I agree with this. One thing I found is the large number of sources that are no longer there or not accurate. It’s easy to source something to a book that can’t be easily verified and I have found many sources to links no longer working.
It’s a good place to start
See, individual fuck-ups (eventually) get corrected; when I was working on an article about gold, the article I was editing listed the top gold reserve-holding nations as the US, Germany, Italy, and France. I checked Wikipedia just to see what it said and saw China listed with about 600 tons more than Germany, which threw me for a loop until I followed the link to Wikipedia’s source document for that data, which had China with 1,000 some odd tons instead of 4,000. By now, it’s back to correct. But if I’d relied on Wikipedia at the time, I would have been screwed.
And here’s a thread where I busted someone on using a poorly-sourced statistic from Wikipedia instead of clicking through to the source report and coming up with really crappy statistical arguments as a result. (Again the Wikipedia article seems to have been more or less corrected in the 6 months since).
Not only that, but what reference gives you a chance to understand the arguments of the different views? For example, certain hard-core libertarians (generally von Mises scions) consider Abraham Lincoln to have been a despot and every now and then these accusations pop up on Honest Abe’s wiki page. These are argued back and forth on the discussion page and allows others to understand the accusations and realize they are bunk.
I love wiki and use it for any non-controversial topic. Controversial articles (like AGW) sometimes get mired in edit wars.
What’s “really bad?” I think most really egregious errors of fact probably get weeded out quickly in all but the most obscure of topics. But that still leaves room for smaller errors. An example I’ve used before is the Battle of Bouvines. The following line…
Philip now took the offensive himself, and in maneuvering to get a good cavalry ground upon which to fight he offered battle (27 July), on the plain east of Bouvines and the river Marque.
…is frankly wrong and gives a very misleading idea of the events as they transpired. The odd thing is that the citation at the bottom gives a perfectly good description of the battle. What actually transpired is Philip was caught moving away from Otto ( who had moved with unexpected speed ) with his army deployed halfway across a bridge, with his rearguard already brushed aside - he had to rapidly decide whether to sacrifice some of his best troops or turn to fight in a tough tactical spot. He turned and won, but it is questionable if before that he could have been said to be “taking the offensive” ( he had good reason not to engage ) and he certainly didn’t seek out the imperial army spoiling for a fight. Last of all he did not choose to fight at Bouvines - it was forced on him. If he had chose to fight you’d have thought he would have picked a better spot than his back to river.
This is the kind of errors you find and rather fewer I think every year ( a couple of other entries I checked have been cleaned up since last I checked them ). Doozy? Probably not quite. But enough that I wouldn’t trust it any more than any other encyclopedia and I don’t trust encyclopedias for more than a first look ( years ago I picked several factual errors out of a regular encyclopedia entry on Genghis Khan someone had posted - they’re really no better ). Wikipedia is very, very useful. I love it, but it should always be used with a grain of salt.
Well, there is the page on the Phillipine-American War where it’s claimed the US may hav committed genocide and killed 1.4 million Phillipino civilians.
In actuality, about 16,000 Phillipino soldiers were killed by US military forces and it’s possible that 200,000 civilians died from a cholera outbreak.
And the “Armenian Genocide” page, which is now locked to prevent edits, is a complete mess and most of the cites are from obviously biased sources and polemics.
This topic comes up regularly.
Define “that bad”.
For purely factual, uncontroversial, popular hard science or history subjects, it is quite good.
For anything else, it is better than a “I’m Feeling Lucky” search of the same type using Google.
But not very much.
And I’m willing to test this if you like. We can generate some random subjects of this type and compare the Wikipedia and “Lucky” pages.
Define a “really bad” error.
I will use the examples I use every time this subject comes up.
The **Hemp **article is, IMO, the Gold Standard for crappy, error ridden, Wikipedia articles that can not be made correct with any amount of effort. Consider the following statements:
“Hemp is one of the faster growing biomasses known,[6]”.
No, it isn’t, and reference 6 makes no such claim. All it says is that hemp produces a lot of fibre. I can’t even find the word “biomass” in reference 6. Moreover what does “one of the faster” mean? That it isn’t one of the two slowest growing? That it is in the top 100 fastest growing? The statement is either meaningless or incorrect, and it is not supported by the reference that the author presents as support.
