Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Jonathan,

I spent some time to carefully read the FAQ about research and the
guidelines linked from there.

As for the progress of the project, I will provide regular updates on Meta.

Best,
Kiril

On Wed, Feb 26, 2020 at 12:56 AM Jonathan Morgan 
wrote:

> Hi Kiril!
>
> Thanks for all the patient and thoughtful clarifications and elaborations
> :)  I left a couple of comments inline, below.
>
> Good luck with your project! As you can tell, we are a curious and
> thoughtful group here on wikiresearch-l. If you have methodological
> questions in the future, please don't hesitate to ask them here.
>
> On Tue, Feb 25, 2020 at 11:43 AM Kiril Simeonovski <
> kiril.simeonov...@gmail.com> wrote:
>
> > Hi Jonathan,
> >
> > You have correctly deduced from the description that there will not be
> any
> > interaction with editors and all the data for the research will be drawn
> > from the publicly available edit histories. The most confusing part that
> > gives impression of intervening is perhaps the "experiment", which
> > conceptually differs in the social sciences from its more common meaning
> in
> > a laboratory environment. That said, this research is not going to
> consume
> > editor's time for surveys nor it is going to convert Wikipedia to a
> > laboratory or ask people to change their behaviour.
> >
> > I came here with the announcement after creating the proposal on Meta and
> > following the guidelines regarding research projects with the goal of
> > getting some useful input from other researchers subscribed to the
> mailing
> > list and learning how to administratively proceed with the proposal on
> Meta
> > (What should be done next on Meta? Will there be an appointed WMF
> > researcher to contact regarding this research?).
> >
>
>
> Putting your research proposal on Meta is best practice for all research
> projects related to Wikimedia. It is not a required step, but it's useful
> for increasing awareness of your project among the broader Wikimedia
> communities (researchers and everyone else).
>
> There is nothing else you need to do at this point, although we appreciate
> it if you would keep your project page up to date as you perform your
> research. When you're done, we always appreciate it if you link to any
> preprints, demos, code repos, slide decks, etc from that page as well.
>
> Your project won't automatically be assigned a WMF contact. The Wikimedia
> Foundation itself does not officially monitor or screen new research
> projects that are published on Meta, or review them for support. However,
> if you believe your research furthers the goals of the Wikimedia Movement,
> you might consider applying for a grant (example
> <
> https://meta.wikimedia.org/wiki/Grants:Project/Misinformation_And_Its_Discontents:_Narrative_Recommendations_on_Wikipedia%27s_Vulnerabilities_and_Resilience
> >).
>
>
> Individual WMF teams (including my team, Research) do occasionally partner
> with external researchers
>  and those partnerships
> can include access to non-public data (under a Non-Disclosure Agreement).
> Some partnerships do involve funding, but this is not common. All
> partnerships are at the discretion of the team manager. In the case of my
> team, that manager is Leila Zia, Head of Research.
>
>  Quick question: when you say "the guidelines regarding research projects"
> above, what document are you referring to? There are a lot of these pages
> in the Research namespace and they are not always up to date,
> unfortunately.
>
>
> > My request for help from research community regarding this research will
> be
> > mostly technical (e.g. smart random sampling of editors, existing tools
> for
> > research purposes, etc.)
> >
>
> We can definitely help you with these questions! You can also post
> questions related to data access and data infrastructure to the analytics
> mailing list .
>
>
> >
> > Best,
> > Kiril
> >
> > On Tue 25. Feb 2020 at 17:06, Jonathan Morgan 
> > wrote:
> >
> > > Taking a quick step back from all the very enthusiastic questioning of
> > the
> > > researcher's motives...
> > >
> > > Kiril,
> > >
> > > Regarding your methods, Your proposal states
> > > <
> > >
> >
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> > > >
> > > that for this study "The editors will be sampled from the pool of
> > > contributors to all language editions over Wikipedia's entire history
> and
> > > will be classified into groups based on their longevity on the
> project."
> > > But it says little more than that.
> > >
> > > When I read this description, it does not sound to me like you will
> > > necessarily be contacting editors for this study, or intervening in any
> > way
> > > into Wikipedia. Stuart and Pine's questions seem to assume that you
> will
> > be
> > > in some way recruiting editors as participants, 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Jonathan Morgan]

Hi Kiril!

Thanks for all the patient and thoughtful clarifications and elaborations
:)  I left a couple of comments inline, below.

Good luck with your project! As you can tell, we are a curious and
thoughtful group here on wikiresearch-l. If you have methodological
questions in the future, please don't hesitate to ask them here.

