Anecdotal Evidence

[Tonyman64]

Why do people bring up anecdotal evidence in response to scientific studies?

[Kasierith]

Because they lack the patience to actually understand the study itself, and to comprehend the difference between controlled study and random events

[Shoomoo]

I don't know. I find anecdotal evidence to only be sufficient to start an investigation. I do not find it to be good enough evidence to PROVE something so to say. Imagine a product someone is selling. There is the possibility of many false testimonies to whatever it is. The same goes for various bogus claims and anecdotal evidence. Am I saying all of the claims or testimonies never happened? No. But with the idea of it being not sufficient for everyone, or repeatable, or can easily be faked, I do not see why it should be on the rank of say empirical evidence. I guess ones experiences sway depending on the situation.

[Spectral]

http://rationalwiki.org/wiki/Dunning-Kruger_effect

[Tonyman64]

As much as I love the Dunning-Kruger effect, I don't this is really a case of it in action.

It's just that I often see people respond to studies by saying that their personal experience contradicts the study.
Do you think it's an inherent human issue or one that could be solved with more education in statistics and the scientific method?

[damntree]

As much as I love the Dunning-Kruger effect, I don't this is really a case of it in action.

It's just that I often see people respond to studies by saying that their personal experience contradicts the study.
Do you think it's an inherent human issue or one that could be solved with more education in statistics and the scientific method?

Anecdotal evidence stating there is a problem with people using anecdotal evidence.


9/10

[Spectral]

As much as I love the Dunning-Kruger effect, I don't this is really a case of it in action.

It's just that I often see people respond to studies by saying that their personal experience contradicts the study.
Do you think it's an inherent human issue or one that could be solved with more education in statistics and the scientific method?

I disagree that it's not a Dunning-Kruger type situation. As near as I can tell, only people that are under the impression that anecdotes refute solidly done science are people that are grossly ignorant of science. As a result, these are the people that present anecdotes with the most frequency, are most impressed with their value, and most convinced that they're an adequate refutation of solid science.

I don't think it's an inherent human issue, but I think it's one that would be difficult to fix in people that are basically unintelligent. I'm a bit of a cynic when it comes to people's ability to become informed; most of the people that have any desire at all to become informed of how the world works are already aware that anecdotes aren't impressive.

[Tonyman64]

Anecdotal evidence stating there is a problem with people using anecdotal evidence.


9/10

I'm not saying anecdotal evidence is pointless in all situations. I'm talking about situations in which people claim it contradicts scientific studies. I have yet to see a study on rates of anecdotal evidence use, so I don't believe I did anything wrong in making the statement I did. I merely want to discuss the issue.

---------- Post added 2012-07-03 at 12:55 AM ----------

I disagree that it's not a Dunning-Kruger type situation. As near as I can tell, only people that are under the impression that anecdotes refute solidly done science are people that are grossly ignorant of science. As a result, these are the people that present anecdotes with the most frequency, are most impressed with their value, and most convinced that they're an adequate refutation of solid science.

I don't think it's an inherent human issue, but I think it's one that would be difficult to fix in people that are basically unintelligent. I'm a bit of a cynic when it comes to people's ability to become informed; most of the people that have any desire at all to become informed of how the world works are already aware that anecdotes aren't impressive.

But I don't think the people doing this are really aware of the fact that they are doing it at all. I don't think they realize that they are attempting to refute science with anecdotes. Although I guess then you could make the argument that it is the situation you described for that reason.

[Gheld]

Why do people bring up anecdotal evidence in response to scientific studies?

I actually know a few people who have won arguments using anecdotal evidence.

But my serious answer would be that only the empirical sciences are absolute. Everything else is open to subjective interpretation.

So unless the study leads to the ability to make a 100% accurate mathematical prediction 100% of the time, neither party is in a position to speak in absolutes.

So if say the study is that in 9 out of 10 cancer patients apples were found to improve libido, you are 10% wrong to say that apples improve the libido of cancer patients. If your friend says "I know a guy with cancer whose libido wasn't improved by apples, therefore apples do not improve the libido of cancer patients" then he is 90% wrong. But he's also 10% correct in this instance.

At the same time the results of some studies, although consistent, are very low in their level of significance. For example in Canada the prevalence of lung cancer in males in the general population is 62 per 100,000. Studies have shown that among smokers only the prevalence of lung cancer is double. So 124/100,000 (which although it doesn't sound like a lot, I think 124 families beg to differ).

So if you're spouting lung cancer statistics, and he says "My grandpa smoked a pack a day his whole life and lived to be 100" he's not necessarily completely off base, because 99.876% of smokers don't get lung cancer.

The fact is in this scenario the scientific study over-simplifies the situation to the point of stupidity because it's a red herring to other major health concerns with smoking, and fails to address the fact that the most of the 62 non smokers with lung cancer will probably live long enough to die of something else, whereas nearly all 124 smokers with lung cancer will die young and from the cancer itself.

But if you simply state "smoking gives you lung cancer, here's the proof" you're packaging the study shit poorly and I wouldn't blame anybody who throws anecdotal evidence at it.

So

Because they lack the patience to actually understand the study itself, and to comprehend the difference between controlled study and random events

The person citing the scientific study may also be presenting it poorly.

[Spectral]

But I don't think the people doing this are really aware of the fact that they are doing it at all. I don't think they realize that they are attempting to refute science with anecdotes. Although I guess then you could make the argument that it is the situation you described for that reason.

This is what I'm shooting for, although I may be communicating it poorly. The people most inclined to do this are so oblivious that they're not even capable of qualitatively distinguishing the difference between science and anecdote.

[Tonyman64]

I actually know a few people who have won arguments using anecdotal evidence.

But my serious answer would be that only the empirical sciences are absolute. Everything else is open to subjective interpretation.

So unless the study leads to the ability to make a 100% accurate mathematical prediction 100% of the time, neither party is in a position to speak in absolutes.

So if say the study is that in 9 out of 10 cancer patients apples were found to improve libido, you are 10% wrong to say that apples improve the libido of cancer patients. If your friend says "I know a guy with cancer whose libido wasn't improved by apples, therefore apples do not improve the libido of cancer patients" then he is 90% wrong. But he's also 10% correct in this instance.

At the same time the results of some studies, although consistent, are very low in their level of significance. For example in Canada the prevalence of lung cancer in males in the general population is 62 per 100,000. Studies have shown that among smokers only the prevalence of lung cancer is double. So 124/100,000 (which although it doesn't sound like a lot, I think 124 families beg to differ).

So if you're spouting lung cancer statistics, and he says "My grandpa smoked a pack a day his whole life and lived to be 100" he's not necessarily completely off base, because 99.876% of smokers don't get lung cancer.

The fact is in this scenario the scientific study over-simplifies the situation to the point of stupidity because it's a red herring to other major health concerns with smoking, and fails to address the fact that the most of the 62 non smokers with lung cancer will probably live long enough to die of something else, whereas nearly all 124 smokers with lung cancer will die young and from the cancer itself.

But if you simply state "smoking gives you lung cancer, here's the proof" you're packaging the study shit poorly and I wouldn't blame anybody who throws anecdotal evidence at it.

So


The person citing the scientific study may also be presenting it poorly.

But if the person is attempting to refute the study with the anecdote about their grandfather then that is what I take issue with. The fact that not everyone who smokes gets lung cancer doesn't mean that someone saying smoking isn't dangerous isn't wrong.

Nonetheless, this is the kind of discussion I was hoping for. Discussing with someone who either believes the opposing side or can effectively play devil's advocate for it is useful.