Wikipedia Tourist Agency
September 18, 2020 8:53 AM Subscribe
An experiment by economists at the Collegio Carlo Alberto in Turin, Italy, and ZEW in Mannheim, Germany found that adding just two paragraphs of text and a single photo to a city's article increased the number of nights spent there by about 9% during the tourist season.
"The encyclopedia anyone* can edit!"
*We regret to inform you that a single power-user has reverted all of the edits performed by an entire team of researchers.
It's just Milkshake 'Pedia, every time.
posted by explosion at 9:21 AM on September 18, 2020 [9 favorites]
*We regret to inform you that a single power-user has reverted all of the edits performed by an entire team of researchers.
It's just Milkshake 'Pedia, every time.
posted by explosion at 9:21 AM on September 18, 2020 [9 favorites]
Has anyone read the actual paper the article is about? I don't see how they could possibly control the variables for this "experiment" to be able to confidently make this assertion.
posted by star gentle uterus at 10:14 AM on September 18, 2020
posted by star gentle uterus at 10:14 AM on September 18, 2020
This looks like the paper. The clever thing that they did is that they only improved the entry in one of the languages, so they could see an increase in e.g. tourists from Germany over Italy and France when they edited that page.
It seems robust enough.
posted by ambrosen at 10:24 AM on September 18, 2020 [10 favorites]
It seems robust enough.
posted by ambrosen at 10:24 AM on September 18, 2020 [10 favorites]
Full text (at least of a version from 2019).
"a well-targeted two-paragraph improvement.. may lead to a 9% increase in tourists’ overnight visits. The median monthly number of hotel nights spent by tourists from the three effectively treated countries to the cities in the control group was about 3,000 (during the six months from May to October). This implies an increase of about 270 nights per month. Even if there were no tourists in the remaining six months, this implies about 1,600 additional hotel nights per year."
Also, "In 2015, international tourists spent 270 million nights in Spain."
So this might be the equivalent of moving visits from Springfield to Shelbyville. I don't think it's going to shift visits from Madrid to Seville.
posted by Mr.Know-it-some at 12:05 PM on September 18, 2020
"a well-targeted two-paragraph improvement.. may lead to a 9% increase in tourists’ overnight visits. The median monthly number of hotel nights spent by tourists from the three effectively treated countries to the cities in the control group was about 3,000 (during the six months from May to October). This implies an increase of about 270 nights per month. Even if there were no tourists in the remaining six months, this implies about 1,600 additional hotel nights per year."
Also, "In 2015, international tourists spent 270 million nights in Spain."
So this might be the equivalent of moving visits from Springfield to Shelbyville. I don't think it's going to shift visits from Madrid to Seville.
posted by Mr.Know-it-some at 12:05 PM on September 18, 2020
I feel like this would at least be true for me, as a sucker for a good city Wikipedia page and picture.
posted by Corduroy at 12:37 PM on September 18, 2020 [1 favorite]
posted by Corduroy at 12:37 PM on September 18, 2020 [1 favorite]
*We regret to inform you that a single power-user has reverted all of the edits performed by an entire team of researchers.
I'd like to push back on this. While it is true that there are major problems within wikipedia of achieving the dream of democratizing information, I do not agree that this problem should prevent people from editing.
I was very frustrated that the page for my current city kind of... stopped talking about the Native Americans who lived here before white settlement. Over the course of a year, I kept my eyes open for info and finally put together a clear picture: "In 1821, the Council of Three Fires signed the first treaty of Chicago, ceding to the United States all lands in Michigan Territory south of the Grand River, with the exception of several small reservations."
I did the research and re-wrote the section to include this new information. It's still there.
I'm sure that if someone with a beef wanted to fight about it, they would. But that's true of everything having to do with how we talk about the places we live.
posted by rebent at 1:10 PM on September 18, 2020 [4 favorites]
I'd like to push back on this. While it is true that there are major problems within wikipedia of achieving the dream of democratizing information, I do not agree that this problem should prevent people from editing.
I was very frustrated that the page for my current city kind of... stopped talking about the Native Americans who lived here before white settlement. Over the course of a year, I kept my eyes open for info and finally put together a clear picture: "In 1821, the Council of Three Fires signed the first treaty of Chicago, ceding to the United States all lands in Michigan Territory south of the Grand River, with the exception of several small reservations."
I did the research and re-wrote the section to include this new information. It's still there.
I'm sure that if someone with a beef wanted to fight about it, they would. But that's true of everything having to do with how we talk about the places we live.
posted by rebent at 1:10 PM on September 18, 2020 [4 favorites]
Yeah; I'm a sucker for a clever experimental design, and this is one. The Spanish authorities track hotel nights by city by country of the tourist. Four of the countries they track are France, Germany, Italy and the Netherlands; each of those countries has a single primary language. The authors found 60 cities in Spain that had robust tourist data and Wikipedia articles shorter than 24K characters (for reference: 2K characters are described as a paragraph, this modest article about Torrent is around 6K as of my typing this.) This eliminated major destinations (the German article about Barcelona is 53K characters) and places that were too small for any tourist activity.
