Talk:Anecdotal evidence

From Wikipedia, the free encyclopedia

This is the talk page for discussing improvements to the Anecdotal evidence article.
Please sign and date your posts by typing four tildes (`~~~~`). Put new text under old text. Click here to start a new topic. New to Wikipedia? Welcome! Ask questions, get answers. This is not a forum for general discussion about the article's subject.	Be polite Assume good faith No personal attacks Be welcoming	Article policies No original research Neutral point of view Verifiability

	This article is being improved by WikiProject Rational Skepticism. Wikiproject Rational Skepticism seeks to improve the quality of articles dealing with science, pseudosciences, pseudohistory and skepticism. Please feel free to help us improve this page. See Wikipedia:Contributing FAQ.
B	This article has been rated as B-Class on the Project's quality scale. (If you rated the article please give a short summary at comments to explain the ratings and/or to identify the strengths and weaknesses.)

[edit] a struggle for understanding

":The only time a sample size of 1 is statistically valid is when you're disproving a conclusion that claims to be correct for 100% of the population in question (which really only happens in logical arguments, not statistical arguments). So, if you're trying to make the argument that cigarettes don't cause cancer, the argument is flawed because the statistical evidence you're arguing from is too weak to support the conclusion. However, the exact same anecdotal evidence could be used to support a very strong argument that cigarettes don't always cause cancer, so to say that an argument is fallacious just because the evidence it's based on is anecdotal is clearly wrong. --Flatline 17:05, 3 January 2006 (UTC)"

All single statistic points are valid if gathered factually. The conclusion gained from the statistic is what is invalid

"gathered factually" and "statistically valid" are two very different things. For a single data point to be statistically valid, you have to be able to show what population the data point came from (which can be very difficult to do after the sample has been taken), demonstrate that the sample was randomly chosen from that population to prevent bias (impossible to show in the case of an anecdote), and, since your sample size is so small, you need to be able to make some sort of argument that the population in question is homogeneous (impossible to make without taking more samples). --Flatline 16:37, 14 January 2006 (UTC)

>Statistically valid" used that way, is not used on the evidence, it is used on the conclusion. The conclusion isn't statistically valid because the statistics are incomplete, not because the evidence is invalid. If a statistic is invalid it is incorrect. All statistics are valid unless incorrect. The statistics gathered may be valid but insufficient to produce any valid statistically conclusion. <

--Eric Norby 11:11(PST), 2006, January 17 (AD)

And as you say it can be valid if you are using the single datum to point out that not 100 percent of the points are inside some conclusion.

In that case, the datum is being used as a counter example, not as statistical evidence. It is statistically invalid. I was mistaken when I called it statistically valid in a previous comment. --Flatline 17:45, 15 January 2006 (UTC)

>The datum used in a counter example will only be invalid if it is a "flyer", a point that may have come from measurement or experimental error. That is determined at the time of the test/sample. If the statistic is undeniably correct, it is not a flyer, it is a valid statistic. Counter examples are used to invalidate a statistically incorrect conclusion< --Eric Norby 11:11(PST), 2006, January 17 (AD)

It is not the evidence that is invalid. It is the conclusion.

A person falling from a height of 2000 feet up didn't die. Does that mean that falling from great height doesn't lead to great death? No but it is valid that falls from great height don't kill everyone.

It is anecdotal to say that falls from great height don't kill.

I've never heard "anecdotal" used this way before. You have some anecdotal evidence that falls from great heights don't kill, but you could just as easily have anecdotal evidence that says falls from great heights do kill, so to apply the word "anecdotal" to the conclusion doesn't tell me anything about the correctness of the conclusion. I did a google search on "anecdotal conclusion" and found lots of examples of people using the phrase to mean "conclusion based on anecdotal evidence". Would I be correct if I said that you would claim that it means "conclusion that is verifiably false but supported by (carefully chosen) correct evidence that misrepresents the true state of things"? --Flatline 17:41, 15 January 2006 (UTC)

>My example was a rewrite of the cancer one. I'm going to try another rewrite of the current definition from what someone else has put forth. Anecdotal evidence should be short data that supports a point. Like anecdote means short story to support a point. It has come to mean short evidence that appears to support but doesn't a conclusion. That conclusion may be false. < --Eric Norby 11:11(PST), 2006, January 17 (AD)

Is it statistically valid that I dropped a hammer and it fell. Yes.

It is a fact that you dropped a hammer and it fell. It is meaningless to ask if it's statistically valid without first identifying "for what purpose?". There is a difference between a fact and a statistic; they are not the same thing. Similarly, there is a difference between scientific data and mere collection of facts (hence the cliche: "The plural of anecdote is not data"). --Flatline 17:41, 15 January 2006 (UTC)

>Exactly. Statistical validity is a term used on a conclusion, not on the statistics. There is no difference between statistics and facts. If taken using proper scientific tools both will be scientific facts. Using facts in statistical mathematics is using facts statistically. Statistics is a collection of scientific facts.

I suppose you have a point that there are "Lies, dam lies and statistics." ;)

What do you consider to be the definition of a statistic?< --Eric Norby 11:11(PST), 2006, January 17 (AD)

Is it wrong to say, "Therefore gravity must still exist." No. Does the single datum give conclusive evidence to that statement? Yes and no. It does support it, and isn't conclusive.

Being a single datum doesn't lead to statistically erroneous data. Using a single point in an attempt to support an incorrect conclusion, is using the data anecdotally, and a weak statistical conclusion. The datum is still correct and valid.

The statistical evidence isn't too weak to support the conclusion. The conclusion is weak because the evidence given doesn't fully support it. The evidence is just as correct as it always was. There is nothing wrong with the evidence.

Anecdote is a short story given to support a larger paper. Anecdotal evidence is supposed to be short evidence to support an anecdote. It has turned into, short evidence that doesn't support an anecdote.

"Charlie's grandmother lived to 95", is anecdotal if you use it to support the point that smoking doesn't cause cancer, however, it isn't anecdotal if used to support the point that not all smokers die of lung cancer.

For you information, that evidence doesn't support "that cigarettes don't always cause cancer", because that person may have had cancer but had not yet grown to significance, or if she had lived long enough she would have gotten it.

" :I agree that it is incorrect to say that anecdotal evidence is pseudoscientific. It would be more correct to say that anecdotal evidence is unscientific. However, generally speaking, it is correct to say that it is statistically unverified since if the anecdote were statistically verified, the argument would be based instead on the far stronger statistical evidence and not the anecdotal evidence. --Flatline 17:05, 3 January 2006 (UTC) "

How can a statistic be "statistically unverified". Is it not a statistic? Is it not verifiable? The evidence given is absolutely correct. The error is assuming it supports something that it doesn't.

" :You're correct: the conclusion is too strong for the evidence given. But this does not support your argument that anecdotal evidence necessarily supports fallacious conclusions. --Flatline 17:05, 3 January 2006 (UTC) "

But it does support my conclusion that the "support" is fallacious, therefore the conclusion is unwarranted.

It is a fallacy of support. The evidence is correct, the conclusion often is not, and the support is fraudulent. It is a scientific fallacy of fraudulent support, assuming logical support where none is apparent.

" :It is a false dichotomy to claim that data can't be both anecdotal and false. --Flatline 17:05, 3 January 2006 (UTC) " A false dichotomy is where two choices are given when there are three or more.

You should either call a fallacy by its most correct name or not. I don't see a valid reason for calling a fallacy by more than one name.

If you use false data for supporting an anecdote, it is much more correct to point out the falsehood of the data.

"Charlie's grandmother died at age 95....etc.", if that data were incorrect, it wouldn't support the anecdote. The anecdote would be left null and void. It would be far wiser to point out that Charlie's grandmother was diagnosed with cancer a week before she was hit by a car, and her health was poor a result of heart problems and emphysema.....and her lungs were found full of charcoal, ash, and tar, than it would to say, "that is anecdotal". False data doesn't make it to the anecdotal fallacy stage.

Sure it can be used incorrectly that way, but the best and most correct defense is derived from false data fallacy rather than the anecdotal fallacy.

The anecdotal fact must be indisputable by both parties, or the evidence will be under attack, the support lost from the beginning. Good data that appears to support an unsupported claim is what anecdotal evidence fallacy is all about.

I don't know Charlie's grandmother, I don't know what her health was, I've never heard of anyone smoking till 95 and being in good health. That is the correct defense, if the data given were "weak or poorly collected". If good data, the only defense is to enter wider more comprehensive data.

If you don't use that defense, the person giving the anecdote can say, 'The data refute it either, and it sure seems to refute the alternate claim.' Your point, that his point is wrong, because his data is incorrect, is lost.

