Do you feel that people trust the internet too much for information?

[welsummerchicks]

I honestly have absolutely NO problem with small sample studies.

None.

As long as it's the kind of data that 'behaves' in small samples.

If someone does research on a neurotransmitter in 14 fruit flies, I'm fine with that. There's a fair body of evidence that already indicates that fruit fly neurotransmitters work in very consistent ways. No problem there. And as far as we know how similarly that neurotransmitter works in humans, I'm not even perturbed about applying that information to humans, if they have the same neurotransmitter and it does the same thing. If there's very little variation in that neurotransmitter in humans. No problem there.

But again, if you're looking at a fairly 'ill mannered' set of data, you have to do all sorts of things to handle your exceptions, your anomalies, and all your just 'what the heck is that' data.

Some kinds of data are very, very well behaved. You can make generalizations from them. They are cooperative. They are not like Math. They are nice. Nice data. Good data. Read - VERY SIMPLE STUPID DATA. IT's MEASURING SOMETHING VERY - well - they call it - 'discrete'. That means, 'EASY TO MEASURE'. LOL.

OK, so here's an example of The Data that Would Not be Tamed.

When Data Will Not Be Tamed, when someone beats their chest and says they tamed it, you can bet some things that are pretty naughty are at work - like, people WANTING the data to back their beliefs. People having a product that that data COULD help sell, for example.

And this example is even with a BUNCH of studies, all with the same conclusions. And for many reasons, certain groups were not only highly motivated to BELIEVE, they were extremely motivated that everone ELSE believe, too!

The moral of THIS story is - Be awfully, awfully careful when the conclusion of a study is something that makes you HAPPY.

Remember that old saying about how to beware the guy that doesn't agree with you, but run like he** from the guy that AGREES with you?

Well, as you read the example keep this in mind - Very few scientists EVER write ANY conclusion to ANY study except this one: "The results are interesting and point out the need for further research". Then their write up tends to ramble on and on about how you should NOT generalize their results and why. Because they did THIS in the study, because they did THAT in the study. So it's interesting - really titillating, honestly, this little result we got, but it's just this one little bit of mortar of the process of developing new knowledge.

It is the media that writes in the newspaper, 'High Heels cause Cancer!' NOT the researchers.

So.

Here's what happened. A study identified a concept in mental illness treatment called 'Expressed Emotion'. Sociologists observed that parents of schizophrenics that were laid back, relaxed, didn't ask a lot of questions about 'are you taking your medicine' and 'you aren't going out to drink with your buddies, are you', had family members that were hospitalized less, were less sick overall, had fewer severe symptoms - the whole nine yards. Houses where there was less arguing, less nosing into the sick person's business, less tension, all that - those people were healthier. Everyone in the whole FAMILY was happier.

Those studies went on a long time. They really 'built up' and counselors even started advising parents to take a very laid back approach, just don't worry, be loose, be cool, relax. Don't monitor meds, even in some cases. JUUUUST be cool.

Then someone asked a very, very simple question.

Which was - didn't it ever occur to anyone that the families that were the most upset about their sick family member, that were the most intrusive, tense, trying to control their family member, asking lots of nosy questions - did it ever occur to anyone that they did that because the person was simply - sicker to start with? Maybe these families are doing all these things because they're dealing with a different sort of case.

What about all these families that were low EE and they had real sick family members? With MRI's that went totally to pot in the midst of all sorts of wonderful gemutlichtheit (cheery laid back surroundings full of low EE).

Gradually, it started creeping up and tapping people on the shoulder, that this EE deal really wasn't the magical answer to all our problems with mental illness. People started thinking - questioning - wondering.

What about all those families that had really low EE til the third time their person landed in the hospital after a quick ride in a police cruiser? What about all those families that had been just cruising along peachy in the low EE world, until Son burned the house down, or killed someone? Or til they suddenly found out, well well, Mary isn't just being a pita, she actually is SICK.

What about that family with the guy who is the alcoholic, and dad gives him a breathalizer test and counts his pills every week, and constantly reminds him what can happen if he doesn't take his meds, or if he starts drinking again? And he's doing so much better than he was last year when dad let him frolic and be free? Now dad asks him who he's going out with every time he leaves the house, and lectures him firmly about what can happen if he screws up. And if he does mess up, all heck breaks loose - lots of EE popping around all over.

