Statistical inference - invalid research

[Guest guest]

>>Hiring a statistician doesn't get you there. How can you calculate

percentages of Yin and Yang, of Jing and Qi, of one meridian or

another? Calculus has to be a tool that describes Yin/Yang,

Jing/Qi/Shen, Liver Yang, Heat/Cold, Upward Directing/Downward

Directing. How shall calculus address these things? You have to

start with methods and tools of CM. If you add a method or tool to

the situation, you have to characterize the effects of that tool on

the entire enterprise. What exactly are units of Qi? If I calculate

the percentage of Qi in anything, will this have meaning? I need to

show my units to present statistics. So do we go for millivolts

here? How about micrograms of .... Jing?

 

I'm a statistician. Hire me. But please give me units! Otherwise

there's no statistics.

 

Emmanuel Segmen

Merritt College>>

 

Emmanuel is making a very important point here. If I've understood

him correctly, this is the point: We're kidding ourselves if we think

we can apply a statistical methodology to the study of CM in any

methodical manner. For a CM practitioner, the differentiations in CM

involved in diagnosing pathology and determining treatment are all

important, and these variables are precisely those things that can't

be measured quantitatively. The entire premises of scientific method

utilising statistics, that you study phenomena by studying changes in

quantifiable and measured variables, can't be applied to CM if the

variables can't be quantified and measured. It's almost a joke to

suggest that they can. So you're pretending to do science without

controlling for, and measuring, all the SIGNIFICANT VARIABLES

involved.

 

This is well known to be faulty scientific method, and there have

been prominent examples of research going wrong for this reason.

 

Emmanuel's insight is that the problem of not being able to quantify

variables in CM fundamentally invalidates any scientific research in

which these variables are, or might be significant.

 

 

Wainwright

[Guest guest]

How can you calculate percentages of Yin and Yang, of Jing and

Qi, of one meridian or another?

 

...We're kidding ourselves if we think we can apply a

statistical methodology to the study of CM in any methodical

manner.

 

You guys sound like you're having too much fun waxing to be practical,

however, you are talking about the issue of measuring the diagnostic

parameters of CM rather than the outcomes of the treatment choices to

which they lead. It is quite possible to quantify treatment outcomes

within the context of accepted, albeit subjective, outcomes

questionnaires developed for western " scientific " research without

changing any aspect of diagnosis or treatment according to CM. If 50

people have spleen qi xu duck stools and they take bu zhong yi qi tang

and a statistically significant greater number of them get better,

perhaps even at a faster rate than a placebo group, then what is wrong

with that as a measure of the value of CM?

 

Unless you don't believe in the validity of CM diagnostic parameters

such as yin/yang, then why test to determine whether they exist rather

than relating their value to the thing that matters most in the end

anyway, the patient's wellbeing.

 

So you're pretending to do science without controlling for, and

measuring, all the SIGNIFICANT VARIABLES involved.

 

You don't need to control all variables. That's impossible. Isn't that

why 'n' is part of the equation?

 

 

This is well known to be faulty scientific method...

 

Redundant " faulty scientific method " would be Obiewan. Yet valuable

slivers of relevant reality and truth can still emerge despite the

limitations of the process. JMHO.

But please wax on...

Stephen

 

 

 

Chinese Herbal Medicine, a voluntary organization of licensed

healthcare practitioners, matriculated students and postgraduate

academics specializing in Chinese Herbal Medicine, provides a variety of

professional services, including board approved online continuing

education.

 

 

 

 

[Guest guest]

Wainwright,

 

Whew! Thank you.

 

Emmanuel Segmen

-

wainwrightchurchill

Monday, October 27, 2003 11:10 AM

Statistical inference - invalid research

 

 

>>Hiring a statistician doesn't get you there. How can you calculate

percentages of Yin and Yang, of Jing and Qi, of one meridian or

another? Calculus has to be a tool that describes Yin/Yang,

Jing/Qi/Shen, Liver Yang, Heat/Cold, Upward Directing/Downward

Directing. How shall calculus address these things? You have to

start with methods and tools of CM. If you add a method or tool to

the situation, you have to characterize the effects of that tool on

the entire enterprise. What exactly are units of Qi? If I calculate

the percentage of Qi in anything, will this have meaning? I need to

show my units to present statistics. So do we go for millivolts

here? How about micrograms of .... Jing?

 

I'm a statistician. Hire me. But please give me units! Otherwise

there's no statistics.

 

Emmanuel Segmen

Merritt College>>

 

Emmanuel is making a very important point here. If I've understood

him correctly, this is the point: We're kidding ourselves if we think

we can apply a statistical methodology to the study of CM in any

methodical manner. For a CM practitioner, the differentiations in CM

involved in diagnosing pathology and determining treatment are all

important, and these variables are precisely those things that can't

be measured quantitatively. The entire premises of scientific method

utilising statistics, that you study phenomena by studying changes in

quantifiable and measured variables, can't be applied to CM if the

variables can't be quantified and measured. It's almost a joke to

suggest that they can. So you're pretending to do science without

controlling for, and measuring, all the SIGNIFICANT VARIABLES

involved.

 

This is well known to be faulty scientific method, and there have

been prominent examples of research going wrong for this reason.

 

Emmanuel's insight is that the problem of not being able to quantify

variables in CM fundamentally invalidates any scientific research in

which these variables are, or might be significant.