" :Pseudoscientific data is unscientific data that is packaged to appear as scientific data. I don't see how this is relevant to the definition of anecdotal evidence.

Not anecdote.

It can be personal or drawn from worldly scientific study. It is the way the data was used not the data.

This is wrong. It is the origin (or lack of a verifiable origin) that determines if evidence is anecdotal. It has nothing to do with how the evidence is used. According to your definition, it would be possible to have two arguments based on the exact same evidence, but the evidence would only be anecdotal in the case of the fallacious argument. That's absurd! --Flatline 17:05, 3 January 2006 (UTC) "

Data is either scientific or false.

Packaging data to support a pseudoscientific claim, is called anecdotal, suppression of data, or false data, or unreliable data, depending on which flaw the pseudoscience is based on. Again, it is the claim that is pseudoscientific. Taking data pseudoscientifically produces false or unreliable data. Anecdotal evidence translates directly to short data or incomplete data.

The origin of a set of data determines its reliability not its anecdotal quality. If you measure gravity with poor equipment a calendar, you will get a poor value for "g". However, none of that information is anecdotal. If "Charlie's grandmother died at age 95...." the reliability is easily checked, but the use of it can be in anecdote or correctly. Note: It used to be that saying that proves not all smokers die of lung cancer was using the data anecdotally, but it has come to mean a scientific fallacy. The fallacy is one of support.

The strongest defense in that case is to show how the data is unreliable. It can still be an anecdotally incorrectly supported claim if the data is 100% accurate. The point is the correctness of the data raises the fallacy from poor data to another fallacy, that can be suppression of data, or unsupported anecdote.

The statement "anecdotal evidence" now implies the evidence is in error, but in reality, it used to imply that the evidence is short.

" :But not an orthogonal label. Any unverifiable anecdotal evidence would be considered poor data. Any anecdotal evidence, even readily verifiable, would be considered poor data if used to support a statistical conclusion since it is not framed in a meaningful statistical context. --Flatline 17:05, 3 January 2006 (UTC) " Again, it is the conclusion that is poor, not the evidence. If the evidence were poor, it would be false data.

" :Since when is it a fallacy to reference a relevant authority? Unless you have reason to doubt the authority of the reference or can demonstrate that there is no consensus of experts in the field, I am inclined to trust this reference. --Flatline 02:58, 5 January 2006 (UTC) " It is not a fallacy to reference a relevant authority. It is a fallacy to conclude that the reference to the relevant authority's conclusion, is proof that the conclusion is correct or even relevant.

You must present the authority's logic or data to prove your point. The fallacy was that you used his conclusion as a definitive statement. The debate here gains nothing. In fact I only have your word, here, that the "authorities" conclusion is relevant or a correct use of his words. I don't even know if he is the kind of authority you claim him to be. I also used that statement in punishment to the fact that you removed my earlier post.

Note: My earlier post just referenced your logical flaw by presenting logic and data. You are the one that chose to remove it.

Do a search on logical fallacies and look up Ipse dixit to get an alternate explanation to mine.

" :The data was incorrect. Your new article is similarly incorrect. --Flatline 02:58, 5 January 2006 (UTC) " The data was correct. My article is more correct than the one before it. It is evolving and improving. I'm thinking of adding the idea that anecdote is short story, hence, anecdotal evidence originally meant short data it has come to mean incomplete data, given to produce and erroneous conclusion.

" :I gave my reasoning on this very discussion page. In fact, you are responding to it now. --Flatline 02:58, 5 January 2006 (UTC) " If so I apologize.

It appeared to disappear, rather than become debated.

" :Please refrain from insults and name calling. It is inappropriate and distracts from the task of improving the article. --Flatline 02:58, 5 January 2006 (UTC) " Since you have now shown a willingness to respond to debate, I will delete the challenging personal parts of my posts.

--Eric Norby 09:57(PST), 2005, 29 December (AD)

[edit] The definition is incorrect.

Trying again.

Example: My grandmother smoked her whole life and lived to 96 without dying of lung cancer. That shows cigarettes don't cause cancer. Is an example of using anecdotal evidence. However, there is nothing scientifically nor statistically wrong with the data.

The only time a sample size of 1 is statistically valid is when you're disproving a conclusion that claims to be correct for 100% of the population in question (which really only happens in logical arguments, not statistical arguments). So, if you're trying to make the argument that cigarettes don't cause cancer, the argument is flawed because the statistical evidence you're arguing from is too weak to support the conclusion. However, the exact same anecdotal evidence could be used to support a very strong argument that cigarettes don't always cause cancer, so to say that an argument is fallacious just because the evidence it's based on is anecdotal is clearly wrong. --Flatline 17:05, 3 January 2006 (UTC)

So the definition that the data is often pseudoscientific or statistically unverified is incorrect. In fact, the evidence given can be shown to occur often, so the "singular" part of the definition not valid.

I agree that it is incorrect to say that anecdotal evidence is pseudoscientific. It would be more correct to say that anecdotal evidence is unscientific. However, generally speaking, it is correct to say that it is statistically unverified since if the anecdote were statistically verified, the argument would be based instead on the far stronger statistical evidence and not the anecdotal evidence. --Flatline 17:05, 3 January 2006 (UTC)

The fallacy is that the evidence doesn't support the conclusion. The conclusion that smoking doesn't cause cancer is erroneous. The correct statement would have been, Smoking doesn't always cause cancer many people including my grandmother have lived to 96 or so and not died of lung cancer while being smokers all their lives.

You're correct: the conclusion is too strong for the evidence given. But this does not support your argument that anecdotal evidence necessarily supports fallacious conclusions. --Flatline 17:05, 3 January 2006 (UTC)

Additional evidence showing that 90 percent of lung cancer cases are smokers relates the more relevant statistic. A large range of other experimental discoveries, and statistical, linking studies shows the mountain of proof that smoking is hard on your health. Science is about verifying things from a wide range of study, not one specific study, no matter how correct that study is.

At the worst anecdotal evidence is good solid data from a single study. All that was studied was those people living past age 90 that didn't die of cancer. Poor scientific conclusion. Perhaps poor study habits (Hand picking data, similar to data suppression but without intent.). But the data is absolutely correct and collected reliably.

If it weren't it would be false data.

It is a false dichotomy to claim that data can't be both anecdotal and false. --Flatline 17:05, 3 January 2006 (UTC)

If it weren't it would be pseudoscientific data.

Pseudoscientific data is unscientific data that is packged to appear as scientific data. I don't see how this is relevant to the definition of anecdotal evidence.

Not anecdote.

It can be personal or drawn from worldly scientific study. It is the way the data was used not the data.

Poor data has an entirely different label.

But not an orthogonal label. Any unverifiable anecdotal evidence would be considered poor data. Any anecdotal evidence, even readily verifiable, would be considered poor data if used to support a statistical conclusion since it is not framed in a meaningful statistical context. --Flatline 17:05, 3 January 2006 (UTC)

Anecdotal evidence is evidence used to falsely support a conclusion, given in an anecdote. The evidence given isn't wrong, the implied support is lacking.

--Eric Norby 2005 January 02, 10:10, (PST)

[edit] What is the difference between anecdotal evidence and witness evidence?

This is not explained. Please explain. 86.11.84.3 23:00, 2 July 2006 (UTC)

Ignore this comment as trolling. It has been well explained. An RFC is still pending for the above user, who has returned two months after failing to respond to it: see Wikipedia:Requests for comment/86.10.231.219. Tearlach 23:24, 2 July 2006 (UTC)

It is not explained. The above comment from Tearlach is aggressive, presumptious as to bad faith, unable to explain the difference and attempting to dismiss valid comment. 86.11.84.3 23:50, 2 July 2006 (UTC)

There is a law section in the article that begins "Witness testimony..." as well as a larger section on the analysis by Miller & Miller. These clearly explain the relationship, therefore I view the request to explain as disingenuous given your past edit history. Tearlach 08:10, 3 July 2006 (UTC)

Indeed, what is unexplained is the perfect defence which the poster promised, but did not deliver. I propose he should be blocked - a community ban rather than going through rfarb. Midgley 12:28, 3 July 2006 (UTC)

[edit] In television

In an episode of Dilbert, the cartoon show, Dogbert uses anecdotal evidence to persuade the public that they have work disabilities. Dilbert repetitively argues against the use of anecdotal evidence, but is proven wrong as Dogbert's evil plan unfurls and succeeds once again.

