Wisdom of the Crowd?
Quality Control on Wikipedia Fails the Test
Everyone loves Wikipedia and most of us can’t assume it away from our everyday life. But is Wikipedia really a reliable source of information? Can it even be used as a source for academic work? Advocates of the online encyclopedia speak enthusiastically about the potential to correct any mistakes thanks to the large number of users. The “wisdom of the crowd” would correct false information immediately. The Wiki community, they claim, operates an efficient system of self-control, which erases all mistakes and intervenes rigorously in every act of vandalism. Contrary to the traditional print encyclopedias, quality control is thus delegated from a panel of editors to the users. That way a new form of expertise has been emerging that is not embodied in a person but rather constituted by a process.
A simple experiment
But does this really work? In the fall semester 2013, I wanted to put it to the test with the help of my history students at the University of Zurich. The bachelor candidates were given the task to plant some false information in the Wikipedia entries on the revolutions about which they were going to write their semester theses. The aim of the vandalizing experiment was to test the highly praised control mechanism and to find out how long it takes the Wiki community to correct the mistakes. To come straight to the point: Wikipedia does not get full marks. The community was rarely wiki (Hawaiian for “quick”) at eliminating the misinformation.
One of the few exceptions was the entry on Rivoluzione cubana (Cuban Revolution) in the Italian Wikipedia, where we claimed that the fall of the dictator Fulgencio Batista in early 1959 had brought Raul Castro to power. Within only one minute, his brother Fidel was reinstituted as the legitimate leader of the Cuban Revolution. Contrary to the Italian version, or the English one for that matter, any change to an entry in the German Wikipedia has to be approved by an administrator, before it is visible for the users. For example, a grave hunger crisis in the southeast of Iran, which had allegedly supported the political rise of one Ayatollah Khomeini, was deleted from the entry Islamische Revolution (Islamic Revolution) within nine minutes. The rationale, however, was not entirely convincing: the hunger crisis — a complete fabrication of my students — had not been decisive for the course of the revolution according to the examining administrator.
From severe sanction to negligence
One student had quite an unpleasant Wikipedia experience: He was banned as author for “unreasonable editing”, only one minute after he had made the CIA responsible for the murder of Egyptian President Anwar as-Sadat, in fact perpetrated by Muslim extremists in 1981, in the German entry Geschichte der Republik Ägypten (History of the Republic of Egypt). The metamorphosis of the anti-fascist protest singer Zeca Afonso, whose song „Grandola, Vila Morena“ had acted as a torch for the Portuguese Carnation Revolution, into the anti-fascist punk singer Zeca Afonso was reversed within a little bit more than a day and thus in a passable period as well. All the other false information, absurd inventions, and historical fairy tales that we had planted in the free encyclopedia were still uncorrected on Wikipedia at the end of the semester, a full six weeks after the beginning of the experiment. In the German entries, the misinformation had even been examined and adjudged fit for Wiki by an administrator.
The putsch of the Marxist Derg against the Ethiopian Emperor Haile Selassie, which launched the Ethiopian Civil War, now took place on January 16, 1974, according to the corresponding English Wikipedia entry. (Traditional historiography dates the coup d’état on September 12, 1974.) And in the German entry Pariser Kommune (Paris Commune) readers could learn that, during the days of the Paris Commune in spring 1871, the guillotine on Place Voltaire was destroyed by a sword in a symbolic act. Mel Brooks would certainly take pleasure in our imaginative student Michael, yet the historical reality is a little bit more prosaic: the guillotine was burned.
Surprising transition in Chile
Astonishing things could be read for several weeks about Chile’s transition to democracy on the German entry Geschichte Chiles (History of Chile). History books have it that about 56 percent of Chileans voted against a further term for President Augusto Pinochet in a plebiscite in October 1988 and that the military junta accepted the verdict of the people teeth-gnashingly. Moviegoers could recently follow the events around the referendum in the movie No starring Gael Garcia Bernal. On Wikipedia, however, the story has read somewhat differently: „On 5 October 1988 a majority of 78.39% voted for another eight year term for Pinochet in a referendum mandated by the Chilean constitution. Only when it was recognized that the intimidation of officials had influenced the results of the popular vote, Pinochet’s recall from office was announced on 17 October 1988. By then it had turned out that 65.99% had in fact voted against Pinochet’s reelection in the first referendum. In a new plebiscite, a majority of now 67.85% voted against a new term of office for Pinochet.” Two of my students had brazenly changed the outcome of the referendum, created a second popular vote, and thereby rewrote Chile’s recent history.
Grade: F
My favorite manipulation is admittedly caused by Pau and Ladina. In the German entry Mexikanische Revolution (Mexican Revolution), they flagrantly relabeled the policy of “pan o palo” (carrot-and-stick policy) pursued by the regime of Mexican President Porfirio Diaz as „pau o ladina“. To be sure, the sample of our experiment is way too small to make a scientific assertion on the effectivity of quality control by the users. Nonetheless: There are reasonable doubts regarding the crowd’s potential to correct mistakes and the suitability of Wikipedia as a source of information — at least in an academic context.
