250,000 Jehovah's Witnesses have died refusing blood (page 28)

steve2

Marvin: 3.3 cannot be a coefficient of the data study because a coefficient ranges only from 0 to 1.

Marvin Shilmer

-

“When he says "JWs", he's referring to those JWs who live within the area of the four hospitals, even if they weren't actually participants in his study.”

As a general rule I’d have to disagree with that statement. Nevertheless, in this case we have every reason to think severely anemic JWs refusing blood in New Zealand have a reduced mortality compared to just about anywhere else in the world. Yet my conclusions do not take advantage of this to statistically increase mortality. We also have no reason to think the statistical incidents of severe mortality are anyways increased in the New Zealand study at issue (that are left unaccounted for). Moreover, my conclusions assume there was not as much as a single death at any other trauma center in the given region of a severely anemic JW refusing blood. Again this is to avoid anything that would statistically increase extrapolated mortality values.

Marvin Shilmer

besty

@marvin

besty,

An annual sample size of 12,700 with an annual population of 3.3 (that would be 33 over 10 years) presents a ratio of 3848-to-1.

In year 1998 there were 5,544,059 JWs. You do the math and tell readers what a ratio of 3848-to-1 gives us for the year 1998 alone.

Can you do that, and put this hard number in writing for readers to watch your math?

Marvin Shilmer

how about you go first and tell me what I am missing, instead of moving the goalposts. As a reminder:

- Beliaev and Marvin agree 10x mortality for anemic JWs

- 0.2/100,000 is the accepted mortality rate for anemic non-JW's

- 2/100,000 mortality would be a reasonable assumption for anemic JW's

What am I missing?

Marvin Shilmer

-

“Did I zone out for a minute, where did the 3.3 come from and how is that a 'population'?

“Are you sure you don't have population and sample size back to front?”

Statistically a sample size is the entire number you’re looking at (in this case comparing something to) and the population of that sample size is the portion of the sample that meets whatever criterion is at issue. In this case the sample size is the annual population of JWs in New Zealand (avg. 12,700 for period 1998-2007) and the 3.3 is the annualized figure of statistical deaths attributed to severely anemic JWs refusing blood among the annualized 12,700 JWs in New Zealand.

Marvin Shilmer

besty

@marvin - you said:

I appreciate you persisting in the particular you do above. It’s made me revisit and revisit and revisit the calculator used in my assessment of this information, which is a good thing. Just now I noticed an algorithm that’s not as labeled. I’m going to look into why this is the case to see if I mislabeled or miscalculated, or perhaps something else. I’ll get back as soon as this reassessment is performed.

Thanks again for pushing on substance

let me know when you have 'looked into this'

Marvin Shilmer

-

“If my numbers are right then you are claiming that over 20% of JWs die because of refusing blood transfusions.

“Is this correct?”

That is an absurdity and is nothing I’ve claimed.

“Says the people who signed the study as " © 2011 International Society of Blood Transfusion”

That presumes those who use blood transfusion advocate the practice based on preference rather than sound science.

Marvin Shilmer

-

“Marvin: 3.3 cannot be a coefficient of the data study because a coefficient ranges only from 0 to 1.”

My usage of coefficient was colloquial meaning in this case it’s a property that’s inherent to the data set.

Marvin Shilmer

adamah

Simon said-

Anytime I see stats along the lines of "every 5 minutes someone ..." I always remember not to assume it's different people. It could be just some really unlucky guy!

Yup, and that's why I took efforts to reinforce that very point of acccounting for single patients who are known to need multiple transfusions (in red):

What I found on the Red Cross site was a claim that every two seconds someone in the U.S. will need blood.

That works out to 16 Mil people per year (assuming no one goes back for 2nds), which given 300 Mil, works out to only 5%, max. (As the other bullets point out, patients with sickle-cell and leukemia will require many transfusions during the course of their treatment, sometimes daily).

