Why the world reads Wikipedia: beyond English - Wiki-research-l

19 Jan 2019

Hi all,

As some of you know, we started a line of research back in 2016 to
understand Wikipedia readers better. We published the first taxonomy
of Wikipedia readers and we studied and characterized the reader types
in English Wikipedia [1]. During the past 1+ year, we focused on
learning about the potential differences of Wikipedia readers across
languages based on the taxonomy built in [1]. We've learned a lot, and
today we're sharing what we learnt with you.

Some pointers:
* Publication: https://arxiv.org/abs/1812.00474
* Data: https://figshare.com/articles/Why_the_World_Reads_Wikipedia/7579937/1
* (under continuous improvement) Research page on meta:
https://meta.wikimedia.org/wiki/Research:Characterizing_Wikipedia_Reader_Be…
* Research showcase presentation:
https://www.mediawiki.org/wiki/Wikimedia_Research/Showcase#December_2018
* A series of presentations to WMF teams and community: Look for tasks
under https://phabricator.wikimedia.org/T201699 with title "Present
the results of WtWRW" for link to slides and more info when available.
* We will send out a blog post about it hopefully soon. A blog post
about the intermediate results is at
https://wikimediafoundation.org/2018/03/15/why-the-world-reads-wikipedia/

In a nutshell:
* We ran the taxonomy of Wikipedia readers in 14 languages and
measured the prevalence of Wikipedia use-cases and characterized
Wikipedia readers in these languages.
* While we observe similarities in terms of the prevalence of the use
cases as well as the way we can characterize readers, we can see that
Wikipedia languages lend themselves to different distributions of
readership and characteristics. In many cases, the one-size-fit-all
solutions may simply not work for readers.
* Intrinsic learning remains as the number one motivation for people
to come to Wikipedia in the majority of the languages, followed by
media.
* In-depth reading and the reading of scientific oriented topics is
highly and negatively correlated with the socio-economic status and
Human Development Index of countries the readers in these languages
are coming from. Long articles that may seem just too long for the
bulk of our audience in US, Japan, and the Netherlands is in high
demand in India, Bolivia, Argentina, Panamá, México, …
* ...

This research was not possible without the extensive contributions by
our formal collaborators: Florian Lemmerich (RWTH Aachen University)
and Bob West (EPFL). On the WMF end, I was fortunate to work with
Diego Saez on this project as well as more recently, Isaac Johnson.
And all those in the Reading Web and Legal team who supported us
throughout the process. I also want to underline the amazing work that
the volunteers in the languages in the study did to support us heavily
to learn more about their languages, not only through help with
communications within their communities but also with the translation
task which was not an easy one as they were asked to offer their time
not only to translate but also do in-person meetings with us for us to
make sure the intent of the question is translated the same way across
the languages. Usernames Strainu, Tgr, Amire80, Awossink, Antanana,
Lyzzy, Shangkuanlc, Whym, Kaganer, عباد_ديرانية, Satdeep_Gill, Racso,
Hasive: Thank you!

Next we are going to extend this study to include demographics
information. More information about it coming out in the next few
weeks. (And I will send out a separate email to wikimedia-l about this
topic and future research over the weekend. I need some time to
finalize the message to make the message most useful for that
audience.:)

Best,
Leila

[1] https://arxiv.org/abs/1702.05379

--
Leila Zia
Senior Research Scientist, Lead
Wikimedia Foundation