Assessing the Reliability of Scientific Information on Wikipedia

[Astronuc]

This is even better.

Congress 'made Wikipedia changes'
http://news.bbc.co.uk/2/hi/technology/4695376.stm

Online reference site Wikipedia blames US Congress staff for partisan changes to a number of political biographies.
Computers traced to Capitol Hill removed unpalatable facts from articles on senators, while other entries were "vandalised", the site said.

An inquiry was launched after staff for Democratic representative Marty Meehan admitted polishing his biography.

Online reference site Wikipedia blames US Congress staff for partisan changes to a number of political biographies.
Computers traced to Capitol Hill removed unpalatable facts from articles on senators, while other entries were "vandalised", the site said.

An inquiry was launched after staff for Democratic representative Marty Meehan admitted polishing his biography.

Freedom of the press?

 

[rachmaninoff2]

How infuriating! Meehan should fire those staffers at once. And resign. And join a monastery and take a perpetual vow of silence.

This is not an acceptable political tactic. Good thing its clearly impossible to get away with - only an idiot senator would even try.

 

[ZapperZ]

I seldom blame the person who is the center of anything. I seldom blame the politician, the evengelist, etc. I tend to blame the people who are duped into buying what these people were selling. These people would not get anywhere without the hordes of masses backing them.

Just look at what is going on here. Even when there are clear evidence that Wkipedia CAN be highly unreliable, you have nearly 80% who participated in this poll who TRUST the info they get from there.

When something like this occurs, all I can do is throw my hands up in the air and walk away. When you set such low standards for yourself, then you get what you deserve and there's no one and nothing else to blame.

Zz.

 

[Astronuc]

'Tis true that one is ultimately responsible for what one consumes.

On the other hand, I believe that people would like to be able to trust other people. It would be sad to live in a world where one is not able to trust others.

I think one should however be aware that Wikipedia is uncontrolled and that the information be verified if one wishes to ensure accuracy.

 

[ZapperZ]

'Tis true that one is ultimately responsible for what one consumes.

On the other hand, I believe that people would like to be able to trust other people. It would be sad to live in a world where one is not able to trust others.

I think one should however be aware that Wikipedia is uncontrolled and that the information be verified if one wishes to ensure accuracy.

But there is a difference between trust based on knowledge, and a blind trust in spite of evidence showing otherwise. You might forgive someone if he/she lies to you once, maybe twice. But after repeatedly being lied to and you still trust that person, then you deserve to be taken advantage off for all you are worth.

I do not trust Wikipedia. Period. I have seen nothing to change my opinion on it.

Zz.

 

[luckycharms]

Wikipedia is just a form of peer review writ large. Isn't that exactly how scientists organize information dissemination?

The whole framing of the question is silly. Trust or not-trust is an awfully inflexible and simplistic way of viewing Wikipedia. The better question to ask is "what percentage of information do you think is correct?" Considering the sheer number of facts, the answer to that is probably arguing whether it is 99% or 99.9% or 99.99%, maybe slightly less for highly specific information such as science articles.

To completely discredit it because there are some mistakes, or to quibble over technical details, or to complain of over-simplification is to ignore the whole point of the exercise. For the vast amount of information, it is correct. Obviously, subjective material is going to be tainted by opinion, people will lie for personal gain, and highly technical articles will be lacking detail or may be incorrect altogether. I own plenty of graduate level textbooks that contain mistakes or inaccuracies and that's to say nothing of supposedly refereed journal articles. That doesn't mean I don't trust them.

 

[ZapperZ]

Wikipedia is just a form of peer review writ large. Isn't that exactly how scientists organize information dissemination?

Er... peer-reviewed means that the info in reviewed by EXPERTS in that particular area. Where does it say in Wikipedia that only someone who is an acknowledged expert in a particular area are the only one who can review the entry? So how is this "exactly" like peer-review system?

The whole framing of the question is silly. Trust or not-trust is an awfully inflexible and simplistic way of viewing Wikipedia. The better question to ask is "what percentage of information do you think is correct?" Considering the sheer number of facts, the answer to that is probably arguing whether it is 99% or 99.9% or 99.99%, maybe slightly less for highly specific information such as science articles.

To completely discredit it because there are some mistakes, or to quibble over technical details, or to complain of over-simplification is to ignore the whole point of the exercise. For the vast amount of information, it is correct. Obviously, subjective material is going to be tainted by opinion, people will lie for personal gain, and highly technical articles will be lacking detail or may be incorrect altogether. I own plenty of graduate level textbooks that contain mistakes or inaccuracies and that's to say nothing of supposedly refereed journal articles. That doesn't mean I don't trust them.

