How Wikipedia and Plane Crashes Prove Collective Memory

Airline crashes are traumatic, violent, and dominate the news cycle from the moment a plane veers off radar. It’s no surprise that nearly 1500 Wikipedia articles exist documenting disastrous airline crashes.

But these Wikipedia articles double as evidence of collective society memory, triggering a collective cascade of related memories, according to research published on Wednesday in Science Advances.

“For the first time in the history of human society, we can quantify, measure, and mathematically model our collective memory,” lead author Taha Yasseri tells Inverse.

Planes crashes and Wikipedia offer a perfect combination of factors for studying collective memory. When a plane crashes, people search for it online — but simultaneously, the volume of searches for other plane crashes from the past is even larger.

And while individual memory is studied all the time, research on collective memory is sparse; historically there hasn’t been a foolproof way to collect data. But over the last decade, online activity has become inextricable from our daily lives. Online repositories like Wikipedia are so fertile and active that researchers can use them to systematically analyze our behavior and memory as a society.

Yasseri and his colleagues studied this phenomena with the Germanwings Flight 9525 tragedy two years ago, when the flight’s pilot slammed into a mountain, taking 149 lives with him. The team looked at page statistics for searches on other crashes and topics associate with them. They began to identify how people not only search for a crash after it makes the news, but navigate from one disaster to another.

They modeled the path of searches that begins with “source” articles — a crash that was recent at the time — and ends up at “targets,” or crashes that had taken place farther in the past. To their surprise, they found that target articles garnered 142 percent more page views than the source ones. The Germanwings plane went down in March 2015, yet searches immediately increased for an unrelated plane that had crashed in November 2001 — despite there being no hyperlink from the former to the latter.

Factors like size and date of crash affected the snowballing of searches, though location of airline company did not. Plane crashes make for an unfortunately apt focus with this kind of study because even though the fear they inspire is wildly out of proportion with their actual likelihood (they remain pretty much the safest form of travel on the planet), there are nevertheless a lot of them to search for, and human morbid curiosity is a constant.

Yasseri says one of the main critiques against social media and online technologies has been that they make our attention span stunningly short. But he points out that this provides us with the opportunity to go back in time and educate ourselves on topics we might have otherwise forgotten about: While researchers had previously found that online searches for a crashed plane typically dip after a week, this study showed that collective memory — at least through Wikipedia searches — lasts about 45 years.

“Our short memory has become shorter, but our long memory has become longer,” Yasseri says. “This is not a contradiction — just the new mechanism by which we remember [events] now.”

Abstract

Recently developed information communication technologies, particularly the Internet, have affected how we, both as individuals and as a society, create, store, and recall information. The Internet also provides us with a great opportunity to study memory using transactional large-scale data in a quantitative framework similar to the practice in natural sciences. We make use of online data by analyzing viewership statistics of Wikipedia articles on aircraft crashes. We study the relation between recent events and past events and particularly focus on understanding memory-triggering patterns. We devise a quantitative model that explains the flow of viewership from a current event to past events based on similarity in time, geography, topic, and the hyperlink structure of Wikipedia articles. We show that, on average, the secondary flow of attention to past events generated by these remembering processes is larger than the primary attention flow to the current event. We report these previously unknown cascading effects.