On Tue, Feb 25, 2020 at 11:43 AM Kiril Simeonovski <
kiril.simeonov...@gmail.com> wrote:

> Hi Jonathan,
>
> You have correctly deduced from the description that there will not be any
> interaction with editors and all the data for the research will be drawn
> from the publicly available edit histories. The most confusing part that
> gives impression of intervening is perhaps the "experiment", which
> conceptually differs in the social sciences from its more common meaning in
> a laboratory environment. That said, this research is not going to consume
> editor's time for surveys nor it is going to convert Wikipedia to a
> laboratory or ask people to change their behaviour.
>
> I came here with the announcement after creating the proposal on Meta and
> following the guidelines regarding research projects with the goal of
> getting some useful input from other researchers subscribed to the mailing
> list and learning how to administratively proceed with the proposal on Meta
> (What should be done next on Meta? Will there be an appointed WMF
> researcher to contact regarding this research?).
>


Putting your research proposal on Meta is best practice for all research
projects related to Wikimedia. It is not a required step, but it's useful
for increasing awareness of your project among the broader Wikimedia
communities (researchers and everyone else).

There is nothing else you need to do at this point, although we appreciate
it if you would keep your project page up to date as you perform your
research. When you're done, we always appreciate it if you link to any
preprints, demos, code repos, slide decks, etc from that page as well.

Your project won't automatically be assigned a WMF contact. The Wikimedia
Foundation itself does not officially monitor or screen new research
projects that are published on Meta, or review them for support. However,
if you believe your research furthers the goals of the Wikimedia Movement,
you might consider applying for a grant (example
).


Individual WMF teams (including my team, Research) do occasionally partner
with external researchers
 and those partnerships
can include access to non-public data (under a Non-Disclosure Agreement).
Some partnerships do involve funding, but this is not common. All
partnerships are at the discretion of the team manager. In the case of my
team, that manager is Leila Zia, Head of Research.

 Quick question: when you say "the guidelines regarding research projects"
above, what document are you referring to? There are a lot of these pages
in the Research namespace and they are not always up to date, unfortunately.


> My request for help from research community regarding this research will be
> mostly technical (e.g. smart random sampling of editors, existing tools for
> research purposes, etc.)
>

We can definitely help you with these questions! You can also post
questions related to data access and data infrastructure to the analytics
mailing list .


>
> Best,
> Kiril
>
> On Tue 25. Feb 2020 at 17:06, Jonathan Morgan 
> wrote:
>
> > Taking a quick step back from all the very enthusiastic questioning of
> the
> > researcher's motives...
> >
> > Kiril,
> >
> > Regarding your methods, Your proposal states
> > <
> >
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> > >
> > that for this study "The editors will be sampled from the pool of
> > contributors to all language editions over Wikipedia's entire history and
> > will be classified into groups based on their longevity on the project."
> > But it says little more than that.
> >
> > When I read this description, it does not sound to me like you will
> > necessarily be contacting editors for this study, or intervening in any
> way
> > into Wikipedia. Stuart and Pine's questions seem to assume that you will
> be
> > in some way recruiting editors as participants, asking them to change
> their
> > behavior, asking them questions, etc.
> >
> > *Will you be performing any of the above activities?* If not, the
> questions
> > asked so far may be beside the point. Anyone is free to perform analysis
> on
> > publicly available and free-licensed data.
> >
> > If you do plan to intervene in Wikipedia in some way, or work with
> editors
> > as research participants or co-researchers, and you would like the
> members
> > of this mailing list to 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Pine,

You can see my reply to Jonathan about the nature of the research and what
does "experiment" refer to in the context of this research.

I think the answer to your question on how this research will benefit the
community is highly dependent on your expectations and the scope of the
research. While there are indicative research projects with the aim to tell
whether the community is ready for introducing a novelty or implementing a
major change, there are also projects that focus on the social aspects of
the collaborative environments and how people behave in different
situations. The latter group of projects seems not to offer direct benefit
for the community but it doesn't mean they are useless and shouldn't be
done at all.

I know it's natural that people with different backgrounds have
difficulties to comprehend the complexity of economic modelling and there
should be an interdisciplinary way to present the findings so that they
could practically reach to a wider group of people. This is something that
will be addressed with this research.

Yet, it's still possible to give you a direction of where the benefit for
the community would come from. Similar researches in the past offered
explanations to many relevant real-world questions such as how much time
should people spend at work, why companies don't produce as much as they
can what people want, why people tend to be more risk-averse towards
unknown risks or how changes in the environment affect the decision-making
by individuals.

