I feel like an old alcoholic who has fervently sworn off the bottle but can't resist taking one more drink!
But let me take one more try, making it as simple as I can.
1. James Bacque in his Other Losses
(1991 revised edition, Pima Press) asserts that the German POWs, held in US prison camps in Europe, were:
....exposed to conditions that killed them at the rate of over 30% per year.
2. The sole source proffered for this assertion is Bacque's Appendix 2.
3. Appendix 2 contains facsimiles of two Tables, IX and X, taken from the typed manuscript of "Medical History, European Theater of Operations" to be found in the National Archives. I have no question but that the Tables are authentic. They are based on a survey by the U.S. Army Medical Corps in the European Theater in may-June 1945.
4. As indicated in my prior post, the caption to Table IX states that it is a comparison of (a) the number of [hospital] admissions and (b) death rates per 1000 per annum for POWs in the ASCZ, with ETO (less UK) Troops [ie essentially US Troops]. Table X is a listing by causes and number of deaths of the "Chief Causes of Death Due to Disease in ABSZ Prisoner of War Enclosure For Six Week Period Ending 15 June 1945".
5.Table IX shows a number of 2,754 POW deaths from disease and a 34.2 death rate per 1000. In light of Table IX's caption, one would think that the 34.2 death rate per 1000 was a per annum rate. The caption says so.
Bob Lembke, however, disputes this and insists that the stated rate was only for the 6 weeks reviewed, which, if true, would elevate the annual death rate to 296.5 per 1000, or 29.65%. ( 52/6 = 8.67 x34.2 = 296.5) BUT:
(a) Bracque himself states that the rates per 1000 shown in Table IX are annual
6 week rates:
.....the rate per thousand for hospital admissions for injuries was 468 per year.
[my emphasis]Appendix 2, p.211.
(b) If the rates shown are on a 6 week basis, then, for example, the POW rate of 5,003 per 1000 for total hospital admissions would be 43,376 per 1000 per annum (52/6 = 8.67 x 5,003 = 43,376), or 43+ admissions for each POW per year, or about 1 every 8-9 days!!! That stikes me as far beyond the stretch of credibility.
(c) Equally incredible would be the resulting death rate for US Troops. If the 3.8 death rate per 1000 shown for US troops excluding battle casualties were computed on a 6 weeks basis, then the total US death rate per annum would be 32.95 per 1000 per annum, or 3.3% - many times higher than the comparable figures supplied in one of David thompson's previous posts.
(d) In my own albeit limited [Iv'e only been around 75 years] experience, unless specifically otherwise indicated, "rates" when quoted are generally understood to be per annum rates. E.G. "I got a 6% interest rate on my mortgage!"
(e) Conclusion: Bob Lembke's old dog still just wont hunt.
6. One would also think from looking at Table IX that the number of hospital admissions and deaths shown are those which actually took place during the 6 week period in question. That seems to me the fair implication of its caption and the arrangement of of the table that follows. But this would demolish Bacque's assertion of an over 30% POW death rate and so he argues that no, the numbers of hospital admissions and deaths shown are not
the numbers actually experienced during the 6 week period, but rather those 6 week numbers projected forward throughout the year. So that, for example, the number of 37,713 POW hospital admissions for injuries shown on Table IX are the projected
admissions which would occur for the entire year if one were to apply the shown rate of 468 per thousand per annum to the population of POWs surveyed.
7. By cobbeling up that theory, Bacque can play mathmetical mumbo-jumbo and produce a POW death rate of over 30%. Here's how: Bacque first purports to compute the number of POWs in the study by taking his 37,713 assumed hospital admissions per year and dividing that number by the 468 per thousand annual rate stated in Table IX. This results in a computed POW study population of 80,583. Well OK, but so what? If you consider the 2,868 total number of POW deaths shown on Table IX as a projection for the total deaths for the entire year, then if the population of the study is 80,563 the annual death rate is only 3.56% (2,868/ 80,563 = 0.0356, or 35.6 per 1000, as shown on Table IX.)
8. Ah, but wait! Although according to Bacque the POW hospital admission numbers reflect projections for the entire year, the POW death numbers do not! How so? Well, in effect Bacque maintains that the "annual projected" 2,868 POW death total shown on Table IX is simply phony. He gets there by looking at Table X, which shows a total of 2,304 deaths from disease during the 6 week period from the 12 chief causes listed, and which, if you project them out for a full year, results in a total of 19,968 annual deaths from disease. That of course is much better, because against a study population of 80,563 a total of 19,968 deaths per year gives you an annual death rate of 24.78%.
