Vidar Leathershod is kind of an interesting guy and in an exchange I had with him a few weeks ago he is rather good at demonstrating how not to make your point...on slashdot or anywhere else.

He starts off with a post berating someone for requiring a citation for the claim that home schooled children perform better and are more socially well adjusted than their non-homeschooled peers. From a reference in a wiki article he cites:

Among the homeschooled students who took the tests, the average homeschooled student outperformed his public school peers by 30 to 37 percentile points across all subjects.

It was pointed out to him that this research doesn't publish it's error values so it's difficult to take seriously.

His notation on being socially well-adjusted was:

John Taylor later found, using the Piers-Harris Children's Self-Concept Scale, "while half of the conventionally schooled children scored at or below the 50th percentile (in self-concept), only 10.3% of the home-schooling children did so."

One thing is that rather than a mean value a specific ordinal is chosen for some undisclosed reason and again there are no error values here. So again the same problem as above.

It was also pointed out that there appears to be a big problem with selection bias here and while he is maligning someone for not reading the Wikipedia he obviously didn't read the article too closely himself since the evidence presented is insufficient.

Now this is where it gets interesting. Instead of addressing these points Vidar says this:

If you have something to add to the article, why don't you just add it, and cite it. Instead of complaining about the cited content, try and get it removed.

This is a good example of a fallacy of distraction. The fact that something is or is not represented in the Wiki is independent of it being correct. It's also asking someone to take a pretty roundabout method to get proper evidence cited. The question at hand is one of how much weight we can put on the cited research but Vidar suggests that instead of producing information that he claims to have:

Because there have been many published studies on the success of homeschooled children in education vs. their formally-schooled counterparts.

And essentially do his homework for him (or her). He (she?) then continues:

Note that I did not say that Homeschooling always produces good results. But it has a better track record than the alternatives.

Again a distraction from the actual point: "Does homeschooling, in fact have a better track record?"

Instead of trying to fight it, why not try to find out why it is more successful and take lessons from it.

And again a distraction. Since we first need to establish that it *is* better before we "take lessons from it"

Anyway so the two lessons here are:

i) Stay on topic.
ii) Support your points.

Later on I find him saying some other stuff:

If you read what I was replying to again, I want you to see that the person was basically trying to put the onus of proof OP not to discover the truth, but to try and weaken his argument without giving his own evidence.

It's an interesting look into someones mind here. Vidar seems to object to having evidence required of the person if they possess certian motives - which he can somehow discern from the following post:

[Citation Required]

Sorry didn't I mention it? This is the post that Vidar responded to. It had the text of the previous poser that was in question above it but the only user generated content was that.

Vidar responded with:

Maybe you should "require" yourself to read a basic Wikipedia article before trying to "require" citations from them.

Anyway this represents two other logical fallacies. a) Non-support - Vidar rejects that a statement needs support because of b) Special pleading - that there is some special case - here being 'bad motives'.

The lesson here would be avoid appeals that speak to someones motive they are unconvincing - Even if that person asked the question out of the most evil motive in the world (I dunno putting kittens into wood chippers or something) it still may be the right question to ask. Likewise it's difficult to argue that someone possesses a particular motive anyway.

This gets brought up to him and the result is:

Obviously, you like arguing. I do to. But not for a freaking week. You literally have tired me into submission, as I cannot even be bothered to read your latest drivel, everyone who disagrees with you is stupid, I get it, there are no other motives in your heart but the pursuit of knowledge and purity, and the fact that you want to defend someone making a baseless attack on people who choose to homeschool or something.

Whoo almost too many things to list here:
i) Prejudicial language - calling something "drivel" qualifies.
ii) Strawman - The only person who was called stupid was Vidar...seemingly justified.
iii) Appeal to motive - similar to above.
iv) Strawman - All of a sudden wanting a citation is a "baseless attack on people who choose to homeschool" and taking him to task for not reading (well) an article that he chided someone for not reading is defending "anti-homeschooling".

