When a plane crashes or goes missing, we become obsessed with knowing exactly what happened. Such tragedies, though statistically rare, capture our attention because it feels inexplicable to lose so many souls in an instant.

It turns out that our curiosity doesn't end there. New research published Wednesday in Science Advancesshows that plane crashes trigger collective memories of similar past events. That might sound obvious, but the researchers used a clever strategy to understand how this dynamic works. In the process, they revealed new insights into collective memory.

Traditionally, scientists who wanted to research how the public recalls past events spent a lot of time and money conducting interviews and surveys. They looked for patterns in how people remembered the past and built a narrative around it. The rise of the internet, however, has yielded vast amounts of data on human behavior, including what people search for and read when a major event becomes all we can talk about.

For this study, the researchers turned to Wikipedia, which makes its page view statistics publicly available, and identified articles on high-profile plane crashes in the recent and distant past. The team, made up of a sociologist, computer engineer, and two physicists, found that when a plane was lost, interest in a similar incident that took place years or decades before also spiked.

"The most important factor was the original importance of [the] past event itself," says Taha Yasseri, the study's team leader and faculty fellow at the Alan Turing Institute. "That suggests some of the events are naturally more memorable than others."

When a Germanwings flight crashed in the French Alps in 2015, for example, people sought more information on Wikipedia about the crash of an American Airlines flight outside of New York City in November 2001.

Yasseri thinks the interest in the latter flight may be related to the pilot's role in each incident. The American Airlines plane went down due to pilot error; the Germanwings pilot is believed to have intentionally crashed his aircraft, though that was not immediately clear at the time. The increased interest in the 2001 incident lasted for several days, and the researchers confirmed that no hyperlink connected the two Wikipedia articles during that time period.

The researchers collected page view data for Wikipedia articles on minor and major aircraft incidents that occurred between 2008 and 2016. Anything that occurred during that time period was considered a recent event while incidents prior to 2008 were categorized as past events.

To minimize background variability in their model, the team narrowed the list of recent crashes down to 11 high-profile instances, including the Air France flight that plunged into the Atlantic Ocean in 2009, the Metrojet flight destroyed mid-air by a bomb in 2015, and the Malaysia Airlines flight that disappeared on its way to Beijing in 2014. Their statistical analysis evaluated increased page views for articles on past events a week after a recent major crash or disappearance.

Whether or not a new incident triggered memories of something that happened in the past largely depended on the importance of the original event. Many accidents, for example, led to a surge in views of Wikipedia articles on the planes that crashed in the Sept. 11 terrorist attacks.

Unsurprisingly, crashes with many deaths tend to be most remembered. Timing is important as well. Events are more frequently associated if less time has elapsed between them.

Another relevant factor was whether a hyperlink on one Wiikipedia article directed readers to a past event. But even when the team removed such instances from its analysis, a similar pattern of remembering emerged. The team also found that memory of aircraft incidents, both deadly and non-fatal, lasted for about 45 years.

"As far as I know, this is the first attempt to provide a mathematical model to predict the flow of attention towards past traumatic events, and paves the way to a new approach to the study of collective memories," Michela Ferron, a researcher at Fondazione Bruno Kessler in Italy who has studied the role of Wikipedia in remembering the past, said in an email. Ferron was not involved in the new study.

While promising, Ferron believes this technique is just one aspect of understanding how collective memory works. Researchers should also explore patterns in how people reflect on and discuss events online, which leads to a "continuous (re)shaping" of memory, she said.

In the study, Yasseri and his co-authors could not account for how media exposure influenced reader behavior. If, for example, a reader heard about the Germanwings flight along with additional reporting on major aviation disasters, it could prompt them to look up information about past crashes.

It's possible, Yasseri says, that people navigated their way to articles about older incidents by exploring Wikipedia categories featured on those webpages, such as related links on airline accidents involving commercial aircraft or more than 50 deaths.

"This exploratory behavior is very much fitting to the idea of an online encyclopedia that gives you information about everything and allows you to be very creative about what to read," Yasseri says.

The model, he adds, could be adapted to look at how the public remembers terrorist attacks or mass shootings in the wake of similar current events, but he doesn't know if it would work for analyzing memories around positive viral events (think BBC Dad or Batkid).

Negativity, Yasseri says, helps news spread faster and farther, potentially making the incident something people talk about and remember for longer periods of time. In other words, "bad news" travels faster, farther and lasts longer.

So the next time you find yourself reading about the past on Wikipedia because of something that just happened, know that what may seem like a few simple clicks is actually a complex series of choices that triggers old memories.