Gendergap and checking our progress against external databases

classic Classic list List threaded Threaded
14 messages Options
Reply | Threaded
Open this post in threaded view
|

Gendergap and checking our progress against external databases

Jane Darnell
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

koltzenburg
thanks,

how about artists identifying as neither female nor male?

cheers,
Claudia
---------- Original Message -----------
From:Magnus Manske <[hidden email]>
To:Research into Wikimedia content and communities
<[hidden email]>
Sent:Tue, 24 Mar 2015 10:59:18 +0000
Subject:Re: [Wiki-research-l] Gendergap and checking our
progress against external databases

> http://magnusmanske.de/wordpress/?p=278
>
> On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell
> <[hidden email]> wrote:
>
> > Hi all,
> > I am compiling some stats regarding the work done on the
Art & Feminism
> > edit-a-thons for my local chapter and while checking the
state of the wikis
> > regarding female artists I noticed that there are huge
local differences
> > per language wiki regarding "who is notable". One of the
things I love
> > about Mix-and-Match is the way you can easily check the
sitelinks per wiki.
> > You can also download the data with autolist to see
which biographies are
> > popular across different languages. I noticed that in
the case of women
> > this seems to be way different than for men. Women
artists are more likely
> > to be notable in one or two languages only, possibly
because they travel
> > less, making their art known more locally than otherwise
- who knows?
> >
> > In any case, here is something to chew on:
> >
> >
https://commons.wikimedia.org/wiki/File:Women_vs_Men_per_external_db_using_Mix-n-Match.jpg
> >
> > I wish we had more databases from more countries that
only contain artists
> > that we could load into Mix-and-Match!
> >
> > Jane
> > _______________________________________________
> > Wiki-research-l mailing list
> > [hidden email]
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
------- End of Original Message -------


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
Claudia,
Unfortunately there would be no way to differentiate those artists from the Wikidata items that are currently missing the gender property. Plus that would only include modern artists, as everyone who appears in the external databases that were born before 1900 have already been assigned a gender whether it's one they would have accepted or not.
Jane

On Tue, Mar 24, 2015 at 12:08 PM, <[hidden email]> wrote:
thanks,

how about artists identifying as neither female nor male?

cheers,
Claudia
---------- Original Message -----------
From:Magnus Manske <[hidden email]>
To:Research into Wikimedia content and communities
<[hidden email]>
Sent:Tue, 24 Mar 2015 10:59:18 +0000
Subject:Re: [Wiki-research-l] Gendergap and checking our
progress against external databases

> http://magnusmanske.de/wordpress/?p=278
>
> On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell
> <[hidden email]> wrote:
>
> > Hi all,
> > I am compiling some stats regarding the work done on the
Art & Feminism
> > edit-a-thons for my local chapter and while checking the
state of the wikis
> > regarding female artists I noticed that there are huge
local differences
> > per language wiki regarding "who is notable". One of the
things I love
> > about Mix-and-Match is the way you can easily check the
sitelinks per wiki.
> > You can also download the data with autolist to see
which biographies are
> > popular across different languages. I noticed that in
the case of women
> > this seems to be way different than for men. Women
artists are more likely
> > to be notable in one or two languages only, possibly
because they travel
> > less, making their art known more locally than otherwise
- who knows?
> >
> > In any case, here is something to chew on:
> >
> >
https://commons.wikimedia.org/wiki/File:Women_vs_Men_per_external_db_using_Mix-n-Match.jpg
> >
> > I wish we had more databases from more countries that
only contain artists
> > that we could load into Mix-and-Match!
> >
> > Jane
> > _______________________________________________
> > Wiki-research-l mailing list
> > [hidden email]
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
------- End of Original Message -------


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2
For ULAN on Wikidata, 477 out of 36,633 are neither male nor female. Two of those have a non-M/F gender assigned. Statistics on those small numbers would essentially be meaningless.

