Jump to content
IndiaDivine.org

Statistical inference

Rate this topic


Guest guest

Recommended Posts

Ken (, Rory, Emmanuel, et al,)

 

The statistical issue is, I feel, important for us to come to terms

with. There's no doubt that one is taking a cultural and paradigmatic

leap if one is prepared to embrace the use of statistical techniques

in the field of CM. It begs many questions, some of which we've

already alluded to. Yet, if we resolutely refuse to have anything to

do with statistics, where does that leave us? Talking about coyotes

having breakfast?

 

Wainwright

 

 

 

 

-

" kenrose2008 " <kenrose2008

 

Sunday, October 26, 2003 2:46 PM

Statistical inference

 

 

> Emmanule, Wainwright, and All,

>

> An early issue of CAOM, I was

> in Ocean Park so it must have been

> in 2000 or 2001, included an article

> by a mathematician named Ron Bloom.

>

> In it he addresses historical aspects

> and ideological presumptions related to

> statistical inference.

>

> I recommend everyone take a look at it.

> It's an odd little piece, and I took heat from my own

> editorial committee (that's what

> they were paid to do, but there was no

> pay). But I wanted to include it

> into what I had envisioned as an

> ongoing discussion about the foundations

> on which much clinical research is

> still based.

>

> The idea that everything that matters

> can be counted up, and that only those

> things that can be counted matter is

> nothing more and nothing less than a

> prejudice based on philosophical assumptions.

>

> I'm neither a philosopher nor a historian.

> So if the details matter to anyone, someone

> else will have to spell 'em out.

>

> But there is no reason on earth that

> any mind should be wedded to the notion

> that mathematics is the only method or

> only valid method of knowledge.

>

> The ancient Greeks recognized at least

> three approaches to the use of the

> cognitive hardware and software, and

> mathein is just one. (The others are

> gnose, and skene.) That statistics

> rose from the latter 19th throughout the 20th

> centuries to become the standard of

> proof about complex systems like human

> beings engaged in clinical studies is

> a kind of magnificent exuberance of the

> will of the new power elite that evolved

> in that time to take the place of the

> deposed despots of imperial times.

>

> It's really a big mistake to forget

> or ignore that medicine is human action

> and subject to the same laws and principles

> and vagaries of human strengths and weaknesses

> as all others.

>

> The mathematicians and scientific researchers

> who work for pharmaceutical companies just

> naturally devise ways of testing and demonstrating

> the safety and efficacy of their sponsors' products

> that will enable the cash flow of all to continue

> and thrive.

>

> From a businessman's point of view, that is the

> real curative effect of any medicine. It cures

> that deep aching emptiness in the bottom of

> the bank account.

>

> So we wind up with a method of demonstrating

> the efficacy of medical treatment that was

> developed to calculate the effects of fertilizer

> on test fields in England somewhere. But if you

> look at the nameplates on the factories and the

> logos on the brand name medicines, you'll discover

> this really terrible and terribly funny fact.

> Same folks making the deposits, by and large.

>

> And you discover how truly effective those

> medicines are. They keep alive a whole corporate

> culture for longer than the indivdual humans

> that create and sustain it.

>

> This is not a dismissal of the whole enterprise.

> I'm merely trying in dramatic fashion to

> bring to light this odd fact that a method

> developed to test fertilizer, statistical

> inference, the idea that you can count up

> the important variables and nail them down

> quite neatly and cleanly...why should that

> method be applicable to something as messy

> as a human body?

>

> Just now through the silence of an early

> morning here on Thunder Mountain, come

> the sounds of a pack of coyotes killing

> their breakfast down by Matanzas Creek.

> Matanzas means " slaughter " and even the

> animals here seem to be able to read

> the invisible writing that hangs heavily

> in the air.

>

> Or more accurately, the earlier inhabitants

> who coined the words thunder and slaughter

> to describe the local landscape were able

> to read the same messges that the animals

> and the plants and the earth itself speak

> day by day as the eons pass.