I thus likely erred on the side of over-estimating, since the actual figure is likely lower (making the 20% figure all the more questionable).

But as you say, the form of the Red Cross "every 2 seconds" stat is rather questionable, on it's surface, no doubt sacrificing accuracy in the name of making it memorable (AKA trying to create a meme).

BTW: Your blog post on Noah / blood was a great read and very entertaining!

Thanks! Glad you found it entertaining and (hopefully) informative, as it's always challenging to juggle readability and substance when presenting material (which is the same point being discussed in this thread).

Marvin said-

You keep saying things like that yet you have failed to express any reason why this data set is not useful for conservative and/or liberal extrapolations based on certain criterion, such as life expectancy in New Zealand compared to the rest of the world, state of healthcare in New Zealand compared to the rest of the world, etc.

Do we, for instance, have any reason to think there are more or less cases of severe anemia in New Zealand compared to the rest of the world (that are not accounted for in the study)? If so, what are those reasons? My review and conclusions assume the best case scenario in terms of mortality.

Do we, for instance, have any reason to think outcomes for treating severe anemia in New Zealand are better or worse compared to the rest of the world? If so, what are those reasons? My review and conclusions assume the best case scenario in terms of mortality.

You’ve made your assertions above. Now prove it.

Nope, that's not how it works. Expressions of doubt doesn't require proof: instead, the one who makes the CLAIM needs to present their evidence as proof, or the claim fails. The burden of proof lies on YOU to explain why Worldwide extrapolation is supported by the data, and as I've said before, you want to extrapolate the conclusions of the NZ study somewhere even the author never intended for it to go, in the first place (hence the comment about 'fools rushing in where angels fear to tread').

That stated, whaddaya got?

(I was trying to be polite by asking if you contacted the author to ask him his opinion of extrapolating his study Worldwide, but it was only me trying to avoid telling you that you're dead-wrong on this issue.)

Oh, on this:

My use of these particular statistics is done to keep the numbers as conservative as possible!

But apparently not conservative enough, since you fail to grasp that you cannot draw ANY scientifically-valid conclusion from the NZ study's data to go Worldwide. You just can't get there (Worldwide) from here (4 hospitals in NZ), on the basis of a single regional study.... NOW, you CAN use the data to make pseudo-scientific claims if you like, but don't be surprised when/if you get called out for engaging in WTesque tactics.

The solution is pretty simple, really: you only need to note the issue and add a few suitable disclaimers (eg "if the results of the NZ study can be applied to the rest of the JWs Worldwide", etc); then you can make the claim, but it's dishonest to do it as anything but an unsupported extrapolation of a single study. It's all about how you word it (and what 'weasel words' you use).

Adam

Marvin Shilmer

-

“how about you go first and tell me what I am missing, instead of moving the goalposts. As a reminder:

“- Beliaev and Marvin agree 10x mortality for anemic JWs

“- 0.2/100,000 is the accepted mortality rate for anemic non-JW's

“- 2/100,000 mortality would be a reasonable assumption for anemic JW's

“What am I missing?”

besty,

I answer for me. Not for you.

What I’ve written and presented is not based on a 10x mortality. The 10x mortality is a differential between mortality of severely anemic patients who accept blood and severely anemic patients who refuse blood. It is not a factor of how many JWs suffer death the result of severe anemia and blood refusal.

To use the given statistics to extrapolate how many JWs suffer death the result of severe anemia and blood refusal I’ve used the statistical value of deaths over the norm among a given population of JWs (33 over 10 years) to form what amounts to a ratio of deaths over the norm suffered by JWs suffering severe anemia. The ratio is 3848-to-1.

Again:

In year 1998 there were 5,544,059 JWs. You do the math and tell readers what a ratio of 3848-to-1 gives us for the year 1998 alone.

Can you do that, and put this hard number in writing for readers to watch your math?

Marvin Shilmer

250,000 Jehovah's Witnesses have died refusing blood

Share this