Deleted Edits?

Deleted Edits?

Are you just analysing undeleted edits or do you also have access to deleted ones?

Among newbies with less than ten edts deleted edits are likely to be article creation, amongst longer term editors it is also likely that deleted edits would also have been tagging articles for deletion as well as creation of articles or editing of new pages created by others.

WereSpielChequers20:02, 20 October 2010

We are using the stub-meta-history.xml files and they contain meta information about all edits, as far as I know, included the reverted edits. Thus we are not biasing the results by excluding people whose edits have been removed (and thereby raising the bar to become a new_wikipedian or active editor).

Drdee00:49, 21 October 2010

Reverted edits are still available to all and I'm sure you would be using them. But deleted edits can normally only be accessed by admins, hence my question as I suspect you won't be including deleted edits and that would give you a completely different picture re the newbie who comes along creates an article or two and leaves after their article is deleted. I don't know what the stub-meta-history.xml files contain, but if they generally available and are a copy of the edits rather than a count of edits by userid then I would be surprised if they contained deleted edits.

WereSpielChequers13:44, 21 October 2010

Thanks for pointing this out to me. Let me ask about this and let's see how to address this.

Drdee13:46, 21 October 2010

Dear WereSpielChequers:

Yes, you are right: the stub-meta-history.xml file does not contain deleted edits and hence deleted articles do not count as edits. In this regard, we try to stay as close to the numbers reported on stats.wikimedia.org to avoid creating confusion. In addition, what I understand is that the deleting of pages is a more recent trend: it rarely happened in the beginning days but has become more profound recently (see also http://meta.wikimedia.org/wiki/Deletionism).

If time permits and data is readily available we might include this to see how it would affect our analyzes. Thank you very much for bringing this to my attention.

Drdee21:29, 22 October 2010

Thanks, I think we will find that there are several ways in which people start editing Wikipedia, and that creating a new article is one route that has becoming increasingly problematic. Another way to identify these editors is from messages on their user talk page, as usually they will get a templated message when their article is tagged for deletion. The messages vary somewhat, in my view the ones for vandalism, attack pages and spam are usually accurate and I'm not greatly concerned if those editors leave. But many of the others are "good faith" potential contributors even if their first edits weren't necessarily ideal.

WereSpielChequers16:14, 27 October 2010

Our first focus is on namespace 0, hence we do not include Talk pages (yet), the reason is that way we keep close to stats.wikimedia.org. But I completely agree with you that the Talk pages contain a lot of valuable information on the experiences that new editors have. Personally, I think natural language analysis could be a promising road to take to uncover micro dynamics of editors joining / leaving.

Drdee19:02, 27 October 2010