Jump to content

A couple of conversation starters

Emails, private conversations and other unpublished sources may be used by researchers, but they are certainly not accepted as reference in the scientific magazines (at least not in my field). I'm not sure if we should use them, if the literature itself prohibits it. This is in any case highly specialistic content. It is only likely to be an issue for our largest projects such as wp-en or wp-es.

I'll give an example why I think it would be good to have a manual for using scientific publications in natural sciences as a source. Any paper will have a certain structure (abstract-intro-method-results-discussion). Depending on the place in the paper, information may be a primary or secondary source. Alerting the user on how to read and use a paper could decrease the amount of unsound references, which is imho a problem at all larger projects.

Woodwalker15:43, 24 November 2009

I agree about emails and conversations, as they can get outside the realm of verifiability and reliability. So, are you proposing that we have different standards/manuals for each of a number of different subjects? Also, if we are following scientific journals and scientific principles then we perhaps should follow scientific principles of the definitions of primary and secondary research. Perhaps a matrix of what we will and will not utilize from scientific inquiry is in order?

While I do not argue with your points, my concern is that non academics may have trouble with academic guidelines, no matter how sensible they may be. How do we convey academic standards to non academics in a way they can understand?

Bhneihouse03:23, 27 November 2009

Being a bit controversial -- we don't.

We set up a better editing environment, with better odds of more success, and changes that will probably catalyze and feed through into such areas and issues... and then focus on getting our basic raw editors and basic "not fit to eat" articles up to a basic standard. best odds all round. We can't afford to do much more, I think that has to be our first focus.

If we have a focus left over, then general improvements for established editors, and we can debate what else we can squeeze in. However much we value experts, and suffer when they leave, they leave because of general issues with the editing environment and poor editorship and disputes, all general issues. Focus on those, which also affect everyone else too. Don't try to make special "expert only" stuff, or fix specialist issues with academic sourcing. Wikipedia's standards and community aren't yet at a point where these specialist issues are a "top 5" focus.

Apologies if this sounds controversial or such. Discuss if needed, but after thinking through how to make best use of an opportunity like this, that's more and more where it seems to come down to. We have to get the basics needed for improvement. that means specialisms such as experts, FAs and the like are not a priority (this time around). Ruthless but ... sigh... hard choices.

FT2 (Talk | email)04:51, 27 November 2009

@FT2: it doesn't sound controversial. Yet some projects, such as wp-en and wp-de, are certainly at an academic level in some areas. I agree these projects should focus on areas that are behind, but there's also the principle that writing a new article is five times less work than rewriting old stuff. I think there is a serious problem growing inside the larger projects concerning citation of (academic?) literature. At the English Wikipedia, longer articles often have hundreds of references to scientific papers. Following these refs, I often find they are simply unsound.

  • They refer to papers that describe another thing;
  • They refer to papers that try to disprove the very information they should verify;
  • They refer to papers that are secondary sources for the information they should verify;
  • They refer to papers that don't even contain the information they should verify.

I can live with the secondary/primary problem for the moment, but the large-scale malcitation at wp-en is a serious problem that undermines the project's intention to become a trustworthy source in education or academics.

Let's call it 'the curse of the {{citation needed}}-template'. Instead of citing the source used for writing the article, users tend to search for any citation on the internet that verifies the information in the article, just to get rid of the {{citation needed}}-template. They find abstracts of scientific papers (access to scientific literature on the internet is often limited to the abstracts) and cite without understanding or reading what the papers are about. It's a way of being lazy. Instead of going to a library, the contributor uses Google, reads diagonally through an abstract and voilà he has a ref. Anyone can see this method will never create an encyclopaedic source useful for academic use.

I agree that this isn't our top priority, and I agree the projects should first get to a plateau, a general level of quality, before they try to reach an academic level. Yet imagine the titanic task of having to check thousands of references for their value once the project is there. It's also a difficult task tat requires specialist knowledge about the subject. Most likely it's not going to happen at all. We simply don't have enough expert users in our communities that would be ready to help.

I don't like the idea of having different guidelines for normal and specialist contributors. The solution I'd rather see is simple. We could suggest to implement a guideline not to use a source when one does not understand it well. Only use a scientific paper as a source when you have a thorough knowledge of the subject yourself. If you don't have that specialistic knowledge, go to a library, search for a textbook (as recent and summarizing as possible - a truly sound secondary source) and use that instead. Such a guideline would be a very simple and effective way to make Wikipedia more acceptable for academic use.

Woodwalker08:20, 27 November 2009

Maybe a simple solution would be to have a category and sourcing guidelines ("Category:Articles whose core content should primarily be from current peer reviewed sources"), for those topics where the topic and availability of references means that core topic material should be cited primarily from strict and limited sources.

Users can agree by consensus if an article should move to that category. Bots can also be used to identify and suggest articles where most cites are already from a list of peer reviewed or other highly reputed sources, for review.

This fits within normal handling. Then for such articles, a bot reviews these every while and highlights those cites which are more than 6(?) 12(?) months from publication so our academic experts can check these and update them or ensure they're still currently valid.

So... essentially isn't this something easily done already, without needing the weight of a taskforce and quality project behind it? Any small group of users could propose a mechanism of this kind and the low level of bot, feed, or other infrastructure needed to support it.

The twist is, we can then use the processes being discussed elsewhere to heavily "push" this along. We can ask readers to ("Help Wikipedia! This topic has 3 cites more than 6 months old. Please update them if they're out of date for the field!") or put these onto a feed or user filters, for users interested in the topics. Cites are tagged with the date of publication so a Citation Check Wizard can ask on editing, "I notice you're about to add a 18 month old cite to this topic. Are you sure this still represents latest knowledge in the field?". And so on. Think about it :)

It also shows how the seemingly simple recommendations we've been discussing can also be used to improve quality in other ways and places - including even specialisms like expert sourcing, which was never previously considered. That's the kind of pervasive effect that we can aim for. Notice we were considering wizards and notifications in a "basic quality" and "new user" context, but once added they can open whole new quality doorways across the wiki.

FT2 (Talk | email)08:46, 27 November 2009

I just read your post in the other thread and this Citation Wizard, wow, I really like the idea! An independent Citation Check Wizard would be great too.

The problem I wanted to point out in my last post (08:20, 27 November 2009) wasn't the problem of outdated sources (Actually, 18 months seems little time for me, 10 to 20 years is already quite good for scientific papers). It's the problem of wrong citations due to a lack of understanding in the subject (see the list of examples I gave in my last post - they are unfortunately quite common). The citation wizard could help, for example by giving advice depending on expertise:

  • You are:
    • An expert for this subject, knowing well the consensus in the literature -> either use primary sources (scientific papers) or secondary sources (textbooks), depending on how general the information is.
    • Just interested but knowing a lot about the subject -> use secondary sources (textbooks), only use primary sources (scientific papers) if you are sure what you're doing.
    • Just searching for a source that can replace the stupid template! -> try to find a secondary source (textbook or neutral website) which has the same subject as this article or section.
Woodwalker10:33, 27 November 2009