Sorry, as someone who has never (?) seen the show I don't understand this. Please rewrite this passage or preferably don't put it back into the article at all. <KF> 01:29, Dec 31, 2004 (UTC)

[edit] it does not follow that the conclusion is fallacious

This idea that anecdotal evidence only supports fallacious conclusions is wrong. Anecdotal evidence is simply evidence that comes from personal experience rather than designed scienfic experiments.

For example, if I know someone who was killed by a gun, then I can use that as anecdotal evidence that guns can kill people which is a verifiably correct conclusion even though it was arrived at through anecdotal evidence.

While it is true that anecdotal evidence is very weak evidence and is dangerous to depend on, this article goes too far when it claims that anecdotal evidence can only support an incorrect conclusion. --Flatline 17:29, 17 December 2005 (UTC)

The new article by Orzetto addresses my concerns. Thank you, Orzetto. --Flatline 14:32, 21 December 2005 (UTC)

[edit] Yes the conclusion is fallacious and the data correct.

The problems we are wrestling with on the definition of Anecdotal Evidence.

Truth: If the evidence is false, the fact that it is used in an anecdote is moot. The evidence is false, not anecdotal. It won't help to have poorly gathered data for the same reason, it is untrustworthy data, not anecdotal data. The data must be true or it supports nothing.

Personal: All evidence is observed from some personal position. It takes a person to observe data. Where the data comes from is unimportant to science. What is important to science is that it is verifiable and repeatable. The reason the data is often irrefutably true is because both sides of the anecdote often know the data is reported correctly from personal experiences. Drop a hammer on Earth and it will fall. It only needs to be true, not personal.

Singular: Often, but not always, users of anecdotal evidence quote more than one case. Therefore, singularity is not a required property of anecdotal evidence. Singularity is anecdotal evidence supporting the point that anecdotal evidence must be singular.

Anecdote: Evidence is science, the anecdote is what the evidence has been chosen to support.

Conclusion: The anecdote is the conclusion that the evidence was selected to support.

If the evidence supports a true conclusion, there is no surrounding evidence to prove the anecdotal nature. The conclusion then becomes a hypothesis or immature point, not an anecdote.

A scientific hypothesis takes data and attempts to fit a model to it. That data must be verifiable as is the data for an anecdote. The difference is, there is data that proves the anecdotal conclusion wrong, where the hypothesis remains a hypothesis until more data supports it or refutes it. If the hypothesis is claimed after refuting evidence is found, it becomes and anecdote supported by the earlier incomplete evidence.

If a conclusion is stated as verified by the data, the conclusion is immature. The cure for the immature conclusion is to point out additional evidence that could also explain the outcome.

Anecdotal evidence is an incorrect premise gained and supported by incomplete scientific data. The cure is to supply the complete data.

Premature conclusion is based on immature scientific data. Data gained without a full scientific investigation. Show additional possibilities.

The two are closely related but not the same.

The fallacy in anecdotal evidence is shown by the complete data.

The fallacy in the incomplete conclusion is shown by alternate explanations.

Smoking doesn't cause lung cancer, because many people have lived to 95 dying from other factors not relating to smoking or cancer, is anecdotal. The more complete evidence, that 90+ percent of the lung cancer cases were smokers, shows the fallacy. The anecdote exposed by more complete data, stops the necessity of more study.

Eating limes prevents scurvy, because Charlie ate them last voyage and didn't get scurvy, is a hypothesis or immature conclusion. There was no other supporting evidence to refute the claim. Being that it lacked falsifying data, more study was appropriate. Charlie ate a new thing he made and calls "aspirin" that voyage also, would lend addition avenues (Now both could be true.).

Quasars are powered, or not, by black holes, are premature conclusions. There is insufficient data to rule out either point. It is hard to know much about energy sources billions of light years away.

Hand picked data leads to false anecdotes.

Incomplete data leads to immature conclusions.

--Eric Norby 12:12, 26 December 2005 (UTC)

You seem to think that anecdotal evidence is some sort of logical fallacy and then go to great lengths to distinguish the "anecdotal evidence fallacy" from other logical fallacies. As such, the article you contributed, while containing some interesting points that should, perhaps, ultimately be included in the "final" article, is so heavily dependent on a misunderstanding that I decided to simply revert. Anecdotal evidence says nothing about the structure of the argument that uses it. Instead, it is a red flag that the premise might not be verifyable and that the argument using it might be relying on unsupportable generalizations (as my stats professor might say, "The plural of anecdote is not data"). Anecdotal evidence isn't a fallacy by itself, it's merely a catagory of evidence that hasn't had the rigorous safeguards of experimental evidence applied. Food for thought: should eye-witness accounts of crimes be considered anecdotal evidence? --Flatline 15:47, 27 December 2005 (UTC)

Okay, flipping through one of my philosophy textbooks from college (Invitation to Critical Thinking, by Joel Rudinow and Vincent Barry), in the chapter "Informal Fallacies of Evidence" they list a fallacy which they call the fallacy of suppressed evidence. From their discription, this fallacy matches the description that you've given for anecdotal evidence, except that in their description, the evidence can come from any kind of source (not just personal anecdotes). What's important to understand is that as a fallacy, this is a type of argument (as opposed to a type of evidence). Other fallacies that commonly use anecdotal evidence include: fallacy of small sample (where conclusions are drawn from a statistically insignificant set of samples), fallacy of unrepresentative sample (where samples are drawn from the wrong population), and (most obviously) unidentified experts ("I met doctor at the bar, and he said..."). --Flatline 17:49, 27 December 2005 (UTC)

"Okay, flipping through one of my philosophy textbooks from college (Invitation to Critical Thinking, by Joel Rudinow and Vincent Barry), in the chapter "Informal Fallacies of Evidence" they list a fallacy which they call the fallacy of suppressed evidence. From their discription, this fallacy matches the description that you've given for anecdotal evidence, except that in their description, the evidence can come from any kind of source (not just personal anecdotes)."

The above is a logical fallacy called "Ipse Dixit", "so his master says".

Since when is it a fallacy to reference a relevant authority? Unless you have reason to doubt the authority of the reference or can demonstrate that there is no consensus of experts in the field, I am inclined to trust this reference. --Flatline 02:58, 5 January 2006 (UTC)

Remeoved Statement

The data was incorrect. Your new article is similarly incorrect. --Flatline 02:58, 5 January 2006 (UTC)

Remeoved Statement

I gave my reasoning on this very discussion page. In fact, you are responding to it now. --Flatline 02:58, 5 January 2006 (UTC)

Remeoved Statement

There is no reason for "Anecdotal Evidence" to be restricted to "personal anecdote". The better the data given the more it appears to support a fallacious anecdote. It is the logical leap to the conclusion where the anecdote errors, not in the data. The data is too incomplete to verify that conclusion.

If we accept your definition, then it is trivial to find argument pairs such that the data is anecdotal in one argument and not in the other. How can this be acceptable? Either the data is anecdotal, or it is not. The context is irrelevant; only the origin matters when determining if said evidence is anecdotal or not. --Flatline 02:58, 5 January 2006 (UTC)

Remeoved Statement

--Eric Norby 09:57, 2005, 29 December (UTC)

Please refrain from insults and name calling. It is inappropriate and distracts from the task of improving the article. --Flatline 02:58, 5 January 2006 (UTC)

[edit] Rewrite

I had a go at a rewrite. The previous version, apart from being a hatchet job on those who use the term, contained a number of errors: particularly the claim that it isn't used in legal discourse. The idea that courts have some mechanism superior to science for checking witness testimony isn't true either: a search on "cognitive bias" + eyewitness will find various ongoing research that suggests the opposite. Courts still place a lot of reliance on unreliable cues such as witness confidence and detail of recollection. Tearlach 02:28, 19 January 2006 (UTC)

There is no category of evidence known in law as "anecdotal". There is witness testimony and there is hearsay evidence, but not anecdotal. It was not claimed that the term "anecdotal" is never used in Court. It very likely is by medical expert witnesses for example, but it is not a category of evidence in law. Accordingly the above claim is wrong on that count.

There is no claim that law has superior or inferior methods to science. Law does however have methods for testing oral witness testimony which neither science nor medicine have or have developed. Again, therefore, your claims are incorrect. The statement that courts "place a lot of reliance on unreliable cues" indicates a misunderstanding of the function of a court. A court has an obligation to make a decision on the basis of the evidence before it, regardless of the fact that evidence from either side might be difficult to assess for reliability or might be of doubtful reliability. Further, by employing methods of testing evidence (which science and medicine lack) the reliability of evidence can be assessed. A court might decide a case by saying with reasons why one witnesses' evidence is "preferred" to another's if there is difficulty assessing the reliability of conflicting claims because of a lack of other sources to confirm or refute. Alternatively, there may be copious other sources to enable oral evidence to be assessed as thoroughly reliable or not.