Best,
Kiril

On Tue, Feb 25, 2020 at 8:42 PM Kiril Simeonovski <
kiril.simeonov...@gmail.com> wrote:

> Hi Jonathan,
>
> You have correctly deduced from the description that there will not be any
> interaction with editors and all the data for the research will be drawn
> from the publicly available edit histories. The most confusing part that
> gives impression of intervening is perhaps the "experiment", which
> conceptually differs in the social sciences from its more common meaning in
> a laboratory environment. That said, this research is not going to consume
> editor's time for surveys nor it is going to convert Wikipedia to a
> laboratory or ask people to change their behaviour.
>
> I came here with the announcement after creating the proposal on Meta and
> following the guidelines regarding research projects with the goal of
> getting some useful input from other researchers subscribed to the mailing
> list and learning how to administratively proceed with the proposal on Meta
> (What should be done next on Meta? Will there be an appointed WMF
> researcher to contact regarding this research?).
>
> My request for help from research community regarding this research will
> be mostly technical (e.g. smart random sampling of editors, existing tools
> for research purposes, etc.)
>
> Best,
> Kiril
>
> On Tue 25. Feb 2020 at 17:06, Jonathan Morgan 
> wrote:
>
>> Taking a quick step back from all the very enthusiastic questioning of the
>> researcher's motives...
>>
>> Kiril,
>>
>> Regarding your methods, Your proposal states
>> <
>> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
>> >
>> that for this study "The editors will be sampled from the pool of
>> contributors to all language editions over Wikipedia's entire history and
>> will be classified into groups based on their longevity on the project."
>> But it says little more than that.
>>
>> When I read this description, it does not sound to me like you will
>> necessarily be contacting editors for this study, or intervening in any
>> way
>> into Wikipedia. Stuart and Pine's questions seem to assume that you will
>> be
>> in some way recruiting editors as participants, asking them to change
>> their
>> behavior, asking them questions, etc.
>>
>> *Will you be performing any of the above activities?* If not, the
>> questions
>> asked so far may be beside the point. Anyone is free to perform analysis
>> on
>> publicly available and free-licensed data.
>>
>> If you do plan to intervene in Wikipedia in some way, or work with editors
>> as research participants or co-researchers, and you would like the members
>> of this mailing list to provide you with feedback or other support, please
>> describe the support or feedback you would like to receive in more detail.
>>
>> If your study is non-interventionist but you still want feedback, we can
>> provide that too. Perhaps you can be more clear about the kind of feedback
>> you want; that will keep the conversation going in an interesting and
>> productive direction that everyone on the list can benefit from.
>>
>> Finally, we the members of this list (whether volunteers or WMF staff) are
>> not peer reviewers, do not speak for the Wikipedia community, and are not
>> empowered to approve or deny research requests.
>>
>> Best,
>> Jonathan
>>
>> On Mon, Feb 24, 2020 at 11:23 PM Kiril Simeonovski <
>> kiril.simeonov...@gmail.com> wrote:
>>
>> > Hi 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Jonathan,

You have correctly deduced from the description that there will not be any
interaction with editors and all the data for the research will be drawn
from the publicly available edit histories. The most confusing part that
gives impression of intervening is perhaps the "experiment", which
conceptually differs in the social sciences from its more common meaning in
a laboratory environment. That said, this research is not going to consume
editor's time for surveys nor it is going to convert Wikipedia to a
laboratory or ask people to change their behaviour.

I came here with the announcement after creating the proposal on Meta and
following the guidelines regarding research projects with the goal of
getting some useful input from other researchers subscribed to the mailing
list and learning how to administratively proceed with the proposal on Meta
(What should be done next on Meta? Will there be an appointed WMF
researcher to contact regarding this research?).

My request for help from research community regarding this research will be
mostly technical (e.g. smart random sampling of editors, existing tools for
research purposes, etc.)