You "trust" refereed journals articles? Really? I don't! All I do with refereed journal articles is to use them as work that have been done and reviewed to not have an obvious error, omission, and mistakes. The whole point of publishing is to have that work examined and verified by independent parties. It is ONLY after such lengthy and tedious process can one use that work as verified, not a second before.

And there is a difference between errors in textbooks and errors in Wikipedia. In textbooks, one is not surprised by typographical errors, and honest errors where something has been overlooked. On Wikipedia, you can't tell if it is that kind of error, or error due to ignorance of the subject matter by the person writing it. How are you to know someone who didn't know any better wrote such a thing? The person who wrote your texts has his/her reputation at stake and will do all he/she can to make sure things are accurate. The person who wrote an entry in Wikipedia doesn't.

So yes, there IS a difference between the two. Your comparison missed a lot of differences that clearly distinguish one from the other.

Zz.

 

[Astronuc]

Wikipedia is just a form of peer review writ large. Isn't that exactly how scientists organize information dissemination?

The whole framing of the question is silly. Trust or not-trust is an awfully inflexible and simplistic way of viewing Wikipedia. The better question to ask is "what percentage of information do you think is correct?" Considering the sheer number of facts, the answer to that is probably arguing whether it is 99% or 99.9% or 99.99%, maybe slightly less for highly specific information such as science articles.

To completely discredit it because there are some mistakes, or to quibble over technical details, or to complain of over-simplification is to ignore the whole point of the exercise. For the vast amount of information, it is correct. Obviously, subjective material is going to be tainted by opinion, people will lie for personal gain, and highly technical articles will be lacking detail or may be incorrect altogether. I own plenty of graduate level textbooks that contain mistakes or inaccuracies and that's to say nothing of supposedly refereed journal articles. That doesn't mean I don't trust them.

Well, I suppose one would have to qualify "Trust", as in does one "trust" Wikipedia 100%. I would have to say, NO!

Also, it's a matter of quality control and those of us who work under mandatory quality control programs, would have to say, strictly no to trusting Wikipedia. There is NO control!

As they say, "the free encyclopedia that anyone can edit."

A refereed journal at least controls who contributes and who referees. That is not necessarily the case at Wikipedia.

And yes, textbooks and articles in refereed journals can and do contain mistakes. But the rate of error is probably much much less than Wikipedia.

 

[Mk]

Er... peer-reviewed means that the info in reviewed by EXPERTS in that particular area. Where does it say in Wikipedia that only someone who is an acknowledged expert in a particular area are the only one who can review the entry? So how is this "exactly" like peer-review system?

Your average idiot doesn't like to write encyclopedia article all day. They're not all experts, but they sure damn know a lot.

 

[Astronuc]

Well, as for trusting Wikipedia or not, how's this?

http://en.wikipedia.org/wiki/Saint_Saens

The neutrality of this article is disputed.
Please see the discussion on the talk page.

How controversial could an article be on Charles Camille Saint-Saëns. Well apparently it can.

The author may not be an idiot, but may be expressing some bias. As for accuracy, I'd have to corroborate with other sources.

I have also found glaring errors in topics in nuclear energy. I haven't corrected them, and frankly I don't need the aggravation.

 

[luckycharms]

Er... peer-reviewed means that the info in reviewed by EXPERTS in that particular area. Where does it say in Wikipedia that only someone who is an acknowledged expert in a particular area are the only one who can review the entry? So how is this "exactly" like peer-review system?

You must be a theorist. Consider it in practice, not in theory. The person writing or editing a wikipedia article is more than likely to have some degree of expertise in the subject matter. Perhaps a pastry chef will suddenly be inspired to write an article on nuclear fission, but I'll bet he'll stick to cannolis. Just because anyone can doesn't mean that anyone will.

And this is the fundamental faith you put into Wikipedia: that complete idiots aren't writing everything. And this faith, naive as it may seem, results in a remarkably high quality of general knowledge reference. How else can you explain that wikipedia gets so many things right? By the skeptic's model, we should expect nothing more than monkeys typing.

Perhaps I shouldn't have said "exactly," but the trust inherent in the peer review process is the same whether the experts are policed centrally or by the community at large. And the results aren't all that different. I think what the Wikipedia model shows is that a disorganized collection of people online will segregate, coalesce, organize and police itself nearly as efficiently as the familiar and formal offline counterparts.