>

> How many white people I wonder have wandered

> through these places without knowing the

> meaning of where they were?

>

> The truth is that even those small patches

> of ground that those Victorian scientists

> sweat over to demonstrate the efficacy of

> one kind of fertilizer over another...but

> who ever stopped to question the whole

> strategy of forcing nature's body to

> produce more than it should...even those

> tiny well defined squares of the vast

> and chaotic earth were too complex to be

> contained in the descriptions of stupid

> white men who could not understand what

> the world was saying.

>

> And still can't!

>

> Rory asked does it make Unschuld's work

> suspect that he relies on some sort of

> statistical inference?

>

> Absolutely.

>

> But his work was suspect long before that.

>

> Everyone's work is suspect. Rory's work.

> My work.

>

> There are always lots of things that I

> want to question Paul about, and that's

> why I place such a high value on these

> opportunities that arise to spend time

> with him. Once again, if at all possible

> people should come and hear him at PCOM.

> Just the idea of formulating a question

> to ask him about these things is a

> worthwhile mental exercise.

>

> And I think we should work together to

> provide more opportunities for him to

> speak and interact with students and

> practitioners and especially those involved

> in education.

>

> Everything must be questioned until you

> are satisfied that you have a personal

> basis for moving on to the next thing

> to question.

>

> Cascades of questions.

>

> It is not an inquisition, amigos. It is

> a great adventure.

>

> And for what it's worth, I say we cannot

> use statistical inference as our main

> engine of logic if we want to produce

> our own comprehensive rendition of a

> subject that has survived twenty time longer

> than the longest lived pharmaceutical

> company on earth.

>

> Don't be confused by my reference to

> Chinese medicine as a single entity.

> There is a totality, and there are

> contemporary expressions of the entire

> cultural legacy known as Chinese medicine.

>

> We're it.

>

> You are it.

>

> According to Confucian metaphysics, the

> only knowing that matters is that knowing

> that takes place when we look into our

> own hearts and act on the results.

>

> And what can you count about that?

>

> Ken

Link to comment
Share on other sites

Wainwright and All,

 

No, not talking about coyotes. Talking about what is CM. If you establish the

methods and tools and honestly believe you can add statistics, great. As yet,

the question of What is CM is still on the table. Statistics is not. Statistics

belongs to calculus, physics, Newton and Leibnitz. When Leibnitz plays mahjong

with the Yellow Emperor, I'll feel just fine about statistical analysis of CM.

 

Here's the test, Wainwright and All. Take out pencil and paper. You have one

hour. What is the integral of e to infinity of the function of Jing raised to

the Qi power. If you can answer this question thoroughly and completely, you

may then move on to apply statistics to Chinese medicine. Please turn in your

papers at the front desk. Thank you.

 

Looking forward to your thoughts,

Emmanuel Segmen

-

wainwrightchurchill

Sunday, October 26, 2003 8:41 AM

Statistical inference

 

 

Ken (, Rory, Emmanuel, et al,)

 

The statistical issue is, I feel, important for us to come to terms

with. There's no doubt that one is taking a cultural and paradigmatic

leap if one is prepared to embrace the use of statistical techniques

in the field of CM. It begs many questions, some of which we've

already alluded to. Yet, if we resolutely refuse to have anything to

do with statistics, where does that leave us? Talking about coyotes

having breakfast?

 

Wainwright

 

 

 

 

-

" kenrose2008 " <kenrose2008

Sunday, October 26, 2003 2:46 PM

Statistical inference

 

 

> Emmanule, Wainwright, and All,

>

> An early issue of CAOM, I was

> in Ocean Park so it must have been

> in 2000 or 2001, included an article

> by a mathematician named Ron Bloom.

>

> In it he addresses historical aspects

> and ideological presumptions related to

> statistical inference.

>

> I recommend everyone take a look at it.

> It's an odd little piece, and I took heat from my own

> editorial committee (that's what

> they were paid to do, but there was no

> pay). But I wanted to include it

> into what I had envisioned as an

> ongoing discussion about the foundations

> on which much clinical research is

> still based.