Best,
Kiril

On Tue 25. Feb 2020 at 17:06, Jonathan Morgan  wrote:

> Taking a quick step back from all the very enthusiastic questioning of the
> researcher's motives...
>
> Kiril,
>
> Regarding your methods, Your proposal states
> <
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> >
> that for this study "The editors will be sampled from the pool of
> contributors to all language editions over Wikipedia's entire history and
> will be classified into groups based on their longevity on the project."
> But it says little more than that.
>
> When I read this description, it does not sound to me like you will
> necessarily be contacting editors for this study, or intervening in any way
> into Wikipedia. Stuart and Pine's questions seem to assume that you will be
> in some way recruiting editors as participants, asking them to change their
> behavior, asking them questions, etc.
>
> *Will you be performing any of the above activities?* If not, the questions
> asked so far may be beside the point. Anyone is free to perform analysis on
> publicly available and free-licensed data.
>
> If you do plan to intervene in Wikipedia in some way, or work with editors
> as research participants or co-researchers, and you would like the members
> of this mailing list to provide you with feedback or other support, please
> describe the support or feedback you would like to receive in more detail.
>
> If your study is non-interventionist but you still want feedback, we can
> provide that too. Perhaps you can be more clear about the kind of feedback
> you want; that will keep the conversation going in an interesting and
> productive direction that everyone on the list can benefit from.
>
> Finally, we the members of this list (whether volunteers or WMF staff) are
> not peer reviewers, do not speak for the Wikipedia community, and are not
> empowered to approve or deny research requests.
>
> Best,
> Jonathan
>
> On Mon, Feb 24, 2020 at 11:23 PM Kiril Simeonovski <
> kiril.simeonov...@gmail.com> wrote:
>
> > Hi Stuart,
> >
> > Thank you for your thoughts so far. I really like how the discussion is
> > progressing.
> >
> > The methodology will, of course, yield other results about editor
> dynamics
> > and growth paths. Paid editing and sock puppetry as systemic risk factors
> > could be included in the model exogenously but it might be possible to
> > endogenise them in any future research. At this stage, the most important
> > thing is to lay the grounds for developing a sensible model that can be
> > later upgraded with new assumptions.
> >
> > As for the editing experience, I've been around since 2008 (this is my
> edit
> > log <
> https://meta.wikimedia.org/wiki/Special:CentralAuth/Kiril_Simeonovski
> > >
> > ).
> >
> > Best,
> > Kiril
> >
> > On Tue, Feb 25, 2020 at 12:37 AM Stuart A. Yeates 
> > wrote:
> >
> > > Hi Kiril
> > >
> > > Let's just say that history has taught us to be risk-averse to
> > > drive-by researchers.
> > >
> > > Can you point us to other research output using this methodology? Do
> > > you (or any of your team) have significant editing experience? Are you
> > > familiar with the firestorm that is paid editing and sock puppetry??
> > >
> > > cheers
> > > stuart
> > >
> > > --
> > > ...let us be heard from red core to black sky
> > >
> > > On Tue, 25 Feb 2020 at 10:43, Kiril Simeonovski
> > >  wrote:
> > > >
> > > > Hi Pine,
> > > >
> > > > The findings from the research will be articulated to draw clear
> > > > conclusions about what causes utility and disutility from
> > participation,
> > > > and how this is perceived by different editors. For instance, it is
> > > natural
> > > > to assume that editors come to contribute by adding content that will
> > > > remain visible, while blocks 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Jonathan Morgan]

Taking a quick step back from all the very enthusiastic questioning of the
researcher's motives...

Kiril,

Regarding your methods, Your proposal states

that for this study "The editors will be sampled from the pool of
contributors to all language editions over Wikipedia's entire history and
will be classified into groups based on their longevity on the project."
But it says little more than that.

When I read this description, it does not sound to me like you will
necessarily be contacting editors for this study, or intervening in any way
into Wikipedia. Stuart and Pine's questions seem to assume that you will be
in some way recruiting editors as participants, asking them to change their
behavior, asking them questions, etc.

*Will you be performing any of the above activities?* If not, the questions
asked so far may be beside the point. Anyone is free to perform analysis on
publicly available and free-licensed data.

If you do plan to intervene in Wikipedia in some way, or work with editors
as research participants or co-researchers, and you would like the members
of this mailing list to provide you with feedback or other support, please
describe the support or feedback you would like to receive in more detail.

If your study is non-interventionist but you still want feedback, we can
provide that too. Perhaps you can be more clear about the kind of feedback
you want; that will keep the conversation going in an interesting and
productive direction that everyone on the list can benefit from.

Finally, we the members of this list (whether volunteers or WMF staff) are
not peer reviewers, do not speak for the Wikipedia community, and are not
empowered to approve or deny research requests.