“hemp is very environmentally friendly (with the exception of chemical fertilizers used in industrial agriculture) as it requires few pesticides[10] and no herbicides.[11]”
This statement is flat out wrong. Hemp requires just as much herbicide and pesticide as any similar crop grown in the same locale. Hemp doesn’t magically kill weeds.
“Hemp seeds contain all the essential amino acids and essential fatty acids necessary to maintain healthy human life.”
No, it does not. No plant food contains “all the essential amino acids a…necessary to maintain healthy human life”. In fact “essential amino acids in [Cannbis] (except lysine and sulfur-containing amino acids) are sufficient for the FAO/WHO suggested requirements for 2−5 year old children.” IOW Cannabis does not contain the lysine and sulfur-containing amino acids necessary to support human life.
I have tried to correct these error sin this article multiple times with references from state and federal agriculuture departments, international agriculture departments, the WHO and FAO and prestigious universities and journals. It always gets removed within days.
To look at another article: “Tropical Rainforest”:
“Rainforests are also often called the “Earth’s lungs” because of the fact that they produce about 40% of the Earth’s oxygen”
This statement is flat out wrong. Rainforests consume more oxygen than they produce. Once again, I have tried correcting tis mistake in the past with references to “Nature” and other journals, Senate Inquiries, reputable academics and other unimpeachable sources. It invariably gets changed back with the comment “ZOMG!!! This can’t be true, everybody knows rainforests produce oxygen”.
Are these “really bad” errors.
Well they are easily provable as being wrong.
They are all errors that I have attempted to correct multiple times with unimpeachable references.
They are always reverted to the erroneous form.
One of the error sis potentially life threatening dietary advice.
In comparison to the importance of the value of the articles themselves they are really bad.
Will those examples do for you?
Would you consider Wikipedia to be a *good *reference source for subjects such as Rainforest or Hemp when it contains errors of that magnitude?
PMSL.
I just checked the references in the “Tropical Rainforest” article
The references (11 and 12) that are supposed to support the statement that “Rainforests are also often called the “Earth’s lungs” because of the fact that they produce about 40% of the Earth’s oxygen.[11][12]”
Those are references that I put there originally over 5 years ago.
Both references explicitly state that rainforests do *not *produce oxygen. One of them is even quoted in the references section as "It took more than 15 years for the “lungs of the world” myth to be corrected. Rainforests contribute little net oxygen additions to the atmosphere through photosynthesis.”
Yet this reference is quoted in the text as supporting the statement that rainforests are the lungs of the Earth and that they produce almost half the Earths oxygen.
How much more evidence do you need that Wikipedia is not reliable?
How much more evidence do you need that the editing process is an utter failure?
I am guessing that, after I gave up, there was an edit war in the article, and somebody restored my references and a statement to the effect that rainforests do not produce much oxygen. That was then the subject of an edit war and that editor also gave up in disgust. So some idiot simply wrote in what they read somewhere else and left the references untouched.
Despite the fact that the references say diametrically the opposite to what is claimed in the text, this edit has stood unchallenged for over 5 years. This is a statement of fact that is easily proved using the references already in the article.
Yet in 5 years not one single editor has done so.
How much more evidence do you want of errors and failure of the editing process in Wikipedia?
What sort of error could I find in the article that you would consider to be a “really bad” error? If it saying that forests produce oxygen when they do not is not “really bad” what is? A statement that trees are really larval elephants?
And this error has stood uncorrected for 5+ years despite the fact that it contradicts both the original edit and the references themelves. How much more evidence do you need that the edit function of Wikipedia is a total failure? In 5 years not one editor has noticed a revision of this magnitude?
And this is not an obscure article like the Battle of Bouvines. Everybody in the world knows what a tropical rainforest is. This is a subject that millions of schoolkids will be studying every day worldwide.
Nor is this a position that might be controversial like the number of Phillipine War dead. This is a position of mathematical fact that is not disputed by any ecologist, biometrician or biogeochemist. It is totally uncontroversial and unanimous amongst scientists.
How blatant, factual and incontivertible does an error have to be….? How important and popular does the article have to be…? How glaring does the error have to be……?
