distinguishing native contributors from helpful strangers

List overview All Threads
Download

newer

older

Asking for your input on Community...

Re: [Wiki-research-l] Database of...

Amir E. Aharoni

5 Jun 2019 5 Jun '19

3:42 p.m.

Hi, There is a phenomenon in Wikipedias in smaller languages: There activity level of people who actually know the language of the wiki and make meaningful text contributions is relatively low, and the activity of people from other wikis who make various technical edits that don't require the knowledge of the language is relatively high. I call the latter group "helpful strangers". They can do things such as fixing categories, fixing invalid wiki syntax, editing templates, adding images, etc.—things that don't require knowing the language well, and can be achieved by copying and pasting, by guessing things from interlanguage links, or by writing language-neutral things, such as numbers or filenames. Now, I've written "relatively low" and "relatively high", but these are just my anecdotal impressions. Has anyone thought of a way to quantify this more precisely? -- Amir Elisha Aharoni · אָמִיר אֱלִישָׁע אַהֲרוֹנִי http://aharoni.wordpress.com ‪“We're living in pieces, I want to live in peace.” – T. Moore‬

Show replies by date

Marc Miquel

5 Jun 5 Jun

4:22 p.m.

New subject: distinguishing native contributors from helpful strangers

Hi Amir, This is an interesting idea. I haven't found a way to detect whether an editor is native or not. My approach to multingual editing is through the concept of having a primary-non primary Wikipedias. Your primary Wikipedia is the one where you have made more edits to (and you are a primary editor there). In the rest of Wikipedias where you have at least an edit, you are a non-primary editor. I'm currently creating a database in which for every Wikipedia I have a table with a column specifying whether an editor is primary from this language or non-primary, another one with the primary language, another one with how many other languages they interacted with and a final one with the total number of edits in all languages. An editor behaves quite differently when he is primary or non-primary in terms of social interactions, topical diversity, etc. To me, this is interesting because it allows me to detect when an editor "exports" content (edits content about their local area, usually politics-related content, in other languages). Assessing the impact of these "technical helpful editors" may not be easy as we'd need to examine the characteristics of the edits. However, quantifying the extent of edits made by non-primary editors is doable. Would that help you? Best, Marc Miquel Missatge de Amir E. Aharoni <amir.aharoni(a)mail.huji.ac.il> del dia dc., 5 de juny 2019 a les 10:54:

...

Andy Mabbett

6:23 p.m.

New subject: distinguishing native contributors from helpful strangers

On Wed, 5 Jun 2019 at 09:42, Amir E. Aharoni <amir.aharoni(a)mail.huji.ac.il> wrote:

...

There is a phenomenon in Wikipedias in smaller languages: There activity level of people who actually know the language of the wiki and make meaningful text contributions is relatively low, and the activity of people from other wikis who make various technical edits that don't require the knowledge of the language is relatively high.

...

Now, I've written "relatively low" and "relatively high", but these are just my anecdotal impressions. Has anyone thought of a way to quantify this more precisely?

It won't answer the question fully, but you can narrow down the results by looking at babel templates to see which languages they self-rate as being proficient in, or otherwise, on their home project(s). I try to act as a "helpful stranger" on non-English projects, for instance by adding images and {{Authority control}} templates. This is usually well received, but there are a couple of projects where the former at least is apparently not welcome, and I've recently been blocked (with no warning; my talk page ink is still red), with no talk page or email access, on Lithuanian Wikipedia. In 2015 I was accused of "vandalism" and "trolling" there. Happy to discuss my experiences - good and bad - off-list, if that will help your research. -- Andy Mabbett @pigsonthewing http://pigsonthewing.org.uk

Kiril Simeonovski

9:03 p.m.

New subject: distinguishing native contributors from helpful strangers

Hi all, I think this is an excellent research topic that might give us helpful insights on how Wikipedias can benefit from the support provided by non-speakers. Discerning the namespaces where this support ends and whether it was made by humans or bots may also give highly useful information. My observations so far regarding any support by the so-called "helpful strangers" can be summarised in the following conclusions: * The larger the community size of a Wikipedia, the higher rules-lawyering applied to the "helpful strangers". This means that: ** Very small Wikipedias (less than 25 active contributors) do not have a strict set of rules nor a native-speaking contributors to watch and every kind of support is welcome (mostly in the form of bot-generated articles and automatic translation of templates). ** Small Wikipedias (from 25-100 active contributors) do have some set of rules and some native-speaking contributors but most kinds of support are still welcome. ** Medium-sized Wikipedias (from 100-1,000 active contributors) do have a clear set of rules and a native-speaking community to take care of everything; the room for support is limited to human editing that abides some rules and sometimes community permission is required (mostly comes in the form of categorisation an correction of templates, while bot-generated stuff is mostly done by native speakers with a bot flag required for strangers). ** Large Wikipedias (over 1,000 active contributors) do have rules about things that could have not been imagined and native-speaking community that easily manages the fields where the strangers could help in, making them not attractive for non-native speakers to come in and help. Another dimension could be a research on the block log of the "helpful strangers" that might explain how these contributors are accepted by the communities they are helping to. Best regards, Kiril On Wed, Jun 5, 2019 at 13:24 Andy Mabbett <andy(a)pigsonthewing.org.uk> wrote:

...