Best,
Jonathan

On Mon, Feb 24, 2020 at 11:23 PM Kiril Simeonovski <
kiril.simeonov...@gmail.com> wrote:

> Hi Stuart,
>
> Thank you for your thoughts so far. I really like how the discussion is
> progressing.
>
> The methodology will, of course, yield other results about editor dynamics
> and growth paths. Paid editing and sock puppetry as systemic risk factors
> could be included in the model exogenously but it might be possible to
> endogenise them in any future research. At this stage, the most important
> thing is to lay the grounds for developing a sensible model that can be
> later upgraded with new assumptions.
>
> As for the editing experience, I've been around since 2008 (this is my edit
> log  >
> ).
>
> Best,
> Kiril
>
> On Tue, Feb 25, 2020 at 12:37 AM Stuart A. Yeates 
> wrote:
>
> > Hi Kiril
> >
> > Let's just say that history has taught us to be risk-averse to
> > drive-by researchers.
> >
> > Can you point us to other research output using this methodology? Do
> > you (or any of your team) have significant editing experience? Are you
> > familiar with the firestorm that is paid editing and sock puppetry??
> >
> > cheers
> > stuart
> >
> > --
> > ...let us be heard from red core to black sky
> >
> > On Tue, 25 Feb 2020 at 10:43, Kiril Simeonovski
> >  wrote:
> > >
> > > Hi Pine,
> > >
> > > The findings from the research will be articulated to draw clear
> > > conclusions about what causes utility and disutility from
> participation,
> > > and how this is perceived by different editors. For instance, it is
> > natural
> > > to assume that editors come to contribute by adding content that will
> > > remain visible, while blocks and reverted edits are risk factors that
> > drive
> > > them away, although different editors have different levels of risk
> > > aversion. Similarly to any other research, the benefit for the
> community
> > > and individual editors is going to be indirect but yet not
> insignificant
> > to
> > > be accepted in the future process of decision-making (if the research
> > > demonstrates the existence of high level of risk aversion towards
> > > something, then it automatically signals that doing that thing is
> harmful
> > > for the environment).
> > >
> > > I know that it's impossible to predict the extent to which this
> research
> > > would make impact because the body of literature is very poor on
> > > volunteer-driven environments in a dynamic setting but it's definitely
> > > worth to start off something that might attract the attention of
> > > researchers in this direction. At the end, the research is not meant to
> > > carve rules in stone that any single editor should respect but rather
> to
> > > suggest something that individuals and communities might find useful
> (the
> > > means of doing this will definitely not turn Wikipedia into a
> laboratory
> > or
> > > put someone's privacy in danger).
> > >
> > > Best,
> > > Kiril
> > >
> > > On Mon, Feb 24, 2020 at 9:43 PM Pine W  wrote:
> > >
> > > > Hi Kiril,
> > > >
> > > > Thank you for sharing your proposal.
> > > >
> > > > I am 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Stuart,

Thank you for your thoughts so far. I really like how the discussion is
progressing.

The methodology will, of course, yield other results about editor dynamics
and growth paths. Paid editing and sock puppetry as systemic risk factors
could be included in the model exogenously but it might be possible to
endogenise them in any future research. At this stage, the most important
thing is to lay the grounds for developing a sensible model that can be
later upgraded with new assumptions.

As for the editing experience, I've been around since 2008 (this is my edit
log 
).

Best,
Kiril

On Tue, Feb 25, 2020 at 12:37 AM Stuart A. Yeates  wrote:

> Hi Kiril
>
> Let's just say that history has taught us to be risk-averse to
> drive-by researchers.
>
> Can you point us to other research output using this methodology? Do
> you (or any of your team) have significant editing experience? Are you
> familiar with the firestorm that is paid editing and sock puppetry??
>
> cheers
> stuart
>
> --
> ...let us be heard from red core to black sky
>
> On Tue, 25 Feb 2020 at 10:43, Kiril Simeonovski
>  wrote:
> >
> > Hi Pine,
> >
> > The findings from the research will be articulated to draw clear
> > conclusions about what causes utility and disutility from participation,
> > and how this is perceived by different editors. For instance, it is
> natural
> > to assume that editors come to contribute by adding content that will
> > remain visible, while blocks and reverted edits are risk factors that
> drive
> > them away, although different editors have different levels of risk
> > aversion. Similarly to any other research, the benefit for the community
> > and individual editors is going to be indirect but yet not insignificant
> to
> > be accepted in the future process of decision-making (if the research
> > demonstrates the existence of high level of risk aversion towards
> > something, then it automatically signals that doing that thing is harmful
> > for the environment).
> >
> > I know that it's impossible to predict the extent to which this research
> > would make impact because the body of literature is very poor on
> > volunteer-driven environments in a dynamic setting but it's definitely
> > worth to start off something that might attract the attention of
> > researchers in this direction. At the end, the research is not meant to
> > carve rules in stone that any single editor should respect but rather to
> > suggest something that individuals and communities might find useful (the
> > means of doing this will definitely not turn Wikipedia into a laboratory
> or
> > put someone's privacy in danger).
> >
> > Best,
> > Kiril
> >
> > On Mon, Feb 24, 2020 at 9:43 PM Pine W  wrote:
> >
> > > Hi Kiril,
> > >
> > > Thank you for sharing your proposal.
> > >
> > > I am concerned about the possibility of Wikipedia being used as a
> > > laboratory for experiments that consume volunteers' time and/or
> > > personal data, and don't benefit Wikipedia or its participants. Does
> > > your research benefit the community, and if so, how? It sounds like
> > > your research intends to develop a model of decision trees for
> > > individual Wikipedians, and at first read I don't understand how the
> > > individual research subjects or the community would benefit.
> > >
> > > Sorry if this sounds defensive, but I hope that you understand why I'm
> > > asking.
> > >
> > > Pine
> > > ( https://meta.wikimedia.org/wiki/User:Pine )
> > >
> > > On Mon, Feb 24, 2020 at 6:00 PM Kiril Simeonovski
> > >  wrote:
> > > >
> > > > Hi all,
> > > >
> > > > I am currently working on a research concerned with modelling user
> > > > behaviour on Wikipedia. The idea is to design a field experiment
> over a
> > > > random sample of Wikipedians in order to examine their risk
> preferences
> > > and
> > > > define (dis)utilities that will be used in a utility-maximisation
> model.
> > > >
> > > > I have already submitted an abstract that got accepted for the
> > > > biennial Foundations
> > > > of Utility and Risk Conference 2020 
> > > and my
> > > > future plans include presentation of the concept at other research
> > > > conferences (including Wikimania 2020).
> > > >
> > > > You can visit the project page
> > > > <
> > >
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> > > >
> > > > of this research on Meta. Your questions and comments are welcome at
> any
> > > > time. Thank you!
> > > >
> > > > Best regards,
> > > > Kiril
> > > > ___
> > > > Wiki-research-l mailing list
> > > > Wiki-research-l@lists.wikimedia.org
> > > > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> > >
> > > ___
> > > Wiki-research-l mailing list
> > > Wiki-research-l@lists.wikimedia.org
> > > 

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Stuart A. Yeates]

Hi Kiril

Let's just say that history has taught us to be risk-averse to
drive-by researchers.

Can you point us to other research output using this methodology? Do
you (or any of your team) have significant editing experience? Are you
familiar with the firestorm that is paid editing and sock puppetry??

cheers
stuart

--
...let us be heard from red core to black sky

On Tue, 25 Feb 2020 at 10:43, Kiril Simeonovski
 wrote:
>
> Hi Pine,
>
> The findings from the research will be articulated to draw clear
> conclusions about what causes utility and disutility from participation,
> and how this is perceived by different editors. For instance, it is natural
> to assume that editors come to contribute by adding content that will
> remain visible, while blocks and reverted edits are risk factors that drive
> them away, although different editors have different levels of risk
> aversion. Similarly to any other research, the benefit for the community
> and individual editors is going to be indirect but yet not insignificant to
> be accepted in the future process of decision-making (if the research
> demonstrates the existence of high level of risk aversion towards
> something, then it automatically signals that doing that thing is harmful
> for the environment).
>
> I know that it's impossible to predict the extent to which this research
> would make impact because the body of literature is very poor on
> volunteer-driven environments in a dynamic setting but it's definitely
> worth to start off something that might attract the attention of
> researchers in this direction. At the end, the research is not meant to
> carve rules in stone that any single editor should respect but rather to
> suggest something that individuals and communities might find useful (the
> means of doing this will definitely not turn Wikipedia into a laboratory or
> put someone's privacy in danger).
>
> Best,
> Kiril
>
> On Mon, Feb 24, 2020 at 9:43 PM Pine W  wrote:
>
> > Hi Kiril,
> >
> > Thank you for sharing your proposal.
> >
> > I am concerned about the possibility of Wikipedia being used as a
> > laboratory for experiments that consume volunteers' time and/or
> > personal data, and don't benefit Wikipedia or its participants. Does
> > your research benefit the community, and if so, how? It sounds like
> > your research intends to develop a model of decision trees for
> > individual Wikipedians, and at first read I don't understand how the
> > individual research subjects or the community would benefit.
> >
> > Sorry if this sounds defensive, but I hope that you understand why I'm
> > asking.
> >
> > Pine
> > ( https://meta.wikimedia.org/wiki/User:Pine )
> >
> > On Mon, Feb 24, 2020 at 6:00 PM Kiril Simeonovski
> >  wrote:
> > >
> > > Hi all,
> > >
> > > I am currently working on a research concerned with modelling user
> > > behaviour on Wikipedia. The idea is to design a field experiment over a
> > > random sample of Wikipedians in order to examine their risk preferences
> > and
> > > define (dis)utilities that will be used in a utility-maximisation model.
> > >
> > > I have already submitted an abstract that got accepted for the
> > > biennial Foundations
> > > of Utility and Risk Conference 2020 
> > and my
> > > future plans include presentation of the concept at other research
> > > conferences (including Wikimania 2020).
> > >
> > > You can visit the project page
> > > <
> > https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> > >
> > > of this research on Meta. Your questions and comments are welcome at any
> > > time. Thank you!
> > >
> > > Best regards,
> > > Kiril
> > > ___
> > > Wiki-research-l mailing list
> > > Wiki-research-l@lists.wikimedia.org
> > > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> > ___
> > Wiki-research-l mailing list
> > Wiki-research-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> ___
> Wiki-research-l mailing list
> Wiki-research-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

