SEPTEMBER 1st First Monthly Service on STATISTICS, LIES AND DAMNED LIES

Postings about each monthly service includes topics discussed and an open forum to further discuss those topics.
Post Reply
tim
Posts: 78
Joined: Mon Feb 06, 2012 12:13 pm

SEPTEMBER 1st First Monthly Service on STATISTICS, LIES AND DAMNED LIES

Post by tim » Tue Sep 10, 2019 2:00 pm

September's service opened with an account of the recent appearance of FFRF attorney Andrew Seidel who was in town to discuss and promote his new book Founding Myth: Why Christian Nationalism Is Un-American. If you have seen the national motto “In God We Trust” on police cars, on your property tax bill and elsewhere you know what is going on. Christian Nationalists' agenda is to thoroughly “Christianize” the United States, allowing invidious discrimination against non-Christians, gays and others, inserting Christian slogans, doctrines and practices wherever possible. They claim anything else is “hostility” towards Christianity! They expect to successfully reverse Roe v Wade, prohibit or at least reduce the availability of many forms contraception and to generally relegate non-Christians and especially atheists to the status of second- or third-class citizens or even de facto non-citizens. Christian Nationalists consider The Handmaid's Tale not as a cautionary story but as a guidebook to the future they want.

Mr. Seidel fielded questions about the nature of religion and whether he thought “In God We Trust” would ever be replaced with a more appropriate national motto such as “E Pluribus Unum” or Ben Franklin's “Mind Your Business.” He said he thought such a change “inevitable,” because of demographics. But this won't happen – or will be precarious – without the furtherance, in the marketplace of ideas, not just of arguments against theism but of arguments for and demonstrative of a positive way forward in which religion becomes understood as something other than supernaturalism. And this, of course, is what the NTCOF promotes. A book signing by Mr. Seidel followed.

We were then reminded that September 1st was the date in 1914 that the last known passenger pigeon died at the Cincinnati Zoo. And uin 1939 Hitler invaded Poland, beginning World War II. This was the most destructive war in history with with 70-85 million casualties including the genocide of the Holocaust and 200,000 killed in the first and – so far – only wartime use of nuclear weapons by the US against Japan which finally brought hostilities to an end. Global maps were redrawn including the partition of Germany and Korea and, within another decade, Vietnam. The United Nations was a result of World War II, as was the Cold War between the USSR and the US and its allies. The conflict has had such profound effects on all the nations of the world that scholars are still studying such questions and are likely to continue doing so for many years.

Our Moment of Science considered the just-published genome-wide assay study of nearly half a million people from the UK and the US that reported on 5 genetic markers for same-sex behavior. Interestingly, none of the markers were on the sex chromosomes. But they did seem to be associated with sex hormone metabolism as well as with the sense of smell.

The study authors wrote:

“Same-sex sexual behavior is influenced by not one or a few genes but many. [There is]overlap with genetic influences on other traits ... and analysis of different aspects of sexual preference underscore its complexity and call into question the validity of bipolar continuum measures such as the Kinsey scale. … Overall, our findings suggest that the most popular measures are based on a misconception of the underlying structure of sexual orientation and may need to be rethought.”

Of course, the interest in a genetic cause of same-sex behaviors is related to the claim that such behaviors are “immoral” and “sinful” and should be criminalized as they once were or should justify invidious discrimination by zealots exercising their “religious freedom.” Now perhaps some unreasonable superstitious zealots will have their beliefs undermined if it is shown that sexual preferences are genetic such that homosexuals “can't help it.” But if it were mostly or even entirely a “nurture” rather than a “nature” effect, how would that change the unchosen character of it? Or might it be “chosen” in some sense if it's influenced by cultural/social acceptance? I suspect most of us here don't feel that we have chosen not to want to eat insects but if we had been raised in parts of the world where this is common we probably would include this in our diets.

So we may well ask: “So what?” If people did just choose to engage in same-sex acts would that make them more or less “moral?” As Freethinkers the answer is almost certainly “no.” But that so many suppose otherwise exemplifies the systemic and continuing outrages that have their source in supernatural doctrines and dogmas. Why should it matter whether same-sex behaviors have a genetic cause or influence or not? We still have no idea whether such things as fingerprints or handedness are genetically determined.

The remainder of the service considered our main topic of statistics, an important – arguably the most important – tool of modern science. This method of analysis relates to both the collection and analysis of data as well as its organization and presentation. The word comes from the same root from which we get the word “state,” reflecting governments' long interest in information about the people and places they rule over. Ominously, the Nazis and other oppressive governments made and continue to make use of statistics to maintain their power.

Despite the importance, power and ubiquity of statistics, it also has many pitfalls. These include problems that arise from small and nonrepresentative sample sizes, biases in the way measurements are taken, the stopping of data collection as soon as a “statistically significant” result is found, and “p-hacking.”

“P” is a statistical measure of how likely it iw that an observed correlation is due to pure chance. Generally, researchers look for p>.05, or the likelihood of a finding being due to chance is less than 5%. But finding such a legitimate measure of “statistically significance” depends on looking for one particular correlation in the data. If one looks for, say, 100 different correlations then by sheer chance several of the, - on average almost five of them – will be found to have p<.05. This, in fact, appears to be the basis of claims that a “nutritional supplement” of a substance first identified in jellyfish has been “scientifically proven” to improve memory. See https://sciencebasedmedicine.org/prevag ... p-hacking/

We discussed work that suggests that many – even most! – published reports suffer from such flaws. Come copies of this article were made available: https://journals.plos.org/plosmedicine/ ... ed.0020124

We reviewed ways that statisical information can – and is – deceptively presented and showed examples of this, including claims of effects that are statistically significant but not practically significant.

Three sources for further reading were suggested:


https://www.amazon.com/How-Lie-Statisti ... exaschurch by journalist Darell Huff is a 1954 classic that was a best-seller in its time.
https://www.amazon.com/Statistics-Done- ... exaschurch is a better and more recent resource by statistician Alex Reinhart.
https://www.amazon.com/Damned-Lies-Stat ... exaschurch is also a more recent work. Its author, Joel Best, a Professor or Sociology and Criminal Justice at the University of Delaware.

Post Reply