evolve2017

I'm unfortunately on mobile and can't write much - however, my understanding is that there is still some debate about the benefits of medication vs sanitation in overall mortality in Western countries. While this doesn't necessarily hold true for children, I thought I'd link to a discussion of this debate. (Can't find the reference I wanted...)

Here's a figure for death from infectious disease by the CDC - note that antibiotics don't appreciably change the rate in decline

https://www.cdc.gov/mmwr/preview/mmwrhtml/mm4829a1.htm#fig1

just a random find - this paper from 1902 mentions the debate, too, with some descriptions of 'natural experiments' https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2401255/?page=1

Also, I am not saying that antibiotics and medications don't work - just that the story actually is more interesting than one might initially guess.

Lastly, these data discuss Western populations - I've heard the opposite may be true in poor countries.

This is actually a great question - take a look at this discussion from 2015.

This isn't my area of expertise, but my understanding is that elephants have many more copies of TP53, a gene that plays a critical role in regulating the cell cycle (and thus uncontrolled growth that may occur in cancer).

One of the linked papers also states that there was no correlation between body size and cancer in their experiments.

https://www.nature.com/news/how-elephants-avoid-cancer-1.185...

I think we're drifting away from the author's article.

His statement about 'lowering the bar' with respect to decreasing false negatives directly implies a relationship between candidate quality and the false negative rate.

I can't comment on your hiring analogy, as you have more experience in this area than I do. My responses were only to highlight the subtle negative character of the author's statement.

I see what you're saying about the bias-awareness intervention.

For your last sentence, I did not write, and did not mean to imply anything about the number of false-positive (FP) 'diversity' candidates vs FP non-'diversity' candidates. It's true that the author hints at this comparison.

I meant to convey that, per the author's statement about 'lowering the bar', Google's practices increase the FP rate in 'diversity' candidates as compared to what would have happened without the practices.

From this perspective, the author is stating that the 'diversity' hires have a higher FP rate than might have occurred otherwise. I think this conclusion - I think it's a fair one, though we may disagree - is what people find somewhat problematic.

Thanks for your response!

This isn't a pharmaceutical trial but there's no reason why we can't follow along with the author's analogy of 'false negatives'.

Here are his words from 'The Harm of Google's Biases':

    I strongly believe in gender and racial diversity, and 
    I think we should strive for more. However, to achieve 
    a more equal gender and race representation, Google has 
    created several discriminatory practices:

He goes on to list his examples, which include the following:

    Hiring practices which can effectively lower the bar 
    for “diversity” candidates by decreasing the false 
    negative rate

I think you may have misread his position earlier; I noted that you misspoke when talking about this particular statement (i.e., whether the false-negative rate applied to 'diversity' vs 'non-diversity' candidates)

EDIT:

Didn't see your statistical comment. Unfortunately, I can't think of a case where 'adjusting a bar' to decrease the false-negative rate does not neceessarily increase the false-positive rate.

The fact that the author uses the phrase 'lowering the bar' suggests that he is also thinking of this statistical requirement.

Just for people who might come along later:

It looks like the author is talking about setting a threshold, as in the context of creating a test. Please have a look at [1].

In this setting, it is not possible to decrease the false-negative rate without increasing the false-positive rate. I think the author is phrasing things this way simply as a matter of rhetoric, not to express the point you are suggesting.

1. http://www.prolekare.cz/dbpic/jp_5403_f_21-x1000_1600

Totally agree! Especially as a PhD student in genetics.

Just wanted to put out the point that race as we understand it does not map cleanly onto distinct phenotypes.

I replied elsewhere but in case you didn't see it:

It looks like the author is talking about setting a threshold, as in the context of creating a test. Please have a look at [1].

In this setting, it is not possible to decrease the false-negative rate without increasing the false-positive rate.

I think the author is phrasing things this way simply as a matter of rhetoric, but his argument clearly suggests that less qualified 'diversity' candidates are being hired.

1. http://www.prolekare.cz/dbpic/jp_5403_f_21-x1000_1600

Please correct me if I err - it's been some time since my college statistics class.

It looks like the author is talking about setting a threshold, as in the context of creating a test. Please have a look at [1].

