When I was a kid, I loved going to the zoo. Our zoo had loads of great exhibits, but the thing I remember best from these trips comes from arguably the most boring one: the insectarium (sorry, bug lovers!). There, tucked away among the shelves of preserved bugs, was a billboard that looked something like this.

Of course, as a kid, you know the answer is the shark. You’ve seen Jaws, and I mean honestly, just look at those teeth! If given a choice between eliminating shark attacks or mosquito bites, you’d go with the sharks 9 out of 10 times, getting the bloodsuckers once out of revenge for their annoying bites.

But if you looked at the text beside the billboard at the zoo, you might think again. Sharks kill about ten people per year; Mosquitos: several hundred thousand!

Forget Sharks: Focus on Mosquitoes!

Though it feels a little silly, this example raises important questions about other risks we face in society. Are we afraid of the right things, or do our instincts often fail us as they do for the shark? Are our public health efforts targeting the most significant risks? Does the media accurately inform us about the risks we face, or does it worsen the mismatch?

While all are interesting and important questions, answering each is too much for this blog. So, we started by investigating the third: Is there bias in media coverage of mortality risks?

First, we had to choose the types of risks we wanted to analyze. We selected fourteen risks, including chronic diseases (responsible for the majority of deaths in the USA) and other more sensational risks. Then, we gathered monthly death counts for these risks in the US and all newspaper articles from several major US publishers from 1999 to 2020. What did this data reveal?

The media covers sensational risks far more often than chronic illnesses.

We consider media bias at three levels: topics, facts, and frames. Starting with topics, we list the number of articles and deaths for each risk in the table below. The last column is the most interesting, showing the ratio between them and revealing dramatic variations in coverage.

For heart disease, there are 323 deaths per newspaper mention. In contrast, homicide is 100 times more newsworthy, with an article for every three deaths. Most glaringly different is terrorism, which has 36 articles for every death!

Overall, there doesn’t seem to be any connection between a risk’s number of deaths and the media coverage it receives. What about within risks? Does the media cover a risk more in time periods when it is more prevalent? To find out, we plot these trends over time in the figure below.

Some risks show similar trends, like Alzheimer's (top left). Others, like cancer (one row right), are totally disconnected. Looking across all risks, statistical analysis revealed a relationship between changes in deaths and media mentions, but deaths account for only about 2% of the changes in media coverage. No matter how you measure it, there’s a major disparity between the risks that cause most deaths in the US and the ones the media focuses on. Next, we examined bias in facts, focusing on mitigation and prevention.

Chronic illnesses are portrayed with less emotion and with individual, non-policy solutions.

We looked at whether articles emphasize collective solutions or individual actions. Collective solutions tend to discuss policies that will reduce a risk, for example: an article that discusses the Brady bill and new policing policies to reduce homicide or another article that discusses a social group advocating for drug laws to reduce overdoses and other drug-related health risks. In contrast, individual actions are about personal steps people can take to avoid diseases, for example: an article discusses how exercise, diet, and weight loss will head off diabetes or another tells the story of a professional athlete who overcame drug addiction after seeking help and going to rehab.

We analyzed about 100 articles per risk (1,540 total), labeling the prevention strategies mentioned in each article. We found that articles on sensational risks often discussed policy solutions, while those on chronic illnesses focused on individual actions. This variation in fact-selection matters because it corresponds to differences in how we view these risks, and how much effort we put into collectively preventing them with legislation and personal actions.

Finally, let's discuss framing. Using sentiment analysis (from -10 negative to 10 positive), we found that sensational risks were described more negatively and threateningly, while chronic illnesses were discussed more neutrally. This means sensational risks are not only mentioned more often but are also portrayed as more serious and concerning than chronic illnesses.

The media focuses on sensational “sharks.” It’s time to refocus!

When it comes to mortality risks, we have been focused on sharks: risks that seem scary but are far less consequential than first meets the eye. Media organizations may help us improve our health systems (perhaps speeding up the march toward preventative care) by giving mosquitos their due and devoting more attention to the true threats plaguing our society.

A concluding thought: digging into the data for this project was a lot of fun, and you can come up with insights of your own by fiddling around with the Media Bias Detector. Check out how much organizations talk about crime or wars in comparison to healthcare or health and wellness. Where else are we misplacing our focus?