9. But we are still not quite at the over 30% rate. How to get there? It's simple - dredge up a Table 23 included in a 1969 Article published in the History of Preventative Medicine in Word War II
which shows the number of POW deaths from disease during the same 6 weeks as 2,754 which when added to the 114 deaths from injury and battle casualty comes to a total death tole of 2,868. When annualized and applied to a study population base of 80,563 the result is a total death rate of 30.86% (52/6 = 8.67 x 2,868 = 24,866/ 80,563 = .3086)
10. But wait! Table IX shows the same number of POW deaths from disease as does Table 23 from the 1969 Article - doesn't that suggest that all
numbers in Table IX are actual numbers experienced for the 6 week period? Of course not, silly! That would mean that Bacque's computation of the size of the POW population underlying the study was all wrong and we couldn't get to our 30% plus death rate. We would be back to that 3.56% rate which, although pretty bad, certainly wouldn't sell any books at all. No, the explanation is simple:
The evidence is clear that the author of this History [Table IX] hid the death rate by suppressing evidence.....[He] simply reproduced the POW death figures for six weeks as if they applied to a whole year. He thus apparently reduced the death rate of 29.7 percent per year for disease to 3.42 percent. This is probably why the author of the History did not show Table X complete. Table X's true total of 2,754 would have revealed that he had deceptively used the same number in Table IX......Because in Table IX the disease death rate has been falsified, the rates for injury and battle casualty have almost certainly been falsified downwards to reduce the death rate......
It's a case of falsification, suppression of facts, cover up!!! That should surely sell some books! BUT
(a) There is nothing whatsoever in Table IX that suggests that the numbers stated are anything other than those actually experienced during the 6 week period reviewed, and certainly no hint that they might be projected annual totals. If the latter was the case, surely the author would have so indicated.
(b) If the numbers stated are annual projection of 6 week actuals, then the figures for US Troops simply make no sense. Let's test Bacque's technique for determining the size of the POW population underlying the study(see 7. above) to determine the size of the US Troops in the ETO European Theater of Operations. The US hospital admissions are shown on Table IX for injuries are 31,070; divide that by the 101 rate per thousand per annum shown and the result is a US troop count of 307,624 in the entire ETO in May - June of 1945. I don't know what the exact number actually was, but it was many times the result of applying Bacque's approach.
(c) Let's test it just once more to make sure the first wasn't a fluke. Take the 1, 162 total US Troop deaths shown and divide it by the total US Troop death rate of 4.1 per 1000 per year. Here we get an even more lidicrous result of 283, 415 for the US Troop comliment in the ETO.
(d) Bacque notices the absurd results produced by application of his methodology to the US TRoop figures, but brushes them away as either statistically unreliable or based on a different survey. And anyway, so what:
These difficulties in Army statistics are typical and usually prevent anyone from discovering the death rate in the camps.
Appendix 2, at p.211.
11. I find it passing strange for a historian to attack the credibility of the only documentary source he has for the theorem he is proposing, but I suppose there is something to the saying that a drowning man will clutch at a sword. As a lawyer, I've been caught myself a few times in desperation to find an argument to save a hopeless case. But I don't think it says much for the integrity of a historian.
13. It seems tolerably clear to me that both Bob Lembke and James Bacque in their different ways have approached Table IX "bassackwards" in the parlance often employed in this part of the world. If you just hone up and apply Occam's well known razor by simply accepting the Table for what it clearly says it is all internal inconsistencies and contradictions vanish into thin air, and one is still left with a shameful 3.56% annual POW death rate, which the Table itself deplores by pointing out that it is nearly 9 times that of US Troops. The only inconsistency is that the POW population derived by applying the various rates per 1000 per annum to the various numbers given the result in each case hovers around 700,000, and that disagrees with a POW population for the study of 70,000 which appears somewhere in the text of the study (which Bacques fails to set out.) Well, its not too hard for me to imagine a tired typist missing a zero - I made the identical mistake on one of my previous posts, and although I'm a superantiquated two finger hunt and peck typist, I at least like to think that others may also be capable of making mistakes too.
14. But then of course without allegations of a shockingly high death rate, outrageous conduct at the highest levels of government, rampant falsifications, suppression of evidence, coverups - how is a book supposed to sell?
15. The above is BORING
beyond belief, and I solemly vow that this is absolutely, definitely, finally, without reservation, cross my heart and hope to die, stick a needle in my eye, my very last post on this topic! (Unless, of course, I am once again overwhelmed by temptation.)