Before we can use it as an ironclad example that Wikipedia simply is not reliable for any topic that is even vaguely controversial?
Well if you can find an examples, in Britannica, of errors of the same magnitude as the ones that I have shown you in Wikipedia, then please do so. I would love to see them
If you can;t then the answer is “Yes, it’s possible, just as it’s possible there is a dragon in my garage. This is a classic argument from ignorance. If you can’t present evidence of such errors in Wikipedia then we assume that they do not exist”.
Did you bother to look up the article on “Race”? I would be astounded if it were not explicitly defined in there.
Remember, Britannica has always been designed to be used as a repository, not as a single article source. They have never defined all terms in all articles because that is impossible. But all important terms will be defined in other articles.
As for the fact that the claim was made, that is not an error. That was the scientific consensus of the day, that there were lesser and higher races. How is an encylopedia in error for presenting the near unanimous position of scientists in its articles?
Do you also consider it an error that the same edition of Britannica said that atoms were indivisible particles and that light traveled though an aether at variable speeds?
And yet I have provided at least one example of a blatant error that has remained undetected for at least half a decade. That error is in a highly popular and regularly edited article. That article has been edited 6 times in the last two days alone. I don’t know how to find stats on how often it has been read, but I guess it gets read hundreds of times each day.
It is an error of provable fact that flatly contradicts even the very references that are supposed to support it.
And this is not an example that I went looking for. It was an example that I discovered by chance based on my experience editing the article many years ago. Essentially a random example.
Based on this, constant critiquing and reading clearly does nothing to prevent errors in Wikipedia.
The bit from the tropical rainforest article is certainly not good, but it was changed on January 6 of this year (not 5 years ago). And has now been reverted.
I gave up 5 years ago following an edit war and it was wrong when I finally admitted defeat. I assumed it had remained stuffed up since then.
But the fact remains that an edit that changes the meaning of a section to the exact opposite of what was there previously, and that is contradicted by the references, managed to remain for over a month. And this in a popular, frequently read and critiqued article.
After a month, there is no real reason to assume it would ever be changed back if it hadn’t come to your attention in this thread.
Yes, by you. 
The hemp nonsense remains. That page is an area where I suspect that no editor wants to tread more than once.
Of course, teachers and college professors make extra money by writing articles for encyclopedias and scholarly journals. So they aren’t exactly an unbiased source.
Rather like asking accountants if computer programs like TurboTax are accurate.
That’s true of a lot of high school teachers as well. The school my parents taught at before they retired adopted such a policy.
Beyond that, I think most who do oppose it aren’t worried about their paychecks so much as the way wikipedia treats so many issues.
I think the historian Max Boot did a good job attacking it. This is where I learned of Wikipedia’s statement regarding casualties during the Philippine-American War. And yes, the 1.4 million figure is still up even though no serious historian claims it. Not even such leftists as Gabriel Kolko and Howard Zinn.
Scholarly journals pay their authors the princely sum of $0.00 per article (well, except for the ones where authors have to pay a fee, often quite hefty, to have their articles published).
In my experience, articles commissioned from academics for specialist encyclopedias are paid for at the rate of about $50 to $200 per article (for what may represent several months of work). Nobody is getting rich, or even making a living, off this sort of work.
Teachers and professors neither object to Wikipedia because it threatens their income (it doesn’t), nor because it contains inaccuracies (it does, but not much worse than many other sources). They object because they know that far too many students will, if they are not actively deterred, rely almost entirely on Wikipedia, either paraphrasing or copying and pasting large chunks of what is there, and will thus learn virtually nothing about either the relevant subject matter or the process of real research. Class assignments are supposed to be part of the learning process, but it is almost inevitable that students will treat them as obstacles to getting their final grade that are to be overcome with the least possible effort. Wikipedia is a threat to good education because makes it far too easy for students to complete many sorts of assignment without really learning anything. Teachers who do their best to deter their students from relying on it (sometimes by exaggerating about its inaccuracy) are the ones who care whether their students actually learn anything.
In other respects it is a very useful resource provided it is understood that (like any encyclopedic work) it may contain inaccuracies, and its treatment of any serious topic is inevitably superficial.