Accordingly, it is clear from the foregoing that the prior edit was not a "hatchet job on those who use the term" but an accurate account of the position. Further it demonstrates that the "rewrite" is a POV job by someone who does not like what is said because it does not sit easily with his point of view. Additionally, User Tearlach pops up elsewhere with others who appear to be in a clique and engage in edit warrring with others whose views he and his colleagues disagree with.

Further, the use of "anecdotal" in relation to oral witness testimony is incorrect and misleading and that is a further reason why this "rewrite" is misleading.

Hence this misleading rewrite if left as it is will be yet another example of the kind of work which damages the reputation and value of Wikipedia and thereby leaves a big questionmark over any pretence there may be of being a reliable source of information. 86.10.231.219 07:26, 21 January 2006 (UTC)

Then you'd better tell Sprenger and Lang, Attorneys, and the NZ Ministry of Economic Development [1] [2]. that there's no such thing. The City of Phoenix Second Generation Disparity Study Final Report has a whole chapter on it. Tearlach 20:13, 21 January 2006 (UTC)

All citations have nothing to do with legal evidence in court. These reports are typical examples of the inappropriate use of "anecdotal evidence", and are prepared by firms of consultants. It is cringing to see this use of such a term in relation to hearsay evidence presented in a form of market survey/research. Where market research/survey evidence is produced in court it is highly unlikely that is will be referred to as "anecdotal" by the lawyers or judge and nor will the hearsay evidence be described as such. This really does make my point that medical, scientific and other categories of individuals simply do not understand evidence and that the use of the term "anecdotal evidence" is inappropriate. It is just not a category of evidence known to law. I cannot legislate for terms Sprenger and Lang, Attorneys want to use on their website. That is up to them. 86.10.231.219 22:52, 21 January 2006 (UTC)

Oh, who am I to believe on the usage of a legal term? Published references from a firm of Washington attorneys, a national government department, and a city survey by a firm of consultants backed by a major Kansas City attorney? Or an unsupported anonymous opinion? When you're in a pit, stop digging. Tearlach 00:48, 22 January 2006 (UTC)

Good advice, which you would be wise to follow. The "who am I to believe" part acknowledges a certain lack of familiarity with the subject-matter which might explain your use of irrelevant "published references" from "digging" on the web. Let me help you.

You will not find a reference in any standard legal textbook on evidence confirming that "anecdotal evidence" is legal term to describe a category of legal evidence. You won't find one because it is not. Whilst there is no scarcity of legal cases in which the phrase "anecdotal evidence" appears, it is not used as a legal term to describe a legal category of evidence and you will more likely than not find it in quotes - "anecdotal" evidence. Anecdotal evidence is not a category of evidence known to law.

Medical and scientific professionals use the term indiscriminately to refer to anecdotes of the kind "I met this guy who told me ..." and to witness testimony. The moment a story such as "I met this guy who told me ...." is committed to paper with facts and details to make it a verifiable account, that is witness testimony. It was also witness testimony before it was committed to paper. So when a patient gives an account of symptoms, that is witness evidence.

Courts don't deal with anecdotes or anecdotal evidence. They do deal in witness testimony, but medical and scientific professionals and the lay public use the term indiscriminately. That is why others have tried to get this page into shape and have failed, because the "anecdote" and witness testimony are the same thing - something a witness says. The only question is whether it is reliable. To find that out, lawyers document the evidence and the legal process tests it in court for reliability. Medical and scientific professionals do not make the distinction - they have a tendency to call all witness evidence "anecdotal".

Surely wouldn't you have realised all of this from the text of this page before you deleted wholesale chunks of what was a correct account of the matter? Did you not realise that? So why did you chop it all up? You carefully did not quote what the page said previously, stating it said the term "anecdotal evidence" ... "isn't used in legal discourse". That was never said nor claimed. Please convince me that I am wrong - are these characteristics of an edit war by a POV warrior or might I be wrong on that? 86.10.231.219 03:57, 22 January 2006 (UTC)

Minor vandalism deleted. The Invisible Anon 10:15, 2 February 2006 (UTC)

Comment restored in edited form I'm sorry about the previous phraseology: however, expressing ideas that 86.10.231.219 dislikes is not vandalism. In fact its removal suggests it hit the nail on the head.

I suspect an agenda here. Claimed proof of efficacy of alternative therapies depends heavily on anecdotal evidence, which has easily analysed weaknesses such as the cognitive bias that applies to all unstructured human observation and recollection. Therefore believers in such therapies might well be expected to spin the term as being ill-defined, and to bury clear description of its characteristics and problems in a mess of obfuscation. 213.130.141.76 13:44, 2 February 2006 (UTC)

Thanks for explaining. I thought it was someone dropping something in to look like reasoned comment that was not. What you say is now clearer.

No clever agenda here. I am not saying that unreliable evidence is to be taken as reliable. I am just making clear that

) anecdotal evidence and witness evidence are/can often be the same thing
) what people mean by anecdotal is meant to be "unreliable" witness evidence - often not collected in any formal way or documented or verified and
) the term "anecdotal evidence" is now used in certain fields and casually to describe witness evidence regardless of whether it is capable of or has been documented and tested for reliability.

There is nothing here to say that plainly unreliable evidence should be relied on. Proof of anything requires reliable evidence. Just because someone claims an alternative remedy cured them, does not mean that is proof it did.

The Invisible Anon 14:38, 2 February 2006 (UTC)

No clever agenda here

I'd believe that if you hadn't replaced the article with unwikified waffle, expunging all reference to issues such as cognitive bias, and framed all the external links with biased captions. 213.130.141.22 02:21, 3 February 2006 (UTC)

I see you have reverted the entire page and provide as justification a brief POV indicating you do not agree. That is perhaps not quite the constructive approach that is generally appreciated. So, why not explain exactly what is troubling you? It is not easy to see from what you have said and you seem to be angry about something. All I can tell so far is that you feel that cognitive bias should be mentioned and in the context of debunking so-called "alternative remedies". There is nothing incompatible with the page as set out before you reverted it and your POV. The Invisible Anon 06:15, 3 February 2006 (UTC)

[edit] Case reports?

The case report is an invaluable tool in medicine, and it is still anecdotal evidence Vandenbroucke JP. In defense of case reports and case series. Ann Intern Med 2001;134(4):330-4. PMID 11182844. Worth linking to. JFW | T@lk 23:39, 21 January 2006 (UTC)

A useful paper indeed. I have noted it. Thank you for bringing it to our attention. Whilst not useful to the discussion of "anecedotal" in law it usefully builds on work others have done in medicine. 86.10.231.219 00:06, 22 January 2006 (UTC)

The intro says "Anecdotal evidence is a term used in medical [...] discourse". It is very useful for this article. JFW | T@lk 08:41, 22 January 2006 (UTC)

The intro to the cited paper does not mention anecdotal evidence. It is User Tearlach's introduction to this page that states "Anecdotal evidence is a term used in medical [....] discourse". That is not in a matter in issue. It is well accepted that science and medicine use this term. What is in issue is that the term is frequently used in science and medicine in relation to any kind of witness evidence when that is inappropriate as demonstrated by legal practice in relation to witness evidence.

What is helpful about this paper is that this is a medical paper which confirms that the often blanket approach in medicine and science of describing witness evidence as "anecdotal" and not accepting it as a category of evidence that can be relied on is inappropriate. This also supports the prior correct text of this Wiki page which User Tearlach deleted wholesale and replaced with wholly incorrect text. Also, this referenced paper cites the use of "anecdotal" evidence twice but not in its own context but from other papers and in the medical and not legal context.

There is a lack of clarity in medicine and science on the matter and a degree of double-think. Witness evidence of the "market survey" type is sometimes used as the basis for a paper in epidemiology. In law "market survey" type evidence from a researcher is often treated with great circumspection. The evidence cannot be tested - for example, the responses of the participants cannot be independently tested nor can their reactions to and understanding of the questions.

