Then you're singularly incurious. And besides, the article begins with a massive portrait of her (colour changed but otherwise clearly representative).
Much of this article is about appearance because our ancestry defines it. I bet most readers would have found a picture extremely relevant to the article.
One test said "2.978 percent African", and she thought she wasn't Black. (2.978%? Three decimal places, really?). Another test said 23% and she thought she was. (23%? So that means like one grandparent? Not really, it's unclear what the "percentage" means or if it even means the same thing from one company's test to another. It means something like "23% of your DNA, the parts we look at, matches some reference sample, that we got from somewhere." What does THAT mean though? What does "matching a reference sample" _mean_ in terms of where your ancestors were from? How has that been validated? And many companies don't even reveal their statistical methods.)
Or maybe it was "0 to 58%".
These "ancestry results", despite being expressed with five significant digits, are scandalously unscientific and of unclear meaning. It is not even clear what "percentage of your ancestry from Africa" _means_. You think people didn't move around and have children with people from all over until the 20th century? Cause they did.
If your great-great-something grandparent lived in what is present day Greece, but THEIR parent was born in what is present-day Egypt (or vice versa; either would be fairly unremarkable)... and you want to know what "percent Greek" your "ancestry" is, like that even means something? Like your DNA could somehow even reveal that great-great-something was born in what is now Egypt but their parent was born in what is now Greece? It can not. All of this is based on a weird fantasy/mythology of being able to separate people into historically continuous categories which remain stable over generations, that just isn't any kind of scientifically or historically justified.
Yeah it is an attempt to correlate genes to geography with far more precision than actually existed and assumptions that populations were hermetically sealed when they weren't. Regional classifications are often even worse in GIGO associations from what I heard with trying to distinguish between say Dutch, Polish, German, and Swiss. It brings to mind sarcastically asking questions about connections to long gone civilizations that nobody really identifies as. How many people globally are Scythian? Babylonian? Carthaginian? Sumerian? Minioan?
It brings to mind the 'smartass genie' technically most accurate way of classifying blackness by just including a count of melanin producing genes.
Everyone who's family has lived in North America for more than a few generations has at least some of both European and African ancestry. This means, that of these people, "black" Americans have European ancestors, and "white" nationalist Americans have African ancestors. We're all mutts, there's no "pure" anyone... just pure human.
Ancedotally: I'm an SAR member with non-tanning skin but have European, African and Asian/native American genes.
The conclusions: races are illusions and racism is completely idiotic. There are bigger things to worry about or look down on than which flavor of human a person ticks checkboxes for identity politics games.
Questions: Let's be honest, what is "diversity?" And then, why do organizations claim to be diverse when they either hire people that all look like them or hire only people that think or believe as do they? Isn't that the opposite of "diversity," but exclusive monoculture or groupthink?
> They said she was 45.306 percent Hispanic, 32.321 percent Middle Eastern, 13.714 percent European and 8.659 percent “other,” which included a mere 2.978 percent African.
Decimals aside, since when is "Hispanic" a race? What could that even mean?
That too aside, 3% African is black according to the still predominant "one drop" doctrine of blackness in this country.
Technically it is an ethnicity but the South American population does have a larger percent of assimilated natives than continental Spain and a long history of a caste system which put born natives lower in the hierarchy than those born in Spain which could arguably make them a more divergent set of genetic traits being messily correlated. One can argue about the name but "Hispanic" is a workable proxy for the broad brush of 'everything below the continental born caste'.
Of course if you want to talk genetics instead of silly human labels you look at the actual contents of their genome.
I realized afterwards that I had meant nationality but that wouldn't be quite accurate either - unless you count former ones like "was once a part of the USSR". And even that would probably be pushing it.
Users flagged it. That's common. Also common is to see it rise again as the tug-of-war between upvotes and flags plays out, or occasionally as moderators turn off flags.
This is an interesting piece from the angle of wondering how we construct our identities and what happens when those constructed identities turn out to be based on "facts" that aren't actually true. Such lines of inquiry have potential to break down barriers by proving the degree to which they are psychological fabrications.
This is an uncomfortable idea for most people. They want to think things like race are objective facts. Learning that racial identity can be a construct based on social messaging that is largely or wholly unrelated to actual DNA evidence calls into question the very basis for much of the racism in the US today.
You are mistaken if you think those DNA percentages are "facts" exactly either, though.
I would say that the "fact" that she has lived her whole life as a Black person is more of a "fact" than the "facts" that one DNA test said "5% African" and another said "23% African" and another said "0-54% African". Those percentages are the output of algorithms with unclear validity and meaning, devised in order to put something on a report to sell tests, they aren't exactly "facts".
But yes, I agree that race is, of course, socially constructed and not a biological "fact" at all... and for that matter so are "ethnic" categories like "Greek" or "Irish"... which is just part of why these DNA tests can't do what people want them to do.