On Tue, Mar 24, 2015 at 11:18 AM Jane Darnell <[hidden email]> wrote:
Claudia,
Unfortunately there would be no way to differentiate those artists from the Wikidata items that are currently missing the gender property. Plus that would only include modern artists, as everyone who appears in the external databases that were born before 1900 have already been assigned a gender whether it's one they would have accepted or not.
Jane

On Tue, Mar 24, 2015 at 12:08 PM, <[hidden email]> wrote:
thanks,

how about artists identifying as neither female nor male?

cheers,
Claudia
---------- Original Message -----------
From:Magnus Manske <[hidden email]>
To:Research into Wikimedia content and communities
<[hidden email]>
Sent:Tue, 24 Mar 2015 10:59:18 +0000
Subject:Re: [Wiki-research-l] Gendergap and checking our
progress against external databases

> http://magnusmanske.de/wordpress/?p=278
>
> On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell
> <[hidden email]> wrote:
>
> > Hi all,
> > I am compiling some stats regarding the work done on the
Art & Feminism
> > edit-a-thons for my local chapter and while checking the
state of the wikis
> > regarding female artists I noticed that there are huge
local differences
> > per language wiki regarding "who is notable". One of the
things I love
> > about Mix-and-Match is the way you can easily check the
sitelinks per wiki.
> > You can also download the data with autolist to see
which biographies are
> > popular across different languages. I noticed that in
the case of women
> > this seems to be way different than for men. Women
artists are more likely
> > to be notable in one or two languages only, possibly
because they travel
> > less, making their art known more locally than otherwise
- who knows?
> >
> > In any case, here is something to chew on:
> >
> >
https://commons.wikimedia.org/wiki/File:Women_vs_Men_per_external_db_using_Mix-n-Match.jpg
> >
> > I wish we had more databases from more countries that
only contain artists
> > that we could load into Mix-and-Match!
> >
> > Jane
> > _______________________________________________
> > Wiki-research-l mailing list
> > [hidden email]
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
------- End of Original Message -------


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
In reply to this post by Magnus Manske-2
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
Exactly! However, I do think this example woman shows a Wikipedia gender bias towards men in that her husband's article has an image and hers doesn't

On Tue, Mar 24, 2015 at 1:54 PM, Magnus Manske <[hidden email]> wrote:


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2
Google image search for properly licensed images didn't give me an image of her, neither did Commons. Maybe I just overlooked it, but maybe there is no image of her in the article because there is no free image of her? In that case, it would not be bias of Wikipedia, but again bias of the interwebs.

On Tue, Mar 24, 2015 at 1:07 PM Jane Darnell <[hidden email]> wrote:
Exactly! However, I do think this example woman shows a Wikipedia gender bias towards men in that her husband's article has an image and hers doesn't

On Tue, Mar 24, 2015 at 1:54 PM, Magnus Manske <[hidden email]> wrote:


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
Yes you have a point there, and it is this "systemic interweb gender bias" which is interesting. Due to our need for "reliable sources" to solve gendergap issues, it would be nice to be able to expose these "interweb" imbalances with graphs like yours for all databases. What does the ODNB look like? That one is completely matched now. I measured 10% women over all.

On Tue, Mar 24, 2015 at 2:17 PM, Magnus Manske <[hidden email]> wrote:
Google image search for properly licensed images didn't give me an image of her, neither did Commons. Maybe I just overlooked it, but maybe there is no image of her in the article because there is no free image of her? In that case, it would not be bias of Wikipedia, but again bias of the interwebs.

On Tue, Mar 24, 2015 at 1:07 PM Jane Darnell <[hidden email]> wrote:
Exactly! However, I do think this example woman shows a Wikipedia gender bias towards men in that her husband's article has an image and hers doesn't

On Tue, Mar 24, 2015 at 1:54 PM, Magnus Manske <[hidden email]> wrote:


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Magnus Manske-2
Updated blog with ODNB (49,419 men, 5,581 women) plot.

On Tue, Mar 24, 2015 at 1:31 PM Jane Darnell <[hidden email]> wrote:
Yes you have a point there, and it is this "systemic interweb gender bias" which is interesting. Due to our need for "reliable sources" to solve gendergap issues, it would be nice to be able to expose these "interweb" imbalances with graphs like yours for all databases. What does the ODNB look like? That one is completely matched now. I measured 10% women over all.

On Tue, Mar 24, 2015 at 2:17 PM, Magnus Manske <[hidden email]> wrote:
Google image search for properly licensed images didn't give me an image of her, neither did Commons. Maybe I just overlooked it, but maybe there is no image of her in the article because there is no free image of her? In that case, it would not be bias of Wikipedia, but again bias of the interwebs.

On Tue, Mar 24, 2015 at 1:07 PM Jane Darnell <[hidden email]> wrote:
Exactly! However, I do think this example woman shows a Wikipedia gender bias towards men in that her husband's article has an image and hers doesn't

On Tue, Mar 24, 2015 at 1:54 PM, Magnus Manske <[hidden email]> wrote:


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
Reply | Threaded
Open this post in threaded view
|

Re: Gendergap and checking our progress against external databases

Jane Darnell
oooo nice!