Hmm, first time using this. Mostly due to clarifying some things said before comments closed on the following thread:

ltbarcly has some odd opinions:

"What you say is true. One symptom is not the disease. However, one symptom is a symptom of the disease, by hypothesis. So he has several of the symptoms (or characteristics which have a high level of correlation with the disease if you like) of ADD, as I previously pointed out."

Except that ltbarcly never pointed this out...what was originally said by ltbarcly was:

"Actually, that isn't 'flow'. It's Attention Deficit Disorder." Not, "this may be a sign of" or "this marginally increases your chance of having"...

Somehow also ltbarcly is claiming more than one symptom now (although later he/she seems to change his/her mind). How ltbarcly came to the conclusion that 'hyperfocus' (which isn't recognised as a symptom by the DSMIV) and the other characterisitcs are 'highly correlated' is left unsaid.

I'm going to step over the jawing about "Wiki being at least as accurate as gospel" since more than a few logical flaws are made and it's majoring (most of his/her text was dedicated to this subject) on a minor.

So after that ltbarcly goes on:

"I am not making a 'directional fallacy'. I am stating that he had one clear symptom of ADD."

Well not even, since (a) There appears to be little evidence to support the assertion that this is a symptom of ADD (see below) and (b) even if we were to write that symptom into the list in the DSMIV you still have the problem that it's unclear if it is being exhibited "to a degree that is maladaptive and inconsistent with developmental level"

So although ltbarcly has changed his/her tune - albeit by pretending he/she said something different than was actually said. This new position isn't quite supported either.

"If there is a higher than average statistical correlation of the presence of a symptom in a person and the presence of the disease,"

Here I'm begining to suspect that ltbarcly doesn't even understand what he/she is typing. What ltbarcly appears to be saying here is if something has greater than 50% correlation with a particular disease then it's a symptom if not it isn't. So something that occurs in 10% of patents isn't a symptom. Which is silly. There are plenty of diagnoses that depend on a series of conditional probabilities which only collectively give you a better than average chance at being diagnoses. Take for example the early screenings for certian birth defects. They rely on two samples of blood word, plus at least one ultrasound. The result? If you screen positive you still have a better than average probabilty of NOT having the disease.

"then if someone does exhibit the symptom they have a correspondingly higher likelihood of having the disease."

Here is another problem with ltbarcly's thinking but this time it's with math. The issue isn't about 'correspondance' it's about significance. Sure a symptom is related to the probability of having a disease but it doesn't have to be to a significant degree.

Take for example a disease that is incredibly rare ( striking say 1 in 1,000,000,000,000 ) but there's a test for it that is right 99.99% of the time. You can say that testing positive for this disease is a 'symptom' of having the disease but the probabily of having said disease even after testing positive isn't terribly signifcant ( you have a far, far, far better chance of not having the disease than having it).

The last bit is sad and funny.

"Besides, the wikipedia article, of you cared to read it, goes over in detail your objections, including hyper-focus not being in DSM - IV."

What medical facts it actually states are a few things.

1) There is no pubmed article with that term.
2) It isn't recognized by the DSM-IV.
3) It states, without a corresponding in-text cite that there is a 'well-recognised' (peacock word) comorbitity with another set of disorders that contain that as a part.

This is where ltbarcly clearly shows which side of the fence he/she is on concerning 'people who don't know how to not believe everything that they read'. That text does nothing to refute the idea that hyperfocus isn't a very good indicator of ADHD.

Essentialy ltbarcly is saying that it's ok to make a diagnosis ("This is ADHD") on the basis of something that isn't recognised by the clincial standard, doesn't appear to be backed by a published study and defend your diagnosis on the basis of a statement of unspecified origin that this disorder is somehow linked to another group of disorders where some people exhibit this aleged symptom.

Perhaps this is why he/she spent so much time defending the idea that the Wiki is as good as gospel.