You are mistaken if you think those DNA percentages are "facts" exactly either, though.
If that is the impression I gave, then my wording probably was sloppy. I've reread it and can't figure out what I "should" have done differently to make it clearer that I am aware that the tests themselves are of questionable merit (at this point in time, when it comes to racial identity), given the issues you pointed out.
I have personally read of cases with devastating personal consequences based on things that are on more solid ground, such as paternity. There was an incident where someone got a DNA test and I think gifted their parents DNA tests and the parents ended up divorced because it revealed that the husband had fathered a child out of wedlock prior to getting married.
I found that especially shocking and bizarre. When I saw the title of the piece, I assumed that it had revealed an illicit affair during the marriage.
Anyway, this article is about someone who was adopted. Adopted children often don't know they were adopted. Even if they do know, it leaves a huge hole in their knowledge about their heritage.
So I think it is inevitable that at some point we will have more reliable tests and someone like this individual will learn that their genetic heritage conflicts with their self image in some important way.
Perhaps we will even see some Neo Nazi members learn that they are not entirely Caucasian either and then what? Do you rethink your racist ideology? Do you kill yourself? Do you live in denial and try to hide the truth from your Neo Nazi brethren?
Sorry, I didn't mean to accuse you of anything or suggest you "should" have said something different. Just trying to engage in the conversation.
But the point I'm trying to make is that I'm _not_ sure that "it is inevitable that at some point we will have more reliable tests". The problem is not test reliabilily, it's that there is in fact no actual reliable way to map from DNA to "not entirely Caucasian", because there's a basic mismatch there. "Caucasian" is (really!) not a genetic category, there is actually no way to _reliably_ determine someone is "Caucasian" (or "Black", or anything else like that) from a DNA test. It just doesn't work like that. (Even though these DNA ancestry companies make money off people thinking it does).
It's true that at some point some neo-nazi, like anyone else, might have results that seem to tell them something that surprises them or that they didn't want to hear. But the problem is in fact with their ideology in the first place that thinks there is such a thing as "Caucasian" encoded in their genes, when there isn't.
This is interesting though, in the case where someone who thought they were Black thinks a DNA test told them they weren't -- my inclination is to say "Ignore the test, it can not tell you that you aren't Black, because the test just can't say that, don't let it make you insecure in your identity" In the case of a neo-nazi who has a similar experience, I'd _want_ it to make them insecure in their identity, because their identity is based on entirely scientifically unsound ideas about what race means.
That might be a contradiction in my OWN ideology. But it's not a contradiction in the science. Either way, the problem is not just that "the test isn't reliable enough yet", the problem is that "Caucasian" or "Black" can NOT be "reliably" determined by a DNA test, because these categories are simply not sound as biologically/genetically determined categories.
This is a bit too deep for me at the moment. I don't know that much about the science. What little I know about DNA is in service to understanding my genetic disorder.
It is an utterly absurd situation that someone like Kamala Harris, who is half-Tamil and partly English, is considered "black". Our present notions of race continue to maintain the repugnant fiction that the hereditary material of some populations is "marked" and those who have even a few such ancestors are "contaminated". Of course, this is very far from accurate. It may reflect present social realities, but there is no rule that says present social realities must not be absurd.
> One said that any percentage not marked “low confidence” was 100 percent certain. Another said each percentage was 99 percent certain. When I asked that representative to check with a supervisor, she did, then returned to tell me that the company’s certainty was 99.7 percent.
Obviously this is a confidence interval covering three standard deviations assuming a normal distribution. If the last assumption can be considered accurate, the correct interpretation of the results' precision is "pretty good", but non-normal distributions occasionally show up and cause the 2007 mortgage collapse.
30 comments
[ 2.3 ms ] story [ 66.4 ms ] threadOne test said "2.978 percent African", and she thought she wasn't Black. (2.978%? Three decimal places, really?). Another test said 23% and she thought she was. (23%? So that means like one grandparent? Not really, it's unclear what the "percentage" means or if it even means the same thing from one company's test to another. It means something like "23% of your DNA, the parts we look at, matches some reference sample, that we got from somewhere." What does THAT mean though? What does "matching a reference sample" _mean_ in terms of where your ancestors were from? How has that been validated? And many companies don't even reveal their statistical methods.)
Or maybe it was "0 to 58%".
These "ancestry results", despite being expressed with five significant digits, are scandalously unscientific and of unclear meaning. It is not even clear what "percentage of your ancestry from Africa" _means_. You think people didn't move around and have children with people from all over until the 20th century? Cause they did.
If your great-great-something grandparent lived in what is present day Greece, but THEIR parent was born in what is present-day Egypt (or vice versa; either would be fairly unremarkable)... and you want to know what "percent Greek" your "ancestry" is, like that even means something? Like your DNA could somehow even reveal that great-great-something was born in what is now Egypt but their parent was born in what is now Greece? It can not. All of this is based on a weird fantasy/mythology of being able to separate people into historically continuous categories which remain stable over generations, that just isn't any kind of scientifically or historically justified.