On Tue, Mar 24, 2015 at 2:55 PM, Magnus Manske <[hidden email]> wrote:
Updated blog with ODNB (49,419 men, 5,581 women) plot.

On Tue, Mar 24, 2015 at 1:31 PM Jane Darnell <[hidden email]> wrote:
Yes you have a point there, and it is this "systemic interweb gender bias" which is interesting. Due to our need for "reliable sources" to solve gendergap issues, it would be nice to be able to expose these "interweb" imbalances with graphs like yours for all databases. What does the ODNB look like? That one is completely matched now. I measured 10% women over all.

On Tue, Mar 24, 2015 at 2:17 PM, Magnus Manske <[hidden email]> wrote:
Google image search for properly licensed images didn't give me an image of her, neither did Commons. Maybe I just overlooked it, but maybe there is no image of her in the article because there is no free image of her? In that case, it would not be bias of Wikipedia, but again bias of the interwebs.

On Tue, Mar 24, 2015 at 1:07 PM Jane Darnell <[hidden email]> wrote:
Exactly! However, I do think this example woman shows a Wikipedia gender bias towards men in that her husband's article has an image and hers doesn't

On Tue, Mar 24, 2015 at 1:54 PM, Magnus Manske <[hidden email]> wrote:


On Tue, Mar 24, 2015 at 12:46 PM Jane Darnell <[hidden email]> wrote:
Well here you get into a problem with data in the databases. Often the records for men are more fleshed out than for women.

Here are two records in the RKD for male and female based on a US-based husband & wife duo never active in the Netherlands, but who were made popular (with articles on 4 or more language wikis)  by a 2014 movie about them:

The ULAN only has one record

One could argue that there is not enough in any of these records to make a Wikidata entry, let alone a Wikipedia article in any language. So just having the record in the ULAN or the RKDartists will not help any language-pedia. The English Wikipedia includes of course articles for both, though they are not sourced to these databases at all.

Which only shows that RKD/ULAN are not the basis for Wikipedia articles. Your "example woman" has 13 references:
https://en.wikipedia.org/wiki/Margaret_Keane

But, the argument could be made that general coverage on the web is biased towards men, and Wikipedia reflects that trend, if only for a lack of sources.


In comparison, here is a US male artist without any Wikipedia article, born the same year as Mrs. Keane, whose entry in the RKD is well fleshed out, because this person was active in Amsterdam

On Tue, Mar 24, 2015 at 12:35 PM, Magnus Manske <[hidden email]> wrote:
Well, a "local bias" would push the total coverage in a certain language, as evidences by high nl.wp coverage for RKD but not for ULAN. But how would that affect the gender bias? No matter how popular RKDartists are on a Wikipedia, unbiased coverage should always put them on the line in my plots. However, only the almost-complete Wikipedias come near that line.

On Tue, Mar 24, 2015 at 11:27 AM Jane Darnell <[hidden email]> wrote:
Magnus,
That is very cool! I also noticed the same trend, which indeed does at first glance appear to be a male bias in the language wikis. However, drilling down into the actual data it's really not so bad. What you actually see is a strong bias from the ULAN for artists who have been active in the US for whatever reason (artworks in US museum, exhibition, or public art). The RKD has a similar strong bias for artists who have been active in the Northern & Southern Netherlands and adjoining countries (nl, fr + de). We should offset this by looking at other databases, and also looking at the same cross-wiki data for all biographies with occupation=artist.

Jane

On Tue, Mar 24, 2015 at 11:59 AM, Magnus Manske <[hidden email]> wrote:

On Tue, Mar 24, 2015 at 8:48 AM Jane Darnell <[hidden email]> wrote:
Hi all,
I am compiling some stats regarding the work done on the Art & Feminism edit-a-thons for my local chapter and while checking the state of the wikis regarding female artists I noticed that there are huge local differences per language wiki regarding "who is notable". One of the things I love about Mix-and-Match is the way you can easily check the sitelinks per wiki. You can also download the data with autolist to see which biographies are popular across different languages. I noticed that in the case of women this seems to be way different than for men. Women artists are more likely to be notable in one or two languages only, possibly because they travel less, making their art known more locally than otherwise - who knows?

In any case, here is something to chew on:

I wish we had more databases from more countries that only contain artists that we could load into Mix-and-Match!

Jane
_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l


_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l



_______________________________________________
Wiki-research-l mailing list
[hidden email]
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l