Marvin said-
When composing my extrapolation I considered the issue of teens and a higher incidence ratio unrelated to refusal of blood product (e.g., Maori ethnicity).
Here’s why I opted not to mathematically account for the issue of teen deaths due to red cell transfusion refusal:
1. I don’t know that there is anything to account for because I don’t know that any of the teens are among the 21 deaths in the 103 JW patients.
Within the framework of science, remember that the rule is, "an absence of evidence is not evidence of absense", since that is what is called, "making an assumption".
Again, your saying "I don't know" as if it's a defense (rather than an admission) is an argument from personal ignorance, and exactly analogous to the believer who says, "Well, I don't know HOW to account for a gap in my personal understanding of how the World actually operates, so I'm assuming God Dun It!" Unfortunatlely, that kind of thinking doesn't fly within a scientific setting, since an absense of evidence or one's personal ignorance does NOT give one carte blanche permission to dismiss all valid criticism(s); thus, anytime you say, "I don't know", it MUST be disclosed as a 'known unknown' if you can't correct the problem IF you want to be honest in your methodology with your readers. "Likely" and "surely" are called ASSUMPTIONS, and are four-letter words in scientific endeavors.
BTW, I'm not quoting any of post-hoc rationalizations (which I don't find very compelling, filled with unproven assumptions), although it's fine to include your reasons (or 'excuses', depending on one's knowledge of the scientific method) in your article to talk your way out of the logistical log-jam of a demonstrable flaw in your extrapolation attempts. It would be advisable to mention the existence of these types of issues in your discussion, as they aren't trivial to anyone who actually KNOWS that studies can be designed to basically show virtually anything, esp when relying on a SINGLE study in NZ (as Mark Twain said, "There are lies, damned lies, and statistics!").
BTW, that's exactly WHY statisticians use the method of meta-analysis: they attempt to compile the results of MULTIPLE STUDIES that are all looking at the same issue, but with different researchers looking at the same question in order to reduce the statistical effects of errors that may exist within a single study. The topic of how to perform a meta-analysis is WAY over my head, as I'm not a biostat (just a retired provider who's worked with a few in the past).
“Not only does the sample population have a higher rate of anemia, but you admit that in the sub-population it has a higher mortality rate if left untreated, and those two factors explain WHY the study is unfit for extrapolation purposes: it's not truly a representative sample of the entire population Worldwide.”
I do not compare results of the Beliaev study with the entire population worldwide. I compare the result of the Beliaev study with the population of JWs worldwide. JW make up the “the sample population” you speak of and the “sub-population” you speak of, and it's mortality among the JW population that is at issue. If I’ve misunderstood you please feel free to restate yourself or otherwise clarify.
I know that, and assumed you'd know that the "population" you're trying to extrapolate to IS the JW Worldwide population: OF COURSE it's not the entire population of BILLIONS of humans who live on Planet Earth. Odd that you'd think I wouldn't know that, having taken a few years of stats more than you?
Anyway, that actually reminds me of another problem:
If you want to claim to generate a "conservative" figure, you'd also need account for the sub-population of those JWs under 18 who are included in the total head count (7.4 mil, or whatever it is) to remove them from the calculations, since they are generally protected as minors in most Countries. I don't know if JWs provide that a figure for under-18s who are included in the total, but that's another factor which you should account for, since to fail to do so would prevent a claim of being "conservative" in attempts to extrapolate from a NZ sub-population (of 12.7k to 7.4 mil, where NZ JWs represent less than 2 out of 1000 JWs worldwide, and those who died in the study represent 1 out of 353,400 of JWs Worldwide).
If ages can be obtained from those patient's charts (an easy solution to the predicament, digging up 21 charts and noting their age at the time they refused blood), then Beliaev's study SHOULD be used as the basis for legislative actions to try and raise the age of consent from 16 to align with the rest of the Western World. However, that change would likely be contested by abortion rights activists, as it would infringe on the right of a 16 yr old female to seek an abortion without their parent's knowledge and/or consent; the same right that allows a 16 y.o. to die as a Martyr for Jehovah also allows a 16 y.o. to seek an abortion without their parents consent! So ironically, those JWs who demand to exercise their right to to die also entails a cost of increasing deaths due to abortion.
Adam