Furthermore, you incorrectly stated previously that there is no record of change. In fact, there is a detailed log of each and every change and addition including discussion and dispute. There is obviously a difference of degree, but there is no reason to assume that people don't value and protect their online reputation just as any scientist would protect his academic one.

Unless you have no faith in market dynamics or statistics, the sheer number of Wikipedia users is reason enough to "trust" the veracity of an article. Its popularity is testament to its accuracy. As with any source, you have to take information with a grain of salt commensurate with the reputation of the source, but you seem to be saying that there is no grain of salt big enough for you to swallow Wiki.

I'm not saying that Wiki is more correct than an expert's textbook or first-hand exploration of a subject. I'm just saying the results are pretty darn good for all the obvious flaws in the model.

 

[ZapperZ]

You must be a theorist.

Obviously, whatever it is that you used to make that conclusion, it is extremely faulty. And this is the proof.

Consider it in practice, not in theory. The person writing or editing a wikipedia article is more than likely to have some degree of expertise in the subject matter. Perhaps a pastry chef will suddenly be inspired to write an article on nuclear fission, but I'll bet he'll stick to cannolis. Just because anyone can doesn't mean that anyone will.

Really? And what made you say that? I can pull out the Superconductivity section of Wikipedia and come up with (i) inacurate description (ii) misleading description, and (iii) confusing description. Forget about getting the facts right. There is ZERO attempt to make every part of it coherent. This is something textbook writers are faced with with producing a book, how to present an idea systematically so that one connects with the other, and better still, they don't appear to contradict each other. And when they do, point those out and why it is a special case.

Science isn't just a series of "facts". They are interconnected. And when you have different parts even within the SAME topic being written and edited by different people ALL THE TIME, you get an incoherent concept being presented by not one, but SEVERAL "pastry chefs", all wanting to present it in their own way. And people "trust" these stuff?

And this is the fundamental faith you put into Wikipedia: that complete idiots aren't writing everything. And this faith, naive as it may seem, results in a remarkably high quality of general knowledge reference. How else can you explain that wikipedia gets so many things right? By the skeptic's model, we should expect nothing more than monkeys typing.

It has nothing to do with how many things it got it right. When you have such a large phase space to work with, even a monkey can pick things right at random. It is how often it gets it wrong, and to what extent it gets it wrong, and to what guarantee does it stand by its content. You don't go to an annonymous person for medical advice without knowing if he/she is a legitimate doctor and has the degree and knowledge to back it up. Yet, you trust knowledge coming from annonymous people whose credentials you haven't checked, and whose track records are unknown.

Call me crazy, but you trust that kind of a source?

Perhaps I shouldn't have said "exactly," but the trust inherent in the peer review process is the same whether the experts are policed centrally or by the community at large. And the results aren't all that different. I think what the Wikipedia model shows is that a disorganized collection of people online will segregate, coalesce, organize and police itself nearly as efficiently as the familiar and formal offline counterparts.

I trust the peer-review process because I know what goes on in it. I am a journal referee myself. I do NOT, however, trust the VALIDITY of the paper appearing in a peer-review journal BECAUSE it is simply the beginning of the verification process. It appears in print, it doesn't mean it is true. So who "peer-reviewed" the photoemission spectroscopy description in Wikipedia?

Furthermore, you incorrectly stated previously that there is no record of change. In fact, there is a detailed log of each and every change and addition including discussion and dispute. There is obviously a difference of degree, but there is no reason to assume that people don't value and protect their online reputation just as any scientist would protect his academic one.

Unless you have no faith in market dynamics or statistics, the sheer number of Wikipedia users is reason enough to "trust" the veracity of an article. Its popularity is testament to its accuracy. As with any source, you have to take information with a grain of salt commensurate with the reputation of the source, but you seem to be saying that there is no grain of salt big enough for you to swallow Wiki.

I think even you know that "popularity" has nothing to do with "quality". McDonalds is "popular", so its food must be better than a 3 star restaurant that has less number of customers in a year? First of all, how is someone who is LOOKING for an answer to something, would KNOW that the thing he/she is reading on Wikipedia is CORRECT? One only needs to look on here at all the absurd references people have made using Wikipedia to back up with faulty idea. So of course they think the info they're looking for is correct!