>

> The idea that everything that matters

> can be counted up, and that only those

> things that can be counted matter is

> nothing more and nothing less than a

> prejudice based on philosophical assumptions.

>

> I'm neither a philosopher nor a historian.

> So if the details matter to anyone, someone

> else will have to spell 'em out.

>

> But there is no reason on earth that

> any mind should be wedded to the notion

> that mathematics is the only method or

> only valid method of knowledge.

>

> The ancient Greeks recognized at least

> three approaches to the use of the

> cognitive hardware and software, and

> mathein is just one. (The others are

> gnose, and skene.) That statistics

> rose from the latter 19th throughout the 20th

> centuries to become the standard of

> proof about complex systems like human

> beings engaged in clinical studies is

> a kind of magnificent exuberance of the

> will of the new power elite that evolved

> in that time to take the place of the

> deposed despots of imperial times.

>

> It's really a big mistake to forget

> or ignore that medicine is human action

> and subject to the same laws and principles

> and vagaries of human strengths and weaknesses

> as all others.

>

> The mathematicians and scientific researchers

> who work for pharmaceutical companies just

> naturally devise ways of testing and demonstrating

> the safety and efficacy of their sponsors' products

> that will enable the cash flow of all to continue

> and thrive.

>

> From a businessman's point of view, that is the

> real curative effect of any medicine. It cures

> that deep aching emptiness in the bottom of

> the bank account.

>

> So we wind up with a method of demonstrating

> the efficacy of medical treatment that was

> developed to calculate the effects of fertilizer

> on test fields in England somewhere. But if you

> look at the nameplates on the factories and the

> logos on the brand name medicines, you'll discover

> this really terrible and terribly funny fact.

> Same folks making the deposits, by and large.

>

> And you discover how truly effective those

> medicines are. They keep alive a whole corporate

> culture for longer than the indivdual humans

> that create and sustain it.

>

> This is not a dismissal of the whole enterprise.

> I'm merely trying in dramatic fashion to

> bring to light this odd fact that a method

> developed to test fertilizer, statistical

> inference, the idea that you can count up

> the important variables and nail them down

> quite neatly and cleanly...why should that

> method be applicable to something as messy

> as a human body?

>

> Just now through the silence of an early

> morning here on Thunder Mountain, come

> the sounds of a pack of coyotes killing

> their breakfast down by Matanzas Creek.

> Matanzas means " slaughter " and even the

> animals here seem to be able to read

> the invisible writing that hangs heavily

> in the air.

>

> Or more accurately, the earlier inhabitants

> who coined the words thunder and slaughter

> to describe the local landscape were able

> to read the same messges that the animals

> and the plants and the earth itself speak

> day by day as the eons pass.

>

> How many white people I wonder have wandered

> through these places without knowing the

> meaning of where they were?

>

> The truth is that even those small patches

> of ground that those Victorian scientists

> sweat over to demonstrate the efficacy of

> one kind of fertilizer over another...but

> who ever stopped to question the whole

> strategy of forcing nature's body to

> produce more than it should...even those

> tiny well defined squares of the vast

> and chaotic earth were too complex to be

> contained in the descriptions of stupid

> white men who could not understand what

> the world was saying.

>

> And still can't!

>

> Rory asked does it make Unschuld's work

> suspect that he relies on some sort of

> statistical inference?

>

> Absolutely.

>

> But his work was suspect long before that.

>

> Everyone's work is suspect. Rory's work.

> My work.

>

> There are always lots of things that I

> want to question Paul about, and that's

> why I place such a high value on these

> opportunities that arise to spend time

> with him. Once again, if at all possible

> people should come and hear him at PCOM.

> Just the idea of formulating a question

> to ask him about these things is a

> worthwhile mental exercise.

>

> And I think we should work together to

> provide more opportunities for him to

> speak and interact with students and

> practitioners and especially those involved

> in education.

