Issues

Media framing is a significant factor in how people and communities interpret, understand, and respond to information. There are two major types of framing:

1. episodic – incidents are reported as isolated events
2. thematic – events are compared, and underlying factors are explored

Articles should include a number of thematic phrases to help contextualize the incident and provide possible remediations:

• Road design
  • Speed limit
  • Number of lanes
  • Lane width
  • Presence of sidewalks, crosswalks, bike lanes
  • Lighting conditions
  • Speedbumps
  • Turn radius
  • Road signage
  • etc.
• Statistics & trends
  • Recent crashes locally
  • Statewide and national trends
• Quotes
  • Local residents
  • Elected officials
  • Transportation safety experts
• Safety initiatives
  • Vision Zero
  • Complete Streets
  • DWI checkpoints
  • High-visibility enforcement

One study found that only about 1 in 20 articles about car crashes employed thematic elements, such as road design or crash statistics. These thematic elements are vital to remind readers that car crashes aren't random events, but natural byproducts of our built environment and the choices people make. Furthermore, readers are more likely to recognize and support preventative measures and solutions when presented with thematic articles.

Counterfactuals are true statements which tend to imply that outcomes would be changed if only people had acted differently. They place a greater burden on the individuals involved, rather than the systemic nature of the underlying situation. Research demonstrates that the presence of counterfactual statements shifts blame toward the victim.

Because car crashes are greatly influenced by systemic factors, such as road design, societal norms and expectations, planning, enforcement, etc., reducing and eliminating counterfactual statements is critical to getting past victim-blaming to achieve comprehensive safety.

Counterfactual statements are a serious anti-pattern because they:

1. obscure the systemic nature of crashes,
2. shift blame onto victims, and
3. absolve actors of irresponsible behavior

Our system is designed to detect a number of counterfactual statements in the following categories:

Counterfactuals typically correspond directly with the absence of thematic elements. The good news is that it's often easy to improve news reporting by observing these relationships:

Status-quo reporting How an incident might be reported today	→	Critical analysis Let's spend a minute to examine this	→	Better reporting Bridging the gap to deliver a higher quality article
The bicyclist was killed when he was hit by an SUV trying to pass him. He wasn't wearing a helmet.		The last sentence is likely to evoke feelings of blame: he should have been wearing a helmet! But wearing a helmet wouldn't have changed the outcome of him being struck. This article makes only factual statements, but there are many things not being said: Is there a safe, protected bike lane provided so bicyclists don't need to share the road with 4,000 lb machines? Is this a popular biking area? If so, are public officials aware of bicyclist safety issues? Was the driver paying attention? Did the driver provide enough buffer space when passing?		The bicyclist was killed when he was hit by an SUV trying to pass him. He wasn't wearing a helmet. The victim was riding along the side of the road which does not have a protected bike lane. While police are still investigating if the driver was focused and provided enough space while passing, nearby residents are sounding the alarm to unsafe road conditions. "So many tourists want to bike along this scenic waterfront route, but the local council for years has been dragging their feet on a bike lane!" says a local. The speed limit is 50 mph, fast enough to make death almost certain should a collision take place.
A pedestrian was left seriously injured after attempting to cross the street. The pedestrian did not use a crosswalk.		The last sentence is likely to evoke feelings of blame: this is what happens when you don't use a crosswalk! The author is certainly sticking to facts, but we should also ask: Is the nearest crosswalk within reasonable walking distance? What is the speed limit at the scene of the incident? Does it provide drivers with enough time to react to pedestrians entering the street to cross? Have there been similar pedestrian-vehicle conflicts in the past? If so, has it led to change? What is the land-use in this area? Does it have schools, restaurants, shops, etc. that people frequently access on foot or bike?		A pedestrian was left seriously injured after attempting to cross the street. The pedestrian did not use a crosswalk. Satellite imagery suggests the nearest crosswalk was at least a quarter mile away – a 5 minute walk – in each direction. A number of pedestrian-vehicle crashes have taken place over the past year, calling into question the safety of a street home to many outdoor restaurants and boutique shops. We spoke with a local restaurant owner: "There's a bus stop across the street, but riders have to run across the street to get to our shop. I told the local bus authority months ago they need to move the bus stop to this side of the street."

