize work in unsafe ways—something they do because it is generally more profitable to do so than to make the work safe.