What about that guy that tried to blow up the mental health building in Texas? He wasn't doing too hot with a little bit of cheerful freedom. He needs someone to be keeping an eye on him, and to be honest, it goes along with a whole lot of EE- nosing into his symptoms, are they changing, why is he whittling that knife out of that frying pan and who's he going to kill with it, all that. He works out better if he is followed around and kept after - big time.

Today, EE is still an unshakeable absolute truth - but only in a few people's minds! To them, EE makes schizophrenia arise, EE makes schizophrenia worse, EE is what gives us all the lousy outcomes. Let them run naked and free, and they won't even be sick!

And today, too, there are varying degrees of 'believingness' in EE.

Around about in the normal middle sort of the field, EE is believed to be a HELP - in some cases - relax, take it easy, accept your ill relative, learn about the illness, find out what efforts work, and take it easy on the ones that don't. Don't feel guilty, if your loved one became sick, or if he has a relapse, sure, maybe there are a few things you can change, but this disease has a natural history- it's genes, it's chemistry, it stinks, but it's not your fault it started, or that it stayed, or anything else.

So EE, which originally was heralded as the answer to everything, as the solution, the cause, EVERYTHING - as something that people could be trained to manage, to buy books about, to attend seminars about, to hire EE counselors to be guided about (hint hint hint), it turns out to be 'hey that's an interesting idea...sort of....'

And so the moral of the story is - 'If something sounds too good to be true, it probably is'.

 

[mom'sfolly]

Dang it, you mean to tell me you can't just "feel good" the mentally ill into a healthy state??????

Yep, I'm sure that when family members are less ill, take their meds without prompting, and are functional, the whole family is happier and saner (I'm not sure if saner is a word, but I'll go with it. My family all gets along pretty well with my mom when she is well....when she is sick, it's a whole 'nother story.

Most data in studies seem to be as slippery as a greased weasel. That is what makes the doubters so happy, they can find plenty of room to play. A well designed study can also get the same weight as a poorly designed study, and then there are those mega studies, that take data from multiple studies and try to reach a conclusion.

 

[welsummerchicks]

"most data seems as slippery as"

I don't find DATA to be at all slippery, honestly.

I find the conclusions that various rubbing-the-hands-together companies and such make from the very reasonable conclusions the people actually IN the study made - to be slippery as a greased weasel!

"High Heels Cause Cancer!"

The very worst, the least understood, the stupidest conclusions, they come from correlation/association studies.

An association/correlation study is what we call a 'trouffle hunt'.

A trouffle is a fungus that grows under the ground. It is tasty, especially with eggs (relevance to BYC). In fact they're so good if you put a little wee piece of trouffle in an omelet you actually drape a dish cloth over your head when you eat your omelet so that not even one little tiny aroma of trouffle gets away.

The reason this metaphor works is that PEOPLE REALLY LIKE TROUFFLES.

You take your trouffle hunting pig, and you go out in the woodsy woods, and you say, 'dig, Trouffle pig!' and it snuffles around - looking for trouffles.

Sometimes he finds a trouffle, sometimes he doesn't. But he sure snuffs around a lot.

That's all an association/correlation study does.

It says 'Hm! There's something that MIGHT be going on here...'

For example, many more patients get diagnosed with colon cancer, are wearing high heeled shoes, than if you just walked down the street and picked out a bunch of people randomly.

Oh my freaking God! High heel shoes cause cancer!

Well maybe high heels AGGRAVATE cancer! Make it worse! Make it start, when it wouldn't otherwise! Blah blah blah!

Not really. Colon cancer tends to occur in older women. Older women are more likely to wear high heeled shoes.

And on the very next day, the headlines in the local paper read, 'High Heels cause Cancer!'

Sure, some day, we might find that it isn't really true that A causes B. Maybe it just so happens that people do A tend, in general, to do D, E, and F. And maybe one of them has something to do with it.

Or not.

Don't get me wrong. If I find mutated proteins from substance A in cell B, and that mutated protein turns out to be what very, very reliably turns off that tumor necrosing factor that normally is my best friend, and so that cell turn cancerous, and those DNA just wind up sticking themselves into other cell's DNA and messing them up too, I'm believing that there is a cause and effect there. That's fine. Don't eat A. Pretty simple.

Even if I thought that if I ate A, there was a 14 percent increase in precancerous changes that COULD lead to cancer, in certain individuals for reasons we don't quite understand just yet, I can do with out A.

And for a scientist to say, 'You know, we're starting to see some pretty uncool things happening with these proteins and DNA, that...really....bother us...', yeah, I don't need to eat it.