I hope that assists to clarify for you. 86.10.231.219 12:50, 22 January 2006 (UTC)

Something about the role of the medical case report would be worth adding to the article. Tearlach 12:45, 22 January 2006 (UTC)

The term "anecdotal evidence" has a wider scope than 86.10.231.219 is suggesting. JFW | T@lk 19:10, 23 January 2006 (UTC)

Exactly the point. To some in medicine and science it includes everything a person says in giving an account of personal experiences. That is a wide scope. If you have another view of that or of what you believe I am saying, I will be pleased to hear it. 86.10.231.219 23:05, 23 January 2006 (UTC)

I don't think so. Anecdotal evidence always has an element of cherrypicking. That is, it is used to support claims about the frequency of occurrence of something, but the "computation" of the frequency is slanted because only one type of occurrence is used in the anecdotes (for example, only stories about people who got better after taking X, but not stories about people who got worse or stayed the same, and no stories about people who didn't take X). ~~I never heard the term applied to "personal experiences" without that element.~~ --Hob Gadling 10:25, 31 January 2006 (UTC)

OK, the SkepDic link "Anecdotal (testimonial) evidence" does exactly that - they confuse two things that don't belong together. But the other link does not. "Anecdotal Evidence". --Hob Gadling 11:11, 31 January 2006 (UTC)

Take care. It is easy to merge or flow in thinking from one evidence concept to another. The cherrypicking point goes to the use of the evidence and its reliability as proof. Evidence is just information. It does not matter what source it comes from. The next issue is whether the information is reliable - is it authentic, is it accurate and is it complete (the whole story). So you test it.

The one factor that differentiates what is normally considered witness testimony from what is often considered anecdote is the lack of formal documentation of the evidence. The point is that it is all witness evidence whether as informal "anecdote" or formally documented testimony. For example the Don Lindsay link states "Suppose I tell you that I thought of a long-lost friend. Just then the phone rang, and it was my old friend. That's a nice anecdote." That is also witness evidence. Whether or not it proves anything depends on the proposition being put and the other evidence available. It would support a proposition that the old friend telephoned or that the old friend either has or had access to a telephone at that particular time. It tends to support the proposition that precognition exists, but alone it does not prove it exists nor does it prove this was anything other than a co-incidence. If such a thing never ever happened, that would support a proposition that precognition does not exist, but again it does not prove it does not. More evidence is needed, and then you have to decide to what standard you want to prove something.

Evidence is a tough topic for many people. It can be just as tough as some topics in philosophy. It shares a great deal in common with the philosophy of knowledge.

This page is dealing with what "anecdotal evidence" is and proof and standards of proof and other kinds of evidence are all part of the matrix of Wikipedia pages dealing with this complex topic. It is all too easy to flow in your thinking from one discrete concept to another. That is one of the reasons people have found it so tough to get this page straight.

I do hope that assists and I also hope it does not come across as patronising. There is a great risk it will and it does not imply I underestimate your ability or overestimate mine. I just want to be as clear as I can on something I know it is far too easy to confuse. The Invisible Anon 02:31, 1 February 2006 (UTC)

[edit] Request for Comments

There has been edit war between this version and this version. [3]

Please place your RfC discussion below. JustinWick 20:16, 6 February 2006 (UTC)

HOAX RfC WARNING - DISPUTED
NOTE: Suggest delete this RfC and a block on 213.130.141.147 - This RfC was posted by an anon 
16:24, 4 February 2006 213.130.141.147 whose only contribution to Wikipedia 
is this RfC [[4]]. 
See also the anon's "contribution" [[5]] from a very similar IP address 213.130.141.72 
(emphasis added) "Proof of efficacy of woo-woo therapies depends heavily on anecdotal 
evidence" 
Anon clearly knows his/her way around Wiki procedures. Believe know who this is - unsuccessful 
edit warrior from other pages.
The Invisible Anon 08:47, 7 February 2006 (UTC)

THE "HOAX" STATUS OF THIS RfC HAS BEEN CHALLENGED BELOW - JustinWick 20:14, 7 February 2006 (UTC)

And CHALLENGE ANSWERED below The Invisible Anon 23:22, 7 February 2006 (UTC)

RfC - I believe this version is better as it contains a references section. I'm a curious as to the existence of any trustworthy sources that assert that the concept of "Anecdotal Evidence" is contraversial. - JustinWick 20:16, 6 February 2006 (UTC)

RfC - The "no universal common usage of or accepted general meaning for Anecdotal evidence" claim should remain if authoritative citations are given to prove this (e.g. quoting valid sources that have conflicting definitions). --Wade A. Tisthammer 20:32, 6 February 2006 (UTC):

see my new comments below (diff [[6]]) setting out the conflicting definitions which show there is "no universal common usage of or accepted general meaning" The Invisible Anon 14:57, 7 February 2006 (UTC)

The differences seem relatively minor. Take for instance these quotes:

"Report of clinical experiences based in individual cases, rather than an organised investigation with appropriate controls, etc."
"Based on casual observations or indications rather than rigorous or scientific analysis"

There are some differences, but the overwhelmingly common theme throughout seems to be individual testimony not backed with rigorous/careful/scientific study (as of the control variables). Perhaps something like a general definition, with the caveat that there are "somewhat different definitions" followed by a few cited quotes (e.g. one mentions clinical experiences and one does not--but perhaps that is because one is defined in the medical context). --Wade A. Tisthammer 19:10, 11 February 2006 (UTC)

Nail on head. I agree 100%, in science & medicine that is the meaning. Consequences not popular - hence this RfC because:-

it means science and medicine dismiss all witness evidence as "anecdotal" (see example of expert witness evidence from a consultant paediatrician cited here [[7]])
witness evidence/testimony is the primary evidence courts admit
others take this as criticism of science and medicine's approach to evidence
they also believe it admits claims that alternative therapies work on the basis of "witness evidence", which they dismiss as "anecdotal evidence"

That is why we have this long debate - no other reason. But it is a very important aspect of the use of the term, as can be seen from the debate here. The Invisible Anon 20:49, 11 February 2006 (UTC)

RfC - The hoax warning on the RfC page worries me. There is a genuine controversy here and the anonymous poster presented it fairly. It might have been a simple oversight, such as forgetting to log in manually after cleaning out one's cookies file. The call to block the IP was an inappropriate suggesion. That said, I'd go with the shorter version of the article until someone contributes a Wikified and referenced expansion. At an article about evidence this type of comment shouldn't be necessary. Durova 10:45, 7 February 2006 (UTC)

if the RfC is not a hoax how does one explain the "woo woo therapies" posting from the almost identical IP.
if there is an "edit war" there does not seem to be much evidence of that
and if there were, what is it about - what are the points in contention?

here are the links to the only dialogues on these versions [[8]] [[9]]

4 the new text covers everything the old text did

5 if there were clear points in contention, then it would be possible to edit to improve the text

6 what are the points in contention that the "edit war" (?) is meant to be about?

7 premature for an RfC? Here it is anyhow - so it would be nice to know what are the issues?

The Invisible Anon 14:15, 7 February 2006 (UTC)

Follow-up The post at RfC says, Talk:Anecdotal evidence - edit war between this version and this version. Thoughts? 16:24, 4 February 2006 (UTC) That looks like a reasonable request. The history page does show reverts over this disagreement and the talk page does discuss this content at length. Neither of these two IP addresses is flagged as a variable AOL IP. In fact neither has any comments on the user talk page, which is the proper place to begin when one suspects vandalism and inappropriate behavior. The "hoax RFC warning" has the appearance of an attempt at poisoning the well. If I stumbled across this page I'd probably flag it. It reads like someone's personal essay. Durova 18:11, 7 February 2006 (UTC)

In Answer Re Hoax No "well poisoning" here. Durova might be kind enough to look again at the talk page and also note there has been just one revert of the prior draft of current text [[10]] and the editor did not take up the offer to discuss why [[11]].

The same anon has been using three different IP addresses on the talk page. These are linked by the comments made and there is the appearance of vandalism (see "woo woo therapies" below). These IP addresses all resolve to 213.130.141.xx and the fourth IP used to make the RfC is 213.130.141.147 which also resolves to 213.130.141.xx.

213.130.141.72 - see the "Proof of efficacy of woo-woo therapies" and "Therefore believers in woo-woo therapies" comments here [[12]]
213.130.141.76 - this confirms by same author [[13]] - "I'm sorry about the previous phraseology:" and this comment "I suspect an agenda here." links this author to the next comment at a third IP address here:-
213.130.141.22 - where the same author continues [[14]] the "agenda" theme with "::No clever agenda here ... I'd believe that if ..."

Ergo, all are linked IPs and this is someone who is very familiar with Wikipedia procedures.