They then sorted the cities into deciles based on tourist popularity, and then they randomly assigned the six cities in each decile group to have one of the six possible "treatments", where they improved articles in two of the languages and not the other two (DE/NL, DE/FR, DE/IT, NL/FR, NL/IT, FR/IT). The premise is that if (for example) the German and French articles are improved for a city but not the Italian or Dutch, they can compare the change in German and French tourists vs. Italian and Dutch tourists. (They also did some analyses comparing British and Spanish tourists to these cities, who are also unaffected by their changes to Wikipedia.) In this way, every city is treated and untreated at the same time - just depending on which language you speak.
One thing that was particularly convincing to me is that their improvements mattered a lot more if the original article was short -- going from a short article (~3K chars or less) to a more fleshed out one had a 20-30% increase in tourist nights, while going from a modest one to a longer one had a negligible effect. (Exactly the Springfield vs Shelbyville effect; I could see someone driving from Barcelona to Valencia trying to pick one of a couple of places to spend the night and the Wikipedia article pushing them one way or the other.)
Some of this is probably related to Google -- they noted that their improved articles finished higher in Google search rankings. When I search places, the picture that Google shows is usually the one from the Wikipedia article; presumably adding a picture to the Wikipedia article in a given language might add a picture to Google's search results. (It would have been interesting to see the results from a "treatment" where they improved the article by adding text and an image, but not oriented to tourists -- add information about industry, politics and local sports instead of culture, attractions and climate, and add an image of the city's crest instead of a photo of the city.)
Fundamentally, the main takeaway is that having basic information that tourists are interested in will increase the likelihood they will stop by, which shouldn't be news to anyone who's ever seen a brochure, visitor's guide or the like.
posted by Superilla at 2:42 PM on September 18, 2020 [11 favorites]
They then sorted the cities into deciles based on tourist popularity, and then they randomly assigned the six cities in each decile group to have one of the six possible "treatments", where they improved articles in two of the languages and not the other two (DE/NL, DE/FR, DE/IT, NL/FR, NL/IT, FR/IT). The premise is that if (for example) the German and French articles are improved for a city but not the Italian or Dutch, they can compare the change in German and French tourists vs. Italian and Dutch tourists. (They also did some analyses comparing British and Spanish tourists to these cities, who are also unaffected by their changes to Wikipedia.) In this way, every city is treated and untreated at the same time - just depending on which language you speak.
One thing that was particularly convincing to me is that their improvements mattered a lot more if the original article was short -- going from a short article (~3K chars or less) to a more fleshed out one had a 20-30% increase in tourist nights, while going from a modest one to a longer one had a negligible effect. (Exactly the Springfield vs Shelbyville effect; I could see someone driving from Barcelona to Valencia trying to pick one of a couple of places to spend the night and the Wikipedia article pushing them one way or the other.)
Some of this is probably related to Google -- they noted that their improved articles finished higher in Google search rankings. When I search places, the picture that Google shows is usually the one from the Wikipedia article; presumably adding a picture to the Wikipedia article in a given language might add a picture to Google's search results. (It would have been interesting to see the results from a "treatment" where they improved the article by adding text and an image, but not oriented to tourists -- add information about industry, politics and local sports instead of culture, attractions and climate, and add an image of the city's crest instead of a photo of the city.)
Fundamentally, the main takeaway is that having basic information that tourists are interested in will increase the likelihood they will stop by, which shouldn't be news to anyone who's ever seen a brochure, visitor's guide or the like.
posted by Superilla at 2:42 PM on September 18, 2020 [11 favorites]
>the page for my current city kind of... stopped talking about the Native Americans who lived here before white settlement
Don't remember which article it is any more, but I found one about a major midwestern sub-region ... once peopled by a number of tribes ... which started it's history in the early 1900s, never mentioning them.
To be fair, it can be complicated. Wikipedia has lots of holes like that, waiting for a subject-expert to come along and fill in the blank. Six million articles, a quarter of them finished ... most likely the ones about politics and celebrities. Stay away from those and the half-life of your contribution goes up a factor of 10.
posted by Twang at 5:21 PM on September 18, 2020
Don't remember which article it is any more, but I found one about a major midwestern sub-region ... once peopled by a number of tribes ... which started it's history in the early 1900s, never mentioning them.
To be fair, it can be complicated. Wikipedia has lots of holes like that, waiting for a subject-expert to come along and fill in the blank. Six million articles, a quarter of them finished ... most likely the ones about politics and celebrities. Stay away from those and the half-life of your contribution goes up a factor of 10.
posted by Twang at 5:21 PM on September 18, 2020
« Older My standout heroes back then were Ray Anh and Quan... | Why David Quammen Is Not Surprised Newer »
This thread has been archived and is closed to new comments
posted by mercredi at 9:00 AM on September 18, 2020 [5 favorites]