>

> Everything must be questioned until you

> are satisfied that you have a personal

> basis for moving on to the next thing

> to question.

>

> Cascades of questions.

>

> It is not an inquisition, amigos. It is

> a great adventure.

>

> And for what it's worth, I say we cannot

> use statistical inference as our main

> engine of logic if we want to produce

> our own comprehensive rendition of a

> subject that has survived twenty time longer

> than the longest lived pharmaceutical

> company on earth.

>

> Don't be confused by my reference to

> Chinese medicine as a single entity.

> There is a totality, and there are

> contemporary expressions of the entire

> cultural legacy known as Chinese medicine.

>

> We're it.

>

> You are it.

>

> According to Confucian metaphysics, the

> only knowing that matters is that knowing

> that takes place when we look into our

> own hearts and act on the results.

>

> And what can you count about that?

>

> Ken

 

 

 

Link to comment
Share on other sites

This is got to be the most absurd issue to deal with on this list. There's 800+

people on this list. Everyone who has had 3 semesters of calculus (not pre-med.

calculus) plus a semester of numerical analysis, please raise your hand. Ah ha!

Just as I thought. Okay, how many people here can calculate the area under a

curve? No, you can not peek in your calculus book on that shelf. Oh ... you

mean none of you have calculus books on your shelf? What a minute. Are there

faculty on this list who teach graduate students who then obtain masters of

science degrees? Can a student be accepted to a graduate science program who

can't do calculus and thus who can't do statistics on behalf of their own

research?

 

Does this discussion have any relevance whatsoever, if only very few people in

any of the CM programs around the country have any clue as to what statistics

is? Z'ev, do your students perform statistical analyses on anything?

 

Can any one here apply integral or differential calculus with regard to any

single idea in Chinese medicine?

 

If yes, I'd like examples. If no, then let's move on.

 

>>>Emmanuel this is just an argument to change school requirements, but at the

same time that is why one can hire a statistician

Alon

 

 

Link to comment
Share on other sites

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

meridian or another?

>>>>>But you can calculate adds and outcomes. You can compare a monkey doing

acupuncture and an acupuncturist, look at the outcomes, even by let say a

neutral CM evaluator that would collect CM information, and then come to

conclusions.

If we cant define what we try to do asking somebody for money, and define what

he/she could expect,and then prove it than this is unethical behavior

Alon

 

 

Link to comment
Share on other sites

By this same reasoning, an MD who has been through a weekend CM course can hire

an acupuncturist drone to stick needles into people based entirely on the clear

CM knowledge of the MD. Do you see what I'm saying here?

 

>>>If the MD is ethical he should understand it before trying to study,

practice, or hire somebody else to perform it for him

Alon

 

 

Link to comment
Share on other sites

Alon,

 

If you calculate the outcome based on measurable Western medical signs, you are

doing Western science. How shall I measure increased spleen qi? How shall I

shall I standardize CM diagnostic features? How will I know the outcome? In

terms of what? If you tell me what happened from the perspective of Chinese

medicine, can you tell me the units of the outcome? Please present a clear

example of statistical applications. I have seen thousands of papers on CM on

Medline and in journals. I have never seen statistics applied to CM diagnostic

features.

 

Emmanuel Segmen

-

Alon Marcus

Sunday, October 26, 2003 10:57 PM

Re: Statistical inference

 

 

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

meridian or another?

>>>>>But you can calculate adds and outcomes. You can compare a monkey doing

acupuncture and an acupuncturist, look at the outcomes, even by let say a

neutral CM evaluator that would collect CM information, and then come to

conclusions.

If we cant define what we try to do asking somebody for money, and define what

he/she could expect,and then prove it than this is unethical behavior

Alon

 

Link to comment
Share on other sites

Emmanuel wrote: By this same reasoning, an MD who has been through a weekend CM

course can hire an acupuncturist drone to stick needles into people based

entirely on the clear CM knowledge of the MD. Do you see what I'm saying here?