The Invisible Anon 23:22, 7 February 2006 (UTC)

I fail to see how three posts to the talk page constitute vandalism of any sort. The woo-woo therapies quote is taken out of context. The IP uses the phrase to mean pseudoscientific medical treatment in a discussion of anecdotal evidence. All of these posts constitute topical debate. I suggest mediation. Durova 05:29, 8 February 2006 (UTC)

I concurr. AFAIK, as long as someone considers an RfC necessary (and attracts "RfCers"), it stands. I think there is a clear dispute here and there is adequate attention on the issue. Anonymous (Ironic?) Editor seems to have vested interest in stopping RfC proceeding - this is suspicious IMHO. My prior comments stand unchanged. - JustinWick 23:49, 8 February 2006 (UTC)

this is suspicious IMHO

I agree. Filibustering is the word that springs to mind. Going back to Wikipedia basics: is the current version encyclopedic? When you open Encyclopedia Britannica, do you see something looking like that? Of course not. Nor will you see it said anywhere that the term has no accepted general meaning, which makes the article - devoted as it is to proving this - original research. Tearlach 21:13, 9 February 2006 (UTC)

In reply All of the definitions are at odds with each other - that there is no accepted general meaning is demonstrated - the evidence is produced - the critics answered. Answering is not Filibustering. I should be praised for directly answering and dealing with the issues and writing an excellent page here in Wiki which deals clearly and thoroughly with a difficult topic others have struggled with. With respect "When you open Encyclopedia Britannica, do you see something looking like that?" is not exactly a specific comment directed at fixing anything. What is your issue?

Here is a comment on just one example of Filibustering of a great many by another editor [[15]]. Now that is what I call filibustering. That editor Midgley whom you are very familiar with is also writing original research over multiple pages with no citations or sources. Plenty more examples from the same editor on the same page. Yet you do not complain about him. I look forward to seeing you tackle all the examples of those contraventions of Wiki policy. The Invisible Anon 08:09, 10 February 2006 (UTC)

All of the definitions are at odds with each other - that there is no accepted general meaning is demonstrated - the evidence is produced

Your supposed evidence isn't demonstrating anything. I hope we agree that there's general agreement as to what a lemon is. Yet the dictionary definitions are all different, just different takes on describing the same thing. This is true for probably all words. Any attempt at consensus or specific attempt at improving the article is pointless unless you drop the mistaken assumption that such differences mean something. Tearlach 20:16, 10 February 2006 (UTC)

1a. A spiny, Asian evergreen tree (Citrus limon) widely cultivated for its yellow, egg-shaped fruit. b. The fruit of this tree, having a yellow aromatic rind and juicy, acid pulp. 2. Lemon yellow. 3. Informal One that is unsatisfactory or defective: The American Heritage® Dictionary of the English Language: Fourth Edition. 2000

1. A pale yellow oval citrus fruit with acid pulp. 2. The tree that bears it. 3.A pale yellow colour. 4.Something disappointing, worthless, unattractive, unpleasant (slang) Chambers Twentieth Century Dictionary

1. a pale yellow oval citrus fruit with thick skin and fragrant, acidic juice; a drink made from or flavoured with lemon juice. 2.the evergreen citrus tree which produces lemons. 3.a pale yellow colour. 4. (informal) an unsatisfactory person or thing. Concise Oxford English Dictionary

In reply

Sigh.

I produce several "definitions" & examples [[16]] each defining "anecdotal evidence" as something different each time.
I carefully show that is the case [[17]]
Tearlach produces three definitions [[18]] which all describe the same thing using only slightly different terms to do so.

That is comparing apples and orangutans. We all know what a lemon is. We don't all know what "anecdotal evidence" is - as demonstrated. And I get accused of filibustering? There's no justice in the world.

The Invisible Anon 21:01, 10 February 2006 (UTC)

Don't be daft. All of these definitions of "anecdotal evidence" are within the normal range of variation you'd expect for different sources describing the same thing. If it were that confused or complex, other references would say so - and they don't. Here's an example:

anecdotal evidence n. Evidence obtained informally from isolated observations rather than from systematic investigation, seldom collected with sufficient care or reported in sufficient detail to be trusted as a basis for generalization, but often a source of hypotheses for further investigation. - A Dictionary of Psychology. Andrew M. Colman. Oxford University Press, 2001.

It does not take 1500 words to explain it. Tearlach 02:23, 11 February 2006 (UTC)

In further reply

They clearly do not describe the same thing.

Let us see if we can get to the heart of this, notwithstanding it is testing my patience to a degree beyond measure:-

1. Do you or do you not agree that (taking just two of the conflicting definitions) [[19]]:-

"'Report of clinical experiences based on individual cases" is by definition

based on facts observed by a skilled clinician
who is duty bound by profession and by law to carefully study the patient in clinic who then documents those observations in the patient's notes for future reference

but this conflicting definition "describes information that is not based on facts or careful study"

does not cover such a careful study in clinic by a skilled clinician
does not cover facts observed of a patient's condition
does not cover observations documented in the patient's notes for future reference

2. Taking from your psychology dictionary definition "seldom collected with sufficient care or reported in sufficient detail to be trusted" do you or do you not agree that

those words are not apt to describe the clinical records made by consultant paediatricians of patients seen and treated by them in clinic?
if a consultant paediatrician referring to his records says "I have treated 527 cases of Otitis media" that is untrustworthy and anecdotal and is still "anecdotal evidence" even if given as witness evidence used in Court under oath?

I am coming to the conclusion that you honestly do not realise all these definitions are completely at odds with each other because I am not allowed to presume that this is now simply a matter of not wanting to lose face in the teeth of the facts. Either way this "dialogue" has become incredibly "one-way" and I am now spoon-feeding the points. I believe I am now being forced to explain to a degree that any reasonable honest and impartial observer will be duty bound to agree is enough. However, I wait to see your response ..... The Invisible Anon 06:23, 11 February 2006 (UTC)

RFC - FX: sigh. I was going to leave this page alone, but as the original poster of the shorter version, I will comment. I think the longer version isn't very informative: it seems to be quibbling about definitions at the expense of a clear concise picture of the topic. I'm also worried about the hoax claim and block request. There clearly is the beginnings of an edit war, and the RFC statement is as neutral as it gets. I don't buy the argument, expressed at length below, that the variation in dicdefs means there's "no universal common usage of or accepted general meaning". Dicdefs of all words and terms show similar variation. Tearlach 12:54, 7 February 2006 (UTC)

The definitions are all different. There is no one common definition - compare these from defns set out here [[20]]:-

"non-scientific observations or studies, which do not provide proof but may assist research efforts"

this foregoing definition of "anecdotal evidence" (full defn below) excludes evidence other than "scientific" even if from careful study:-

"Report of clinical experiences based in individual cases"

is a definition clearly of the "witness testimony" type such as evidence of a consultant clinician of his/her experience treating numerous patients

"describes information that is not based on facts or careful study"

this definition of "anecdotal evidence" is to evidence neither based on facts or study and could be anything

"from personal interviews, public hearings, and surveys." and "from the sense of anecdote `unpublished narratives or details of history'"".

again covers very different categories of evidence

"Information passed along by word-of-mouth but not documented scientifically"

this covers only oral information

The Invisible Anon 14:57, 7 February 2006 (UTC)

RfC - On 16 Dec 2005 an edit was made that completely changed the tone of the article. The motivation for that is unclear, as nothing is in the talk page on the reason Eric Norby made the edits. Quickly, however, uncivil bahavior erupted. It appears that this article is being used as a place for surrogate arguments related to anecdotal evidence. The article up to 12 Dec 2005 was written in NPOV style, though not enough emphasis was given to be utility of anecdotal evidence in hypothesis generation. Currently there has been a shift to legalese. This would be appropriate in an article on evidence in a courtroom setting, but not a general article on what is meant by anecdotal evidence. The arguments about there being different dictionary definitions of a word and thus the term not having meaning is specious. Most words in a dicitonary have several definitions. That merely shows that words have nuances of usage or even different meanings in different contexts. The use of anecdotal evidence was very clear from the initial verisons of the article and did represent the most common usage of the term. Kd4ttc 17:23, 11 February 2006 (UTC)

[edit] RFC - Some Definitions of Anecdotal Evidence - showing contrasting conflicting approaches

Preamble - The current text of this Wiki page brings together in one definition all of the different kinds of use of the term "anecdotal evidence" and provides a comprehensive and accurate definition of how the term is used. This page is solely concerned with what "anecdotal evidence" is in all its forms and meanings. This page does not deal with concepts of proof or reliability of evidence. See the Wiki topic evidence and it will be seen that it is made up of numerous discrete elements spread across numerous pages each dealing with an element of evidence. Evidence is a complex subject and it can be seen from the talk page and from the different versions of the page that some editors have been inadvertently combining and confusing several different evidential concepts and bringing them to this page (in innocence but in error).

Definitions/Interpretations The following are a smattering of examples - there are many conflicting definitions and descriptions of the term.