 

 

Wainwright

 

 

 

[Guest guest]

Emmanuel is making a very important point here. If I've understood

him correctly, this is the point: We're kidding ourselves if we think

we can apply a statistical methodology to the study of CM in any

methodical manner. For a CM practitioner, the differentiations in CM

involved in diagnosing pathology and determining treatment are all

important, and these variables are precisely those things that can't

be measured quantitatively. The entire premises of scientific method

utilising statistics, that you study phenomena by studying changes in

quantifiable and measured variables, can't be applied to CM if the

variables can't be quantified and measured. It's almost a joke to

suggest that they can. So you're pretending to do science without

controlling for, and measuring, all the SIGNIFICANT VARIABLES

involved.

>>>>>>Well both you and Emmanuel need to study more statistics. Please refer to

the science of human behavior and models for longitudinal and multifactorial

models.

Alon

 

 

[Guest guest]

S. M. wrote: You guys sound like you're having too much fun waxing to be

practical,

however, you are talking about the issue of measuring the diagnostic

parameters of CM rather than the outcomes of the treatment choices to

which they lead.

>>>Thank you

Alon

 

Stephen and Alon,

 

I'll call your bluff. Measuring the outcomes is precisely what I'm talking

about. Quantifiable units are in Western medicine. If this is not true, please

state clearly what numerical units of CM you wish me to use. Please do not

leave this thread until you have stated this. This is the crux. Outcomes. Got

it? Go. State away. And don't give me Western medical terms or you flunk. I

want CM units. Fire away.

 

Emmanuel Segmen

 

 

 

[Guest guest]

Your shoulder(s) hurts you (group) can abduct it to x degree you needle st-38

you can raise shoulder to y degree. I think even you can plot this in a two by

two table

alon

 

 

[Guest guest]

I'll call your bluff. Measuring the outcomes is precisely what I'm

talking about. Quantifiable units are in Western medicine. If this is

not true, please state clearly what numerical units of CM you wish me to

use. Please do not leave this thread until you have stated this. This

is the crux. Outcomes. Got it? Go. State away. And don't give me

Western medical terms or you flunk. I want CM units. Fire away.

Emmanuel Segmen

 

 

Emmanuel,

Relax bro. There is no bluff, just an attempt at reasonable dialog.

Life is not black and white, and no one flunks for sharing their valued

opinions.

 

Using the example of an outcomes study of spleen qi xu loose stools, it

is possible to scale the degree of looseness of stool with actual words

rather than numbers or " units " such as watery, like paste, with or

without undigested food, solid, or firm. This scale, or one along these

lines, stays within the terminology of symptoms that helps define spleen

qi xu and are clearly consistent with the context of CM. There is no

need to quantifiably measure spleen qi in order to determine whether it

has been influenced. Unit quantification is not necessary in order to

either diagnose the problem or determine whether it has resolved. Signs

and symptoms work fine. Headache for example can be severe, frequent,

or gone, resolved. These are important factors to determine in

diagnosing or assessing the treatment outcome that are fully consistent

with the practice of CM. They also serve the purpose of established

" scientific " outcomes research quite well.

 

Also, I don't agree that using numbers to scale the degree of severity

of symptoms evaluated according CM somehow makes the CM diagnosis or

treatment less relevant in a research setting. Take for example 50

people with a skin condition manifesting as red spots on the face,

regardless of the fact that it may or may not be called acne in western

medicine. If 3 experts agree that all of them have the CM diagnosis of

heat in the blood, appropriate herbs are prescribed, and we count (yes,

numerically) the number of spots before and after two weeks of either

treatment or placebo, then we have a statistically valid way of

calculating the outcome of the treatment. There is no change in the way

the person was diagnosed or treated within the parameters of CM. Yet,

numbers were used unobtrusively to evaluate the outcome of the CM

diagnosis and treatment of the condition. You do not need to count

units of heat in the blood. That would not only seem ridiculous but

completely unnecessary for the purpose of evaluating the potential for

use of CM as a treatment of red spots on the face related to heat in the

blood. Use numbers. They're ok.

 

Kind regards,

Stephen

 

 

 

 

 

[Guest guest]

At 9:39 AM +0000 10/28/03, wainwrightchurchill wrote:

>If we accept this heterogeneity in approaches as an important reality

>about CM, it seems to me that it makes CM increasingly inappropriate

>for scientific investigation, except perhaps of a very crude kind.

--

 

Wainwright,

 

Depends what you mean by scientific investigation. If you mean double

blind studies looking at bio-medical criteria, then I agree with you,

it would not be appropriate. If, on the other hand, you mean research

on the basis of the science of Chinese medicine in it's own terms,

then it is appropriate and possible. It is quite possible to factor

out heterogeneity, for example by making the study of the work of

only one doctor, or one style, or one type of treatment, or making

the study disease based rather than pattern based, or studies of the

prevalence of certain patterns in particular diseases. How about

comparing the use of a single standard herb formula to treat a (CM)

disease, and comparing that the same disease using pattern based

treatment.

 

In other words, there are plenty of possibilities for research

completely according to the terms of CM, even allowing for the fact

that the practice is heterogeneous. Such research could be of value

to practitioners in clarifying clinical issues, even though not each

research study would of value to all CM practitioners.

 

Rory

 

 

--