___
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Pine,

The findings from the research will be articulated to draw clear
conclusions about what causes utility and disutility from participation,
and how this is perceived by different editors. For instance, it is natural
to assume that editors come to contribute by adding content that will
remain visible, while blocks and reverted edits are risk factors that drive
them away, although different editors have different levels of risk
aversion. Similarly to any other research, the benefit for the community
and individual editors is going to be indirect but yet not insignificant to
be accepted in the future process of decision-making (if the research
demonstrates the existence of high level of risk aversion towards
something, then it automatically signals that doing that thing is harmful
for the environment).

I know that it's impossible to predict the extent to which this research
would make impact because the body of literature is very poor on
volunteer-driven environments in a dynamic setting but it's definitely
worth to start off something that might attract the attention of
researchers in this direction. At the end, the research is not meant to
carve rules in stone that any single editor should respect but rather to
suggest something that individuals and communities might find useful (the
means of doing this will definitely not turn Wikipedia into a laboratory or
put someone's privacy in danger).

Best,
Kiril

On Mon, Feb 24, 2020 at 9:43 PM Pine W  wrote:

> Hi Kiril,
>
> Thank you for sharing your proposal.
>
> I am concerned about the possibility of Wikipedia being used as a
> laboratory for experiments that consume volunteers' time and/or
> personal data, and don't benefit Wikipedia or its participants. Does
> your research benefit the community, and if so, how? It sounds like
> your research intends to develop a model of decision trees for
> individual Wikipedians, and at first read I don't understand how the
> individual research subjects or the community would benefit.
>
> Sorry if this sounds defensive, but I hope that you understand why I'm
> asking.
>
> Pine
> ( https://meta.wikimedia.org/wiki/User:Pine )
>
> On Mon, Feb 24, 2020 at 6:00 PM Kiril Simeonovski
>  wrote:
> >
> > Hi all,
> >
> > I am currently working on a research concerned with modelling user
> > behaviour on Wikipedia. The idea is to design a field experiment over a
> > random sample of Wikipedians in order to examine their risk preferences
> and
> > define (dis)utilities that will be used in a utility-maximisation model.
> >
> > I have already submitted an abstract that got accepted for the
> > biennial Foundations
> > of Utility and Risk Conference 2020 
> and my
> > future plans include presentation of the concept at other research
> > conferences (including Wikimania 2020).
> >
> > You can visit the project page
> > <
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> >
> > of this research on Meta. Your questions and comments are welcome at any
> > time. Thank you!
> >
> > Best regards,
> > Kiril
> > ___
> > Wiki-research-l mailing list
> > Wiki-research-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
>
> ___
> Wiki-research-l mailing list
> Wiki-research-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
>
___
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Stuart A. Yeates]

Removing accounts with low numbers of edits will, of course, blind
your analysis to users who use throw-away accounts, even when they can
clearly be attributed to the same individual.