And I could play this game too. If Wikipedia is THAT great, how come less than 10% of practicing scientists actually actively use it? After all, from the last Nature survey, I believe more than 30% (or was it 50%) of scientists surveyed have heard of it. If it is that great, how come it plays an insignificant role in active research work? I can point out to you many texts that I had, even as an undergrad, that I often still refer too even in my capacity as a physicist. So this is already another difference between Wikipedia and textbooks.

One could argue, as has been mentioned already in this thread, that it isn't MEANT as an indepth source, or a primary source. And I'd say EXACTLY! Not only is the source of the info dubious, it also means that it doesn't attempt to be rigorous. It is meant to be dumbed down and superficial, and often, this is where it gets into trouble at being "simple" and "accurate". Having helped someone wrote a book before, there were many nights where we struggled at getting JUST the right word, or passage, or even phrases so that our simplified idea does not convey the the wrong impression or the wrong idea to someone just barely learning the subject matter. And this is even before the publisher or book editor even got their hands on the manuscript!

Now how many Wikipedia authors do you think put THAT much care and thought into such a thing. Want to give the people the wrong impression about wave-particle duality? Why not? They won't know any better if we tell them they're just cows! That's why they're looking it up!

Zz.

 

[luckycharms]

Are you really equating life and death medical decisions with encyclopedia articles?

Your standard here seems to be whether it compares to a textbook, a standard which no one is claiming. A textbook and an encylopedia are entirely different in scale, scope, and audience.

You don't go to McDonald's for a three star meal. You go to McDonald's because it fulfills a set of expectations such as cost, convenience, and taste. Wiki isn't trying to be a three star filet mignon textbook, it's trying to be the Big Mac of knowledge. You don't go to McDonald's for your anniversary dinner, you go when you need to grab a quick bite. Wikipedia is no different.

I don't understand the relevance of the Nature statistics. I wouldn't expect practicing scientists to use wikipedia for highly specialized scientific research any more than I would expect them to use Britannica to split the atom. That doesn't mean encylopedias are utterly useless.

The only trust I or anyone else seems to be claiming is that one can "trust" Wikipedia to give a basic overview, established facts, and a starting point for further exploration. Nothing more.

 

[ZapperZ]

Are you really equating life and death medical decisions with encyclopedia articles?

Your standard here seems to be whether it compares to a textbook, a standard which no one is claiming. A textbook and an encylopedia are entirely different in scale, scope, and audience.

You don't go to McDonald's for a three star meal. You go to McDonald's because it fulfills a set of expectations such as cost, convenience, and taste. Wiki isn't trying to be a three star filet mignon textbook, it's trying to be the Big Mac of knowledge. You don't go to McDonald's for your anniversary dinner, you go when you need to grab a quick bite. Wikipedia is no different.

Note that it was you who made the comparison of Wikipedia being no different with peer-review journals AND textbooks. Furthermore, you used its popularity as "evidence" of its validity. Being popular and being valid are not mutually inclusive.

I don't understand the relevance of the Nature statistics. I wouldn't expect practicing scientists to use wikipedia for highly specialized scientific research any more than I would expect them to use Britannica to split the atom. That doesn't mean encylopedias are utterly useless.

The only trust I or anyone else seems to be claiming is that one can "trust" Wikipedia to give a basic overview, established facts, and a starting point for further exploration. Nothing more.

And maybe you need to read a bit deeper into the thread and to note that for MANY people, Wikipedia IS their primary source of info. And this isn't just "theory", it is practice! All you need to do is do pay attention long enough to the various science forums on here and you'll have your example soon enough. If the source is shaky, then it shouldn't be used as a primary source of info. It makes no sense!

I know of people who use Wikipedia as a beginning part of simply to know where to look. I have zero problem with that. It automatically means that you are using it to ask for directions and use other sources for your info. It also means that you don't rely on the content (these all, btw, are symptoms of not completely trusting the validity of the content). But look at how many people actually use it THAT way, especially on the 'net! People seem to believe in a source that easily, and these are the very same people who are not well verse on that subject matter that they're looking up.

Again, if one gets suckered into things like this THAT easily, then one deserves what one gets. No one and nothing else is to blame, not even Wikipedia.

Zz.

 

[luckycharms]

Popularity can be a very good judgement of quality in some cases. I didn't mean to imply that it was automatic.

Hey, I'll totally agree with you on the frustration wrought by people using wiki as a primary source. For basic scientific info, say melting points or simple formulas, I don't see any problem with using it as a quick and dirty reference. Unfortunately, many people don't know what to filter out and where to draw the line.