As demonstrated, counterfactual statements obscure the underlying causes of crashes. Counterfactuals tend to preclude discussion about systemic factors and instead shift blame on individuals, including the victims. Statements insinuate a different outcome was possible if only individuals acted differently, and articles with counterfactual statements virtually never go into depth to discuss remediation.

accident

1a. an unforeseen and unplanned event or circumstance

2a. an unfortunate event resulting especially from carelessness or ignorance

Merrian-Webster

The tool marks usages of the word "accident" as problematic because the word accident implies a sense of faultlessness and unavoidability. Over 90% of car crashes are caused by human error, and virtually all causes of these crashes are known and avoidable, mostly by changing behavior and design.

When crashes come down to a handful of causes — most commonly distracted driving, speeding, tailgating, alcohol and drug usage, and following too closely — we cannot conclude the causes are unforeseen, and while carelessness may be a factor, these are not matters of mere chance, fortune, or acts of God.

We subscribe to the mentality that all accidents are crashes, but not all crashes are accidents; therefore, a higher burden of proof is required to use the word "accident". Some news agencies express worry that if they use the word crash, someone might file a libel/defamation suit saying the agency is blaming them for the incident. These fears are unwarranted and unrealistic, but there are other words that are recommended, in order of precedence:

1. crash
2. collision
3. incident
4. accident

There has been notable shift in sentiment toward the use of the word "crash":

• The Associated Press recommends avoiding the word "accident" if negligence is claimed or proven, as of 2016.
• The NYPD no longer uses the word "accident" to describe car crashes, and has renamed its Accident Investigation Squad to the Collision Investigation Squad, as of 2013.
• The British Medical Journal (BMJ) has banned the word "accident", as of 2001.

The tool is context-aware and does not punish uses of the word "accident" if it's part of a direct quotation or a proper noun, such as investigative bodies (e.g. the Boston Accident Reconstruction Unit).

Object-based reference deals with how reporters and readers refer to entities. This gives rise to the dichotomy: object-based language vs. people-based language.

Research shows that when referring to drivers, language is usually object-based instead, giving agency to cars, traffic, etc. On the other hand, vulnerable road users are nearly always referenced in human terms: pedestrian, bicyclist, person, man, woman.

Giving agency to objects instead of people obscures who the actors are, and tends to shift blame toward the victims.

• A bicyclist was hit by an SUV. → A bicyclist was hit by the driver of an SUV.
• The car jumped the curb. → The driver drove over the curb.
• The automobile remained at the scene. → The driver remained at the scene.

Additionally, using object-based language personifies objects in downright weird ways:

• The vehicle fled the scene.
• The automobile co-operated with police investigators.

Agency discovers issues with agentive vs. non-agentive sentences. Research demonstrates that readers place more blame on victims when reading non-agentive sentences, i.e. sentences that don't include the perpetrator or one doing the action.

Sentence	Classification
A pedestrian was struck.	Non-agentive
A driver struck a pedestrian.	Agentive
A pedestrian was struck by a driver.	Agentive

Readers identify agents as having ability to cause harm, which helps share responsibility for a mishap. Ensure that all sentences that discuss crashes include an agent.

Focus identifies who the subject of the sentence is, and hence who is at the center of attention. Readers assign more blame to the focus of the sentence. News articles that cover crashes between vehicles and vulnerable road users place the focus of the sentences on the VRU over 70% of the time.

This difference is illustrated below, with the focus (subject) underlined:

• A pedestrian was struck by an SUV.
• An SUV struck a pedestrian.