"anecdotal" "Report of clinical experiences based in individual cases, rather than an organised  
investigation with appropriate controls, etc." Origin: G. Anekdota, unpublished items, fr. 
An-priv + ekidomi, to publish - [[21]]

"anecdotal" adjective - describes information that is not based on facts or careful study:
anecdotal evidence [[22]]

"1 entry found for anecdotal evidence.
Main Entry:     anecdotal evidence
Part of Speech:         noun
Definition:     non-scientific observations or studies, which do not provide proof but may assist    
research efforts
Example:        This chapter provides anecdotal evidence from personal interviews, public hearings, 
and surveys.
Etymology:      from the sense of anecdote `unpublished narratives or details of history'"
Source: Webster's New Millennium™ Dictionary of English, Preview Edition (v 0.9.6)
Copyright © 2003-2005 Lexico Publishing Group, LLC

"2 definitions found for anecdotal
From The Collaborative International Dictionary of English v.0.48 :
 Anecdotal \An"ec*do`tal\, a.
    Pertaining to, or abounding with, anecdotes; as, anecdotal
    conversation.
    [1913 Webster]

From WordNet (r) 2.0 :
 anecdotal
      adj 1: having the character of an anecdote; "anecdotal evidence"
      2: characterized by or given to telling anecdotes; "anecdotal
         conversation"; "an anectodal history of jazz"; "he was at
         his anecdotic best" [syn: anecdotic, anecdotical]"

"One entry found for anecdotal.
Main Entry: an·ec·dot·al
Pronunciation: "a-nik-'dO-t&l
Function: adjective
1 a : of, relating to, or consisting of anecdotes <an anecdotal biography> b : ANECDOTIC 2 <my  
anecdotal uncle>
2 : based on or consisting of reports or observations of usually unscientific observers <anecdotal 
evidence>
3 : of, relating to, or being the depiction of a scene suggesting a story <anecdotal painting> 
<anecdotal detail>
- an·ec·dot·al·ly /-t&l-E/ adverb"
[[23]]

an·ec·dot·al adj.
  1. also an·ec·dot·ic [ -dtk ] or an·ec·dot·i·cal [ --kl ] Of, characterized by, or full of 
     anecdotes.
  2. Based on casual observations or indications rather than rigorous or scientific analysis: 
     "There are anecdotal reports of children poisoned by hot dogs roasted over a fire of the 
     [oleander] stems" (C. Claiborne Ray).
  [[24]]

Definitions of Anecdotal Evidence on the Web:

  * Information passed along by word-of-mouth but not documented scientifically.
     www.nechakowhitesturgeon.org/sturgeon/glossary/index.php
  * Anecdotal evidence is unreliable evidence based on personal experience that has not been 
    empirically tested, and which is often used in an argument as if it had been scientifically or 
    statistically proven. The person using anecdotal evidence may or may not be aware of the fact 
    that, by doing so, they are generalizing.
    en.wikipedia.org/wiki/Anecdotal_evidence

Anecdotal Evidence: Anecdotal Evidence is information you obtain from a subjective report, an 
observation, or some kind of example that may or may not be reliable. In addition, anecdotal 
evidence is not scientifically valid or representative of a larger group or of conditions in 
another location.
(alleydog link removed 10 December '06 because WP identified it as spam and wouldn't let me save edit otherwise - --Calair 12:30, 10 December 2006 (UTC))

See this helpful webpage which exemplifies further contrasting approaches to "anecdotal evidence" [[25]] The Invisible Anon 23:18, 6 February 2006 (UTC)

[edit] Anecdote and Narrative

Narratives are used to make points, and since we all retain an internal model of the world, and can model another person's mind as well, narrative has a subjective reality to us. Anecdotes represent a transfer of soem piece of narrative, which we can test for consistency against those models.

The statistical examples seem to hve ignored the Rev. Bayes' work, and also the n-of 1 trial which is a statistically sound evidence based medicine method. Since it relies on repeated trials the applicability to 666 m falls may be limited by volunteers.

Overall, this has the apeparance of a silly argument, given that we all know what anecdotes are even before we reach our anecdotage. The Invisible Anon 14:33, 27 February 2006 (UTC)

[edit] Cleanup

There are a few points of cleanup that need to happen.

No lead-in section
Large amounts of preformatted text
No references

joshbuddy^talk 15:09, 1 March 2006 (UTC)

[edit] Comment

I count dialogue to the RfC between two versions as: 4 - older version 1 - uncertain 1 - newer version

If this were a poll that would constitute the 2/3 supermajority to return to the older version. Durova 04:17, 3 March 2006 (UTC)

Improvement is the appropriate course. Reversion is less appropriate for a number of reasons.

Further, the main issue seemed to be disagreement over whether there was an absence of a generally accepted definition of "anecdotal evidence". Subsequent to the comments Durova suggests be treated as "votes", references were provided to demonstrate this in this talk page [[27]] and in the project page [[28]]. Consensus was achieved [[29]] with the comments of Wade A. Tisthammer posted 19:10, 11 February 2006 (UTC). The RfC was also converted to a "factual dispute" by its propounder.

Additionally, irrespective of the issue of definition, the issues regarding a comparison with legal evidence - which relies substantially on the oral testimony of witnesses - with the scientific and medical approach - which discounts it as "anecdotal" should remain. :^Talk - The Original Invisible Anon at 86.10.231.219

Respectfully, since this comes from the author and only supporter of the newer version, it's hard to regard that opinion as definitive. The older article needed minor editing. The newer version is nonstandard format violates WP:NOR and WP:NPOV. Several editors raised these points weeks ago and politely suggested improvements. Nothing has improved.

I have only two interests here: I want Wikipedia to be a quality encyclopedia and I want to help editors work harmoniously. It becomes impossible to compromise when one editor insists that the only demonstration of good faith is wholehearted agreement and interpolates improper conduct into any dissent. I strongly advise "The Invisible Anon" to browse some ongoing dialogues at Wikipedia:Requests for comment/User conduct and Wikipedia:Requests for arbitration. Per Wikipedia:Words of wisdom, "When you start accusing everyone of being in on a conspiracy, you shouldn't be surprised if they decide to confirm your paranoia by banding together against you." I hope this doesn't go there. One way to get on a better track is to look for grains of truth in other people, to suppose they have honest motives and know useful things. Regards, Durova 20:19, 3 March 2006 (UTC)

[edit] Organizing this page

It appears that someone with a very specific POV has been involved in editing this previously. I have tried to write a general introduction that encompasses all possible meanings.

It seems that the massive list of definitions is unnecessary, and it seems this might be best organized by how the term is used in three contexts:

Legal
Scientific
Philosophical

Anyway, This was just a start. Jokestress 20:13, 23 March 2006 (UTC)

Basically the section of Anecdote & Witness Evidence is identifiable as the view of Donald W. Miller and Clifford G. Miller, which appears in various articles arguing that legal standard of evidence is superior to that of science/medicine. Personally, I think you'll get a far cleaner article by going back to this version and expanding it (for instance, to include the role in hyothesis generation). Tearlach 20:39, 23 March 2006 (UTC)

That version is indeed very straightforward but heavily weighted toward a legalistic definition. I agree that it is a good basis, especially for that part of the article. I will try to get to this again next week sometime. The conflict between scientific and legal definitions needs to be explored as well. Jokestress 20:55, 23 March 2006 (UTC)

heavily weighted toward a legalistic definition

In what way? That is - genuinely - interesting as an assessment because I'm an ex-scientist with no remote connection with the legal circuit, and I'd take the Miller & Miller version as legalistic. Tearlach 02:49, 24 March 2006 (UTC)

I too am more familiar with the use in relation to scientific evidence and evidence-based medicine, and I feel that context is not well-covered here yet. I'd like to see the scientific/medical use expanded, and then discuss conflicts between legal and scientific contexts (Daubert standard, etc.). Jokestress 20:42, 28 March 2006 (UTC)

[edit] Wikipedia NEEDS this resolved!

A definition to anecdotal evidence is absolutely crucial to wikipedia. A reference to this article is needed to dispute those who wish to quote that their hair dresser's uncle's cat's psychic told them the world is flat. Anecdotal evidence is a scientific concept, I believe we need some real scientists in on the debate.