It brings to mind the 'smartass genie' technically most accurate way of classifying blackness by just including a count of melanin producing genes.
Ancedotally: I'm an SAR member with non-tanning skin but have European, African and Asian/native American genes.
The conclusions: races are illusions and racism is completely idiotic. There are bigger things to worry about or look down on than which flavor of human a person ticks checkboxes for identity politics games.
Questions: Let's be honest, what is "diversity?" And then, why do organizations claim to be diverse when they either hire people that all look like them or hire only people that think or believe as do they? Isn't that the opposite of "diversity," but exclusive monoculture or groupthink?
Decimals aside, since when is "Hispanic" a race? What could that even mean?
That too aside, 3% African is black according to the still predominant "one drop" doctrine of blackness in this country.
Of course if you want to talk genetics instead of silly human labels you look at the actual contents of their genome.
"African" is similarly vague. A pure Bantu, an Egyptian, and an Amazight from Morocco are all African.
I don't see how DNA analysis can be one of them!
This is an uncomfortable idea for most people. They want to think things like race are objective facts. Learning that racial identity can be a construct based on social messaging that is largely or wholly unrelated to actual DNA evidence calls into question the very basis for much of the racism in the US today.
I would say that the "fact" that she has lived her whole life as a Black person is more of a "fact" than the "facts" that one DNA test said "5% African" and another said "23% African" and another said "0-54% African". Those percentages are the output of algorithms with unclear validity and meaning, devised in order to put something on a report to sell tests, they aren't exactly "facts".
But yes, I agree that race is, of course, socially constructed and not a biological "fact" at all... and for that matter so are "ethnic" categories like "Greek" or "Irish"... which is just part of why these DNA tests can't do what people want them to do.
If that is the impression I gave, then my wording probably was sloppy. I've reread it and can't figure out what I "should" have done differently to make it clearer that I am aware that the tests themselves are of questionable merit (at this point in time, when it comes to racial identity), given the issues you pointed out.
I have personally read of cases with devastating personal consequences based on things that are on more solid ground, such as paternity. There was an incident where someone got a DNA test and I think gifted their parents DNA tests and the parents ended up divorced because it revealed that the husband had fathered a child out of wedlock prior to getting married.
I found that especially shocking and bizarre. When I saw the title of the piece, I assumed that it had revealed an illicit affair during the marriage.
Anyway, this article is about someone who was adopted. Adopted children often don't know they were adopted. Even if they do know, it leaves a huge hole in their knowledge about their heritage.
So I think it is inevitable that at some point we will have more reliable tests and someone like this individual will learn that their genetic heritage conflicts with their self image in some important way.
Perhaps we will even see some Neo Nazi members learn that they are not entirely Caucasian either and then what? Do you rethink your racist ideology? Do you kill yourself? Do you live in denial and try to hide the truth from your Neo Nazi brethren?
But the point I'm trying to make is that I'm _not_ sure that "it is inevitable that at some point we will have more reliable tests". The problem is not test reliabilily, it's that there is in fact no actual reliable way to map from DNA to "not entirely Caucasian", because there's a basic mismatch there. "Caucasian" is (really!) not a genetic category, there is actually no way to _reliably_ determine someone is "Caucasian" (or "Black", or anything else like that) from a DNA test. It just doesn't work like that. (Even though these DNA ancestry companies make money off people thinking it does).
It's true that at some point some neo-nazi, like anyone else, might have results that seem to tell them something that surprises them or that they didn't want to hear. But the problem is in fact with their ideology in the first place that thinks there is such a thing as "Caucasian" encoded in their genes, when there isn't.
This is interesting though, in the case where someone who thought they were Black thinks a DNA test told them they weren't -- my inclination is to say "Ignore the test, it can not tell you that you aren't Black, because the test just can't say that, don't let it make you insecure in your identity" In the case of a neo-nazi who has a similar experience, I'd _want_ it to make them insecure in their identity, because their identity is based on entirely scientifically unsound ideas about what race means.
That might be a contradiction in my OWN ideology. But it's not a contradiction in the science. Either way, the problem is not just that "the test isn't reliable enough yet", the problem is that "Caucasian" or "Black" can NOT be "reliably" determined by a DNA test, because these categories are simply not sound as biologically/genetically determined categories.
Have a happy Sunday.
> One said that any percentage not marked “low confidence” was 100 percent certain. Another said each percentage was 99 percent certain. When I asked that representative to check with a supervisor, she did, then returned to tell me that the company’s certainty was 99.7 percent.
Obviously this is a confidence interval covering three standard deviations assuming a normal distribution. If the last assumption can be considered accurate, the correct interpretation of the results' precision is "pretty good", but non-normal distributions occasionally show up and cause the 2007 mortgage collapse.