In this setting, it is not possible to decrease the false-negative rate without increasing the false-positive rate.

I think the author is phrasing things this way simply as a matter of rhetoric, not to express the point you are suggesting.

1. http://www.prolekare.cz/dbpic/jp_5403_f_21-x1000_1600

Thanks for the ad hominem attack :)

If you look through my post history, you'll see that I made this account to talk about another article about a week ago.

EDIT:

I can't respond to you, unfortunately.

I'm sorry we've resorted to name-calling - however, I stand by my earlier statement. Irrespective of your identity, there has been a noticeable trend of discussions about 'reverse racism'.

I have not referred to you with a pronoun nor made any mention of your race/sex/anything else. I apologize if you feel that I did this.

It seems like you're open to learning more about genetics (I think...) so I'll share some interesting findings that don't really fit with the popular conception of race.

It turns out that the most genetic variability lies between... black people who live in West Africa.

While we can use 'junk DNA' to classify people somewhat well, that doesn't tell us much about what makes a person a person.

This is not a controversial view - here's a discussion of it for example: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1893020/

There's a lot to be said about the current trend of privileged (read Western, perhaps wealthy, perhaps white, perhaps male) people to bemoan whatever concessions are made to the minority group but I'll leave that aside.

I believe that everyone should have a fair shot at a good career. I can't even begin to talk about how that is not the case in America - if we disagree here, I'm not sure we'll ever come to agreement.

It's not clear that the workers favored in this alternate pipeline are incompetent (which is really the implication that everyone is making). What is clear is that people with their perspectives are not as prevalent in the technology field.

Perhaps companies are recognizing a) the value of multiple perspectives or b) the structural difficulties that deny some people the same 'fair shot' we all have.

I'll leave you with this. The SAT should be a test of one's aptitude, perhaps readiness for college. If some individuals receive years of tutoring, but others have to go to the library to get test preparation, is our metric for college readiness fair? Maybe you'll say yes, but if we were talking about any other thing (cars, experimental models, you name it) we would clearly recognize these two groups had different pre-test levels of preparedness.

I'm not sure if you read his document, where he talks about Google having lowered the standards for its employees.

But I'll leave that to the side - there's no reason to think that many of the differences he describes aren't socially constructed. Maybe the way we raise women in America/the West turns them off to pursuing tech careers.

Until we can resolve this, it's ridiculous to use psychology studies to ascertain biology - and even afterwards, our goal should be ensuring access to well-paying careers, not remaining shackled to our biology.

As an example, we have trained runners to break previously 'impossible' race times through training, technology and science - we realize that what we thought was impossible was simply not compatible with the system we previously had in place.

You're highlighting another big problem these days, slavish reverence for the 'best schools'.

I will also get a PhD from Harvard, and from a more selective program than this person. Yet I recognize the limits of my knowledge.

Studying microbes [1] does not give you the standing to discuss sociology. If your very well-qualified dentist wanted to do a code review of your last commit, you might think twice before accepting his advice - namely, you would ask what special knowledge he has, besides being a generally smart individual.

1. https://www.ncbi.nlm.nih.gov/pubmed/?term=Damore%20JA%5BAuth...

I think you're doing contortions here.

He says that Google 'has created' (past tense) a discriminatory hiring practice.

Presumably his coworkers were hired in the past and not the present or future.

Thus, the hiring practices that prevailed when his coworkers were hired were discriminatory, per the author.

First, I agree that non-hormonal contraception decreases a lot of the risk associated with hormonal contraception (almost tautological, but worth stating).

Second, there is no increase in mortality between women who use oral contraceptive pills and women who do not (with the exception of women over 35 who smoke; OCPs are contra-indicated in this group).

For any OCP-users who might read this, there is no overall increase in cancer associated with OCPs, though some cancers are more likely with OCPs and others are less likely with OCPs.

Many of the side effects you're mentioning are either rare (hypertension, stroke/blood clots) benign (maybe you mean hepatic adenoma when you say liver disease?) or things I have not heard of (would appreciate scholarly references; gall bladder disease, heart disease[not sure what you mean here]).