Seriously, this is a term used every day in scientific and medical professions (not to mention the media) how on earth could this be considered to be disputed? There isn't a disputed tag sitting on the evolution article because it isn't disputed in reality. This article should be no different. Factoid Killer 23:13, 31 March 2006 (UTC)

[edit] Cleanup attempt

OK, I have just given this a pretty serious overhaul, keeping in mind comments on this talk page. I believe I have covered most of the major concerns. However, it is far from complete. At this point, I think a casual reader would get a general sense of the concept and its different uses. I have added citations wherever possible, and I'd like to see more published sources for quotations. Please feel free to make additions or suggestions, but let's try to keep this from becoming unreadable. Thanks to all whose suggestions on this page were incorporated into this iteration. Jokestress 04:08, 1 April 2006 (UTC)

I like what you've done with the article. Well done! Factoid Killer 12:27, 1 April 2006 (UTC)

[edit] Oxymoron

I removed "oxymoronic" from this sentence: "Anecdotal evidence is an informal oxymoronic account of evidence in the form of an anecdote". It's an awkward edit, and moreover is a failure of NPOV 101: instead of defining "anecdotal evidence" as oxymoronic, we need to say who says it is and why. As the article notes, there are in fact some forms of anecdote that are accepted as evidence, e.g. case studies in medicine. The veracity of an observation and how one interprets it will vary with the situation. -Jim Butler 21:17, 31 May 2006 (UTC)

Agreed, "anecdotal evidence" isn't an oxymoron. I too would like to see a published source claiming it is. Jokestress 04:34, 1 June 2006 (UTC)

I've commented on this on my talkpage. See WP oxymoron. Mccready 07:37, 1 June 2006 (UTC)

The article you cite says of terms like "anecdotal evidence": "unlike real oxymora, [they] are used in full earnest and without any sense of paradox by many speakers in everyday language." Labeling the term "anecdotal evidence" an oxymoron is a POV that requires a published citation. You can probably find one in a scientific skeptic's book or in a law journal, two views that have disputes over meaning of the term "evidence." Jokestress 08:17, 1 June 2006 (UTC)

I agree that we don't need to say it's oxymoronic. I'll remove this language, again. Friday (talk) 17:36, 1 August 2006 (UTC)

[edit] Why anecdotal evidence is unreliable

The article needs a section on something like this, citing evidence. From the top of my head ... the study where people witnessed a car crash and were asked about it later. —porg es_(talk) 09:25, 3 August 2006 (UTC)

That would be like Loftus & Palmer (1974) and Loftus & Miller (1978)? Although I would suggest that those speak more just to the unreliability of memory generally, not just anecdotal evidence. If you had a videotape of the *whatever*, it still be flawed through availability or representativeness --Limegreen 09:44, 3 August 2006 (UTC)

Yes, but since anecdotal evidence is totally reliant on the rememberances of the witness, that is relevant :) —porg es_(talk) 05:13, 4 August 2006 (UTC)

Depending on exactly which definition of anecdotal evidence from the article you choose to use, then anecdotal evidence can be be reliant on memory. In a witness context, then probably yes. But if it's a doctor presenting anecdotal evidence of a rise in illness X, then there may be written records or even lab test results. There is then no memorial component, but with doubts to the representativeness etc., the evidence is still anecdotal. --Limegreen 00:14, 5 August 2006 (UTC)

[edit] Flying vs driving

It's often reported that "flying is safer than driving" but without qualification this isn't a meaningful claim. Road travel typically involves relatively short trips with the risk spread relatively evenly throughout the entire trip; air travel typically involves longer trips with much of the risk concentrated at takeoff and landing. The result of this is that comparative safety depends on whether you look at risk per mile travelled (appropriate to driving) or per trip (appropriate to flying). Therefore I've removed this from the article. --Calair 12:30, 10 December 2006 (UTC)

[edit] Recent changes questioned

The definition has once again slipped off the deep end of "anecdotal evidence is bad evidence". Although the evidence quoted from personal anecdote, or even general anecdote, can be unscientific, incorrect, and invalid, it can just as easily be very correct and factual but still be used anecdotally and that use can be correct or incorrect. Thus, bad data has other names and the difference from anecdotal evidence must be pointed out. The complication of the definition is that most people use the term erroneously to mean "bad data". There should be some mention of why that is an erroneous use of the anecdotal logical.

It is not the evidence that is in dispute, it is the method, or conclusion. The main story is what is in dispute not the anecdote. The definition must entertain why the main story is erroneous not the anecdote. If the anecdote is in dispute there are other logical fallacies for bad evidence that are much more powerful to use. It would be inefficient to use the anecdotal evidence logical flaw.

Again, someone pointing out that many smokers die at age 95 from something other than lung cancer or smoking related disease, is valid scientific data. People that eat limes while on a long sea voyage don't get scurvy, is also equally valid scientific data. What is lacking from those anecdotes is the idea that there could be many other variables that contributes to the observed conditions of lung cancer and scurvy. Such factors are diet, and genetics. Concluding from the anecdote leaves the other factors to weaken the debate. It is simpler to say "anecdotal evidence" than what about died, genetics, location, type of water, etc...

Data given anecdotally ignores all those other possible variables. In a sense, it is incomplete evidence. In another sense, it is incomplete study to for the conclusion in the larger part of the paper.

The claim of "anecdotal evidence" should permeate that the conclusion is premature or even wrong, but the evidence doesn't support nor refute it.

If you use the personal anecdotal evidence that you grandmother lived to 95 and didn't die of lung cancer, you could correctly use it to prove that lung cancer doesn't kill all smokers. That conclusion should be a green light for science to study the area of 'why some people don't die young of lung cancer'. That type of conclusion from anecdotal evidence should suggest that the previous lung cancer studies, although possibly good, need more study.

Personal anecdote is often very very irrefutable, especially when both parties know the anecdote personally. In a sense, anecdotal evidence and personal anecdotal evidence are irrefutable correct and useful.

Also: There needs to be an explanation regarding the difference between "anecdotal evidence" and "personal anecdotal evidence". It should incorporate the idea that all science is supposed to be personal. "If you don't agree do the experiment yourself." The difference between scientific facts and anecdote should be given. Remember anecdote is a small story inside a bigger one, therefore it is likely that the anecdotal evidence just doesn't tell it all.

Please correct the definition in the main page to reflect the difference between "bad evidence" and "anecdotal evidence." They are not the same, although many think they are. There is a subtle but important difference. As the page reads now it promotes the idea that anecdotal evidence, is the poorest data type ever. Remember anecdote is a small story in the larger story. So the larger story is not proven by the anecdote, even if the larger story is correct and the anecdote correct. It is the larger stories fallacy that is in error by anecdotal evidence, not the evidence itself. If it is the evidence itself, there are other names given to the bad data.

Example: If you grandmother lived to 95 and smoked her whole life and died in a car-wreck, how do you know her lungs weren't riddled with cancer and would have killed her had she lived just a little longer? That is and example of incomplete study. If it is known that she did have lung cancer, it is incomplete or fraudulent evidence. All those are a much greater fallacy than the personal anecdote.

If the person comes back with, "We did a complete extensive autopsy by three independent labs and all concluded that here lungs, although black were in perfect health." Then the information is just anecdotal if used incorrectly, if used correctly it is scientific.

Eric Norby 16:23, 13 March 2007 (UTC)

These analogies and points are all interesting, but we need a published citation to back these assertions up, so this doesn't veer back into original research. If you have specific citations you'd like to discuss, please list them here. Jokestress 16:32, 13 March 2007 (UTC)

They are examples. Analogy is something else. The best I can give you for citations is to look up any definition of anecdote, data, evidence, and truth. Do not look at the definition for anecdote here at Wikipedia, it is somewhat erroneous and not reflective of most other standard definitions. I only post the comment here as help, should you want help it will require a tiny bit of background effort on your part. To blindly ask for proof of standard definitions seems rather dry towards me don't you think? Eric Norby 16:36, 23 March 2007 (UTC)

[edit] Objection to Miller/Miller table

The table lists scientific evidence as "Irrefutable", which is a little misleading. Without getting into the specifics of the scientific method, one of the "general observations" for the reliability prong of the Daubert Standard is that the theory or technique must be falsifiable. Battles of the experts are a ludicrously expensive feature in trial law and are generally avoided (see Expert witness), but that does not mean that they do not take place. Is this table a direct quote from the source or can it be modified (or at least disclaimed)?Somedumbyankee (talk) 21:56, 2 May 2008 (UTC)

[edit] The whole article is wrongly based

The article as it stands today seems to be confusing 2 entirely different ideas:

1. A story about whose veracity there is doubt (specifically, but not only, hearsay).

2. An accurate and reliable report which has no statistical significance (my grandfather smoked 40 a day until he died at 90). (Of course the report may actually also be unreliable, but the point is that it's not statistically significant).

I think only (2) could be called anecdotal evidence; (1) is either unsupported hearsay or just unreliable testimonial evidence. But if it's felt that both these things should be grouped under "anecdotal evidence" a very sharp distinction should be made between them.

Pol098 (talk) 14:07, 2 June 2008 (UTC)