I do not really agree, however, that it is very good (I do not say it is useless) as a portal to more in depth information on the topic. In my experience, the references are often chosen more on the basis of what happens to be openly available online (or where the author of some fragment of the article happens to have picked up some factoid) and, even when they really do support the point being made (and as Blake points out, that is by no means always the case), they are rarely the authoritative or original source of the information, and are quite often merely mentioning it in passing (often without themselves citing any authoritative source). Traditional encyclopedias, with entries written by experts, are usually at some pains to cite authoritative sources (and, very often, only an expert will know which are the authoritative sources). Wikipedia seems to have virtually no mechanisms in place to ensure that cited sources are authoritative. (Incidentally, being a research publication in a prestigious scholarly or scientific journal is by no means sufficient to make something an authoritative source.)
I love Wikipedia. I Wikipedia’d my way through my entire degree and did fine. For obvious reasons, you can’t cite it directly in academic papers, but Wikipedia contains a shitload of solid gold info on countless topics.
That would be the Journal of the American Chemical Society, which is the best chem-only journal you can publish in (the Brits and the Germans try to compete, but they fail.)
I use wikipedia all the time, first to check on things I already know but don’t bother having memorized, and second to quickly check for references on chemistry stuffs.
The article on Anacreon boldly states that he lived 582 BC – 485 BC. Which to me appears rather precise.
The article on the movie Atlantis used to say that it had been filmed in New Zealand. When in fact it was the Danish island Sealand.
The article on Sub rosa seems to be rather dubious.
The article on Hizb_ut-Tahrir used to state for a fact that the Koran has the answer to all of life’s questions.
I’m one of the top-50 contributers on the Danish Wikipedia. But actually I ofte find that the German wiki is the best of the lot, and better than the English. The English appears often to be very political.
Here’s an article from people that are experts in this type of stuff:
http://extension.oregonstate.edu/catalog/html/sb/sb681/#Fertilizer
"When grown under favorable conditions, hemp is very competitive with weeds, and no herbicides are generally used in fiber hemp production. Many authors have commented on the exceptional ability of hemp to suppress weed populations (Dewey, 1901 and 1913; Robinson, 1935; Dempsey, 1975, Van der Werf, 1991). Weed suppression with minimal pesticide use is potentially one of the greatest agronomic and environmental benefits of growing hemp in rotation with other crops. Thick stands of hemp have been reported to suppress aggressive weed species including quackgrass (Agropyron repens) (Wright, 1918), bindweed (Robinson, 1935a), and yellow nutsedge (Cyperus esculentus) (Lotz, 1991). The use of hemp to suppress weed populations may offer conventional and organic growers an effective alternative to current weed control practices. Recent commercial experience in the U.K. has shown that with proper timing of planting, weeds can be almost completely suppressed during the hemp growing season. Uneven stand establishment or planting too early, which may lead to stunting of hemp seedlings by cool weather, can allow spring weeds to compete with hemp (Low, 1995b). Weed suppression by hemp is the result of direct competition with weeds during the hemp growing season, however, and weeds will begin to grow again after the hemp crop is harvested. Residual weed suppressing compounds have not been reported in the soil following hemp production. "
In my experience Wikipedia is useful for non-controversial subjects.
Get into hotly disputed areas in the area of science and health (which is mostly what I’ve consulted it for) and its dependability is shaky.
It’s an adequate starting point for some research, and your implying otherwise reeks of an agenda, so there. 
There are probably many statements that can be made, this is my experience:
For some substantial percentage of topics and given a particular set of goals (it is goal dependent) it is the best (in terms of efficiency) starting point compared to any other I am aware of including a google search excluding wiki or the library.
IIRC that is one of the articles that I referenced in my edits and that were later removed.
And I agree with you entirely. It is sad when a few minutes on Google can return such high calibre references that say that hemp does require herbicide use, yet the article has boldly declared that no herbicide is required for many, many years. Worse yet is the fact that, when a change is made to reflect that fact that hemp does require herbicide use using that reference, it will be reverted almost instantly.
You are quite correct. This is a classic example of the unreliable nature of Wikipedia. An article on agriculture with facts that are easy to verify has been hijacked by pro-hemp idiots who have no idea about agriculture and includes numerous easily-proved inaccuracies that it is impossible to correct.