 

>>>If the MD is ethical he should understand it before trying to study,

practice, or hire somebody else to perform it for him

Alon

 

Alon,

 

Then you've answered my point as I would have. If you were ethical, you would

understand statistics well enough yourself in order to apply them yourself. You

would not " hire " a statistician. Also with your clear understanding of

statistics, you could decide for yourself how to measure CM qualities in units

to which a calculus can be applied. You would not need my understanding of

statistics. However, I would feel free to challenge you if I felt you had not

applied a calculus to units of CM qualities. Okay with you?

 

Emmanuel Segmen

 

 

 

Link to comment
Share on other sites

Why try?

 

As I understand it, CM uses a qualitative approach instead of a

quantitative approach of WM.

 

Doesn't that settle it? Doesn't this exclude the use of statistics in

CM?

 

Alwin

 

 

--- " Emmanuel Segmen " wrote:

> 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?

Link to comment
Share on other sites

If you were ethical, you would understand statistics well enough yourself in

order to apply them yourself. You would not " hire " a statistician.

>>>>Now that is bunk. Most western studies utilize professional statisticians.

The physicians in complex multidimensional models for example do not understand

the statistics. Not all studies are done with simple two by two tables

Alon

 

 

 

Link to comment
Share on other sites

Alwin,

 

In my humble science instructor's opinion .... yes. CM to my limited knowledge

has not presented a numerically quantitative approach ... except where it was

applied to WM disease states and WM made quantitative evaluations.

 

Emmanuel Segmen

-

Alwin van Egmond

Monday, October 27, 2003 12:55 AM

Re: Statistical inference

 

 

Why try?

 

As I understand it, CM uses a qualitative approach instead of a

quantitative approach of WM.

 

Doesn't that settle it? Doesn't this exclude the use of statistics in

CM?

 

Alwin

 

 

--- " Emmanuel Segmen " wrote:

> 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?

 

 

 

 

Link to comment
Share on other sites

If you were ethical, you would understand statistics well enough yourself in

order to apply them yourself. You would not " hire " a statistician.

>>>>Now that is bunk. Most western studies utilize professional statisticians.

The physicians in complex multidimensional models for example do not understand

the statistics. Not all studies are done with simple two by two tables

Alon

 

Alon,

 

With all due respect, I can tell you as a professional scientist that your

position here is in fact absolute and unabashed " bunk " ... to use your term.

Sounds like the former Saturday Night Live, doesn't it. Feel free to call me a

slut next. But I digress, no PhD in science who directs peer-reviewed studies

is even remotely as ignorant as you regarding statistics. All such professional

scientists are trained in statistics from undergraduate school. That's why MDs

are not allowed in professional laboratories most of the time. You are

absolutely correct in your assessment regarding MDs not understanding

statistics. They mostly take a watered down two semester training in calculus

in undergraduate school and never look at statistics again for the rest of their

lives. Is it any wonder that clincal science and basic science are worlds

apart? Alon, I mean no disrespect when I say that you are completely incapable

of designing a peer reviewed science experiment or even knowing what the results

mean. You are no doubt a fine clinician. Asking for statistical studies of CM

is like " fighting for peace " ... it's an oxymoron. You have no idea what you

are asking for, and short of having you as my science student for a few years, I

have no way of telling you what you can't now see.

 

Respectfully (if somewhat slutty),

Emmanuel Segmen

 

 

 

Link to comment
Share on other sites

But I digress, no PhD in science who directs peer-reviewed studies is even

remotely as ignorant as you regarding statistics. All such professional

scientists are trained in statistics from undergraduate school. That's why MDs

are not allowed in professional laboratories most of the time.

>>>>Well we are talking about medical research are we not? Somehow MDs publish

studies daily, the statistics of which they do not understand, but they can

understand P values, reliability, validity, sensitivity and specificity,

likelihood ratios etc.

I see no reason we should not understand at the same level

Alon

 

 

Link to comment
Share on other sites

Join the conversation

You are posting as a guest. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...