|
The MoA Week In Review – Open Thread 2020-18
Last week's posts at Moon of Alabama:
The U.S. decided to run along the red curve while China, with less than 100 new cases today, has moved from the red to the blue curve.
> The effective reproductive number [in Wuhan] dropped from 3.86 (95% credible interval 3.74 to 3.97) before interventions to 0.32 (0.28 to 0.37) post interventions. <
bigger
Other issues:
Inside the World Uyghur Congress: The US-backed right-wing regime change network seeking the ‘fall of China’ – Grayzone
Erik Prince Recruits Ex-Spies to Help Infiltrate Liberal Groups – NY Times
Thousands of U.S. Areas Afflicted with Lead Poisoning beyond Flint's – Scientific American
SUVs and Pickup Trucks Are Now Too Big For Our Already Gigantic Garages – Vice
Use as open thread …
I did some research on lead poisoning two years ago (privately distributed, as my research is as much an attack on current epidemiology as it is on e.g. the beliefs of the race realist movement), and have a brief summary that touches on the US problem.
Firstly, the US practice is to report *geometric means* rather than *arithmetic means* of blood lead levels. This is a scam, that serves to hide the scale of poisoning. To avoid the scam from becoming too obvious, zero readings of blood lead level are recorded by the US government as 6ug/l (5 in some years). Geometric means are only justified if the dose response is logarithmic.
The only studies of which I am aware, that claim a logarithmic dose response of IQ to lead are European blood lead at birth versus IQ, where the logarithm can be an artefact of biological half-life of lead in blood (35 days) after regular exposure ends (beginning of maternal leave, during an era of leaded petrol), given a *linear* underlying dose response.
There is evidence that much of the lead poisoning in the US is sporadic (mean interpoisoning time per Poissonian process much greater than biological half-life)—look at the distribution of blood lead levels in NHANES III and later—high blood lead tails closer to truncated hyperbolic rather than Gaussian). A simple Monte Carlo estimate assuming comparable dose from frequent as from sporadic poisoning, coupled with use of “confirmatory” blood lead test after 12 weeks (iirc) gave a factor of ~8 ratio between geometric and arithmetic means, where the frequent dosing by itself gave a ratio of 1.2. Thus the mean dose effect would be grossly underestimated using the geometric mean dose after the phase-out of leaded petrol versus before.
Secondly, there is evidence that uptake of lead per unit exposure is a function of genes. This leads both to lead poisoning being a “black problem,” and to gross misestimates of dose response.
Sub-Saharan Africa largely skipped the bronze age. At the oldest iron workshop sites, i.e. in Nigeria and southern Africa (see Mining and metallurgy in Negro Africa, George Bantam books), the tool inventory for working iron is Stone age rather than bronze age. Many African ethnic groups did later work copper alloys, for status objects and cult objects, but useful tools were usually made of iron.
As lead is always present in soft metal ores (e.g. gold, silver, copper and tin), bronze ages always lead to mass lead poisoning. As adult lead poisoning leads to fertility problems, a genetic selection is expected during a bronze age, for reduced lead uptake. Also note that empires tend to show up during bronze ages, or where gold or silver is worked (Egypt, Ashanti, South and Central American empires, Mongol empire). Presumably lead-induced popular thuggery leads to empire formation.
As lead exposure in antiquity would be minimal for Sub-Saharan Africans, higher blood lead levels and crime rates would be due to lower evolved resistance to lead uptake. Genes that are good candidates for investigation for lead uptake variation include those identified early on by Davide Piffer in his polygenic IQ models.
IQ is seems as primarily caused by genes in twin studies so parental IQ is included in studies on lead, but the measurement of dose response of lead on IQ is confounded when parental IQ is included in the dose response model, without including the parents’ exposure to lead in their infancy and prenatally. Where a linear regression is used in a model where the underlying dose response is linear, the apparent dose response (\beta_{apparent}) will be related to the underlying dose response (\beta_{actual}) by
\beta_{apparent}=\beta_{actual} (1-R_{Pb:parent,Pb:child} \sigma_{Pb:parent}/\sigma_{Pb:child}) +
\sigma_{\delta IQ}/\sigma_{Pb:child} R_{Pb:child,\delta IQ}
Where R_{Pb:parent,Pb:child} is the correlation between the dose (or proxy) that the parents received in their childhood to the dose (or proxy) received by the children being measured,
\sigma_{Pb:parent} and \sigma_{Pb:child} are the standard deviations of parental and child dose (or proxy),
\sigma_{\delta IQ} is the standard deviation of IQ variation not caused by lead, and
R_{Pb:child,\delta IQ} is the correlation between the non-lead variation in IQ (e.g. gene fluctuation relative to parental mean per individual) and lead dose or proxy (e.g. gene environment correlation, see comment below).
When a study is done that bins individuals according to apparent dose received, individuals with very low dose are probably less exposed compared to their parents, relative to medium exposure individuals. Thus one may find the various funny curves of apparent dose response as an artefact of not including parental dose.
It is not clear how one might distinguish genetic variation in uptake from gene-environment correlation, without doing a controlled uptake experiment.
Canada also has a severe water lead problem, e.g. Montreal, Regina, Moose Jaw and Prince George…
Posted by: Johan Meyer (2) | Mar 9 2020 2:23 utc | 52
|