> Taking pills to inferfere with your normal hormonal and bodily functions is not a great idea, when there are other options with fewer or no side-effects

In general, all medications have risks and benefits. A blanket statement such as this one ignores the risk-benefit calculation that each patient and her physician must assess in picking a medication.

This is definitely a trend but it is one I do not suspect will change any time soon. Just to put things in perspectives for people who have not experienced medical school:

> 50% of students I knew (secretly) spent thousands of dollars on recorded medical lectures

While this may be indeed be true for the students you know, this is also a common practice, in general. The pay-to-play nature of board exam preparation has existed much longer than the active-learning trend. There's no way to know whether these students would have purchased additional materials in another educational system.

> thus did well on the tests (and MCATs) anyway.

You seem to be conflating a couple of things. The MCAT (the Medical College Admission Test) is taken before medical school. There, just as for later examinations there has been a proliferation of prep materials, for which students pay dearly. This is in no way related to curriculum change in medical schools.

Again, you are correct that students who pay more do better.

> "objective peer-reviewed studies" of this learning style won't show score/testing degradation -- because the students are so self-sufficient that they work around the failing format with supplemental lecture materials purchased online

I understand the sentiment, but if students do better after the adoption of this new system, aren't they learning better? Eventually, medical students move from Wikipedia to more authoritative sources like UpToDate or primary literature. If they, themselves, are motivated to expand their knowledge - and as a result do better on tests - this intervention would have been beneficial.

I think you're implying that the huge industry behind medical test preparation materials is new, but this is decidedly not the case.

A recent study that peripherally relates to our discussion came up on my facebook feed:

https://hms.harvard.edu/news/recession-proof

Incidentally, I'll be starting a PhD in a human genetics laboratory, but I don't have a lot of experience with this kind of literature.

This seems like a reasonably accessible (college-biology level) discussion of what evolution would look like in the human gene pool:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2933187/

----

EDIT:

I suspect you'll need a university library to access this, but Dr. Pardis Sabeti is well known in this area and also an amazing person - she played a big role in the molecular biology work that underlay the US response to the Ebola crisis. She wrote a review that explains the tools we currently have to ascertain human evolution:

https://www.ncbi.nlm.nih.gov/pubmed/16778047

I guess I would need to see an analysis of whether there's been any change in human allele frequencies. I suspect there hasn't.

But to your example, one issue with the farmed fish is that they now constitute a closed gene pool - we can easily explain their decreased fitness with the likely rise of harmful alleles.

With the explosion of population, I don't see much relationship between the salmon and humans.

I don't think you can remove a discussion about changes in allele frequencies from environments. Thus, making 'other things being equal' is not a reasonable premise from which to discuss evolution.

I do see the argument you are espousing but I would only suggest that you consider the broader context of what happens when new species arise. I suspect you care more about the elimination of alleles (which is surprisingly hard to do), but this is a good example for why I disagree:

One possibility for speciation is the creation of a new niche (lets say a new type of tree grows on an island, its seeds having been carried there by an unusual current). Here, organisms might have increased reproductive opportunities over their neighbors by exploiting this new tree. In time, one could imagine organisms that live in the new tree, like microbes, could acquire enough new features to constitute a new species.

Consider the case where organisms from the original population become unable to grow on the new tree. If the new species is also unable to grow in the original environment, how can we say either is more or less resilient?

I think you're misunderstanding my point, but maybe I shouldn't have taken vertebrates as an example; we can of course imagine some island where massive selection pressures lead to speciation. I would be interested if you have such an example.

However, dog breeds are not species and, given their non-random mating, don't really tell us much about whether 200 years is long in vertebrate evolutionary time.

As far as humans go, I will grant that allele frequencies may have changed if you pick an isolated subpopulation of humans (say, everyone in Greenland) but even that seems extremely unlikely.

As an aside, there are researchers who are using the caste system in India as a way to explore genetic architectures within a civilization. As these castes have existed for considerably longer than the period we're discussing, meaningful allele frequency changes may arise.

I would be very interested to read an academic paper that made this point. 200 years is a blink of an eye in almost any vertebrate's evolutionary time.

Also, to support your claim, one would need evidence that the child mortality of the past reflected genetic unfitness (as opposed to poverty or the unavailability of antibiotics).