On Wed, 5 Jun 2019 at 09:42, Amir E. Aharoni <amir.aharoni(a)mail.huji.ac.il> wrote:

people

from other wikis who make various technical edits that don't require the knowledge of the language is relatively high.

Now, I've written "relatively low" and "relatively high", but these are just my anecdotal impressions. Has anyone thought of a way to quantify

this

more precisely?

Ziko van Dijk

11 Jun 11 Jun

9:14 p.m.

New subject: distinguishing native contributors from helpful strangers

Hello Amir, Interesting, I called this phenomenon "foreigh helpers" nearly 10 years ago: https://de.wikipedia.org/wiki/Benutzer:Ziko/Handbuch-Allgemeines These people do not speak the language of the language version - otherwise they would be simply a 'normal' part of the community. But they help out with skills, maybe especially those that are not existing in the small community. I have considered also a foreign helper who is supporting translation tools or platforms, like GerardM with providing scripts or people who support translatewiki. But that is usually an activity outside the wiki. Kind regards. Ziko Am Mi., 5. Juni 2019 um 16:04 Uhr schrieb Kiril Simeonovski <kiril.simeonovski(a)gmail.com>om>:

...

On Wed, 5 Jun 2019 at 09:42, Amir E. Aharoni <amir.aharoni(a)mail.huji.ac.il> wrote:

people

from other wikis who make various technical edits that don't require the knowledge of the language is relatively high.

Now, I've written "relatively low" and "relatively high", but these are just my anecdotal impressions. Has anyone thought of a way to quantify

this

more precisely?

_______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

Aaron Halfaker

9:38 p.m.

New subject: distinguishing native contributors from helpful strangers

Looks like these kind of methods would be useful: https://www.aaai.org/ocs/index.php/ICWSM/ICWSM16/paper/viewPDFInterstitial/… I've been looking to implement edit type modeling in ORES. You could use something like this to build a profile on each editor by what types of work they generally do on articles. Right now, this has been deprioritized, but I've got some interested from Wikimedia Product ("Audiences"), so we might prioritize it soon. I'd love to pull in a contractor to do the work to bring it to ORES. On Tue, Jun 11, 2019 at 9:15 AM Ziko van Dijk <zvandijk(a)gmail.com> wrote:

...

whether

it was made by humans or bots may also give highly useful information. My observations so far regarding any support by the so-called "helpful strangers" can be summarised in the following conclusions: * The larger the community size of a Wikipedia, the higher

rules-lawyering

applied to the "helpful strangers". This means that: ** Very small Wikipedias (less than 25 active contributors) do not have a strict set of rules nor a native-speaking contributors to watch and every kind of support is welcome (mostly in the form of bot-generated articles and automatic translation of templates). ** Small Wikipedias (from 25-100 active contributors) do have some set of rules and some native-speaking contributors but most kinds of support are still welcome. ** Medium-sized Wikipedias (from 100-1,000 active contributors) do have a clear set of rules and a native-speaking community to take care of everything; the room for support is limited to human editing that abides some rules and sometimes community permission is required (mostly comes

the form of categorisation an correction of templates, while

bot-generated

stuff is mostly done by native speakers with a bot flag required for strangers). ** Large Wikipedias (over 1,000 active contributors) do have rules about things that could have not been imagined and native-speaking community

that

easily manages the fields where the strangers could help in, making them not attractive for non-native speakers to come in and help. Another dimension could be a research on the block log of the "helpful strangers" that might explain how these contributors are accepted by the communities they are helping to. Best regards, Kiril On Wed, Jun 5, 2019 at 13:24 Andy Mabbett <andy(a)pigsonthewing.org.uk>

wrote:

> On Wed, 5 Jun 2019 at 09:42, Amir E. Aharoni > <amir.aharoni(a)mail.huji.ac.il> wrote: > > > There is a phenomenon in Wikipedias in smaller languages: There

activity

> > level of people who actually know the language of the wiki and make > > meaningful text contributions is relatively low, and the activity of > people > > from other wikis who make various technical edits that don't require

the

> > knowledge of the language is relatively high. > > > Now, I've written "relatively low" and "relatively high", but these

are

> > just my anecdotal impressions. Has anyone thought of a way to

quantify

this

more precisely?

_______________________________________________ Wiki-research-l mailing list Wiki-research-l(a)lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

-- Aaron Halfaker Principal Research Scientist Head of the Scoring Platform team Wikimedia Foundation

1819

days inactive

1825

days old

wiki-research-l@lists.wikimedia.org

Manage subscription

5 comments

6 participants

tags (0)

participants (6)

Aaron Halfaker
Amir E. Aharoni
Andy Mabbett
Kiril Simeonovski
Marc Miquel
Ziko van Dijk