cheers
stuart
--
...let us be heard from red core to black sky

On Tue, 25 Feb 2020 at 08:42, Kiril Simeonovski
 wrote:
>
> Hi Stuart,
>
> Yes, all those terms refer to a Wikipedia account. My plan is to avoid
> sampling accounts with very low activity (probably less then a minimum
> threshold of edits) because of the impracticality to draw any conclusion
> from them.
>
> Best,
> Kiril
>
> On Mon, Feb 24, 2020 at 8:36 PM Stuart A. Yeates  wrote:
>
> > When you say "participant", "user" and "editor" do you actually mean
> > account?
> >
> > I routinely notice what appear to be people attending real-file events
> > using one account but then editing afterwards with a different
> > account.
> >
> > cheers
> > stuart
> > --
> > ...let us be heard from red core to black sky
> >
> > On Tue, 25 Feb 2020 at 07:00, Kiril Simeonovski
> >  wrote:
> > >
> > > Hi all,
> > >
> > > I am currently working on a research concerned with modelling user
> > > behaviour on Wikipedia. The idea is to design a field experiment over a
> > > random sample of Wikipedians in order to examine their risk preferences
> > and
> > > define (dis)utilities that will be used in a utility-maximisation model.
> > >
> > > I have already submitted an abstract that got accepted for the
> > > biennial Foundations
> > > of Utility and Risk Conference 2020 
> > and my
> > > future plans include presentation of the concept at other research
> > > conferences (including Wikimania 2020).
> > >
> > > You can visit the project page
> > > <
> > https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> > >
> > > of this research on Meta. Your questions and comments are welcome at any
> > > time. Thank you!
> > >
> > > Best regards,
> > > Kiril
> > > ___
> > > Wiki-research-l mailing list
> > > Wiki-research-l@lists.wikimedia.org
> > > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> > ___
> > Wiki-research-l mailing list
> > Wiki-research-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> ___
> Wiki-research-l mailing list
> Wiki-research-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

___
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Kiril Simeonovski]

Hi Stuart,

Yes, all those terms refer to a Wikipedia account. My plan is to avoid
sampling accounts with very low activity (probably less then a minimum
threshold of edits) because of the impracticality to draw any conclusion
from them.

Best,
Kiril

On Mon, Feb 24, 2020 at 8:36 PM Stuart A. Yeates  wrote:

> When you say "participant", "user" and "editor" do you actually mean
> account?
>
> I routinely notice what appear to be people attending real-file events
> using one account but then editing afterwards with a different
> account.
>
> cheers
> stuart
> --
> ...let us be heard from red core to black sky
>
> On Tue, 25 Feb 2020 at 07:00, Kiril Simeonovski
>  wrote:
> >
> > Hi all,
> >
> > I am currently working on a research concerned with modelling user
> > behaviour on Wikipedia. The idea is to design a field experiment over a
> > random sample of Wikipedians in order to examine their risk preferences
> and
> > define (dis)utilities that will be used in a utility-maximisation model.
> >
> > I have already submitted an abstract that got accepted for the
> > biennial Foundations
> > of Utility and Risk Conference 2020 
> and my
> > future plans include presentation of the concept at other research
> > conferences (including Wikimania 2020).
> >
> > You can visit the project page
> > <
> https://meta.wikimedia.org/wiki/Research:Modelling_Behaviour_in_a_Peer_Production_Economy_upon_Evidence_from_Wikipedia
> >
> > of this research on Meta. Your questions and comments are welcome at any
> > time. Thank you!
> >
> > Best regards,
> > Kiril
> > ___
> > Wiki-research-l mailing list
> > Wiki-research-l@lists.wikimedia.org
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
>
> ___
> Wiki-research-l mailing list
> Wiki-research-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
>
___
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

Re: [Wiki-research-l] Modelling user behaviour on Wikipedia

[Stuart A. Yeates]

When you say "participant", "user" and "editor" do you actually mean account?

I routinely notice what appear to be people attending real-file events
using one account but then editing afterwards with a different
account.

cheers
stuart
--
...let us be heard from red core to black sky

On Tue, 25 Feb 2020 at 07:00, Kiril Simeonovski
 wrote:
>
> Hi all,
>
> I am currently working on a research concerned with modelling user
> behaviour on Wikipedia. The idea is to design a field experiment over a
> random sample of Wikipedians in order to examine their risk preferences and
> define (dis)utilities that will be used in a utility-maximisation model.
>
> I have already submitted an abstract that got accepted for the
> biennial Foundations
> of Utility and Risk Conference 2020  and my
> future plans include presentation of the concept at other research
> conferences (including Wikimania 2020).
>
> You can visit the project page
> 
> of this research on Meta. Your questions and comments are welcome at any
> time. Thank you!
>
> Best regards,
> Kiril
> ___
> Wiki-research-l mailing list
> Wiki-research-l@lists.wikimedia.org
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l

___
Wiki-research-l mailing list
Wiki-research-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wiki-research-l