It seems silly to me to believe that our races make us inheritently different. Our culture and our ancestry does make us different. To the extent Black people are on average less intelligent than white people. It would be reasonable to believe poor people are less intelligent than well off people. Since more black people live in poverty it is reasonable to believe that that would have an effect on their ability to advance. Likewise though just as Asian and Jewish cultures push their kids to learn more and excel academically, there is a subsection of the Black culture that shows little interest in academic achievement. With the violence issue, yes with Black Americans they are more likely to be violent that cant be denied, but i would argue that is a relatively new phenomena brought on by the crack epidemic in the 80s and 90s. The ripples from the crack epidemic can no doubt still be felt in black neighborhoods. These are cultural issues, i think it is important to be honest the issues facing the Black culture so they can improve. But these issues are not genetic there is no reason to think one race is superior to another.
You make some excellent points about the undesirability of a world using population to determine outcomes. BTW, in the 60s and early 70s there was an understanding that only a limited number of blacks could play for and be on the basketball court at any time. The introduction of race as a classification and analytical frame leads directly to assumptions about capability based on race. If we used other classification schemes we might come to very different conclusions. There is also the problem with trying to explain complex, multi variable human and social outcomes using a single characteristic or factor.
John, I'm writing a second responsive post after listing to a couple of your related and recent conversations (YouTube'd) with Glenn Loury. By the way, please keep posting these longer form conversations, as this present national focus on racism is ripe with layer and nuance.
Murray's latest work, to your point, compellingly presented that the numbers say relative to population-level IQ. The good and hyper-rational news is the vast majority of all four populations (Murray's grouping) are of sufficient and similar IQ such that mental agility (what IQ measures) is NOT a deciding factor for a successful life or community. Other factors such as early childhood education, two-parent present families, delaying children until after sufficient education, property-secure neighborhoods -- other social hygiene factors matter as much or more than mental agility. That is the space where most of the human population lives, works, and plays.
However, the intractable and stark reality about mental agility population-level statistics concludes the distribution -- the right-size high standard deviation outliers correlate with ethnic population membership. The "high right-side," extremely high mental agility groups will have, proportionally and in descending order, Asian, European, Latin, then African members. Therefore, in a just and fair and meritocratic community, high mental-agility professions will populate accordingly. To demand some other proportionality in the spirit of fairness or equity is not only irrational, it is societal suicide as it demands we refuse the better qualified and capable and reward the less qualified and incapable. It is an attempt to thwart our evolutionary imperative of natural selection and survival of the fittest. It ain't gonna work, and in the spirit of Jerod Diamond's "Guns, Germs, and Steel", it will be only a matter of time when an adjacent tribe who doesn't not play this foolish woke fiction will conquer us, sooner or later.
Added: I forgot that Murray had posted a spreadsheet with his data, so mea culpa there. But this still doesn't excuse leaving the numbers out of the book, and his case looked even worse when I re-examined his data.
Note: As for the low sample size, when he said "6 or fewer," he wasn't referring to total numbers of arrests, but of arrests for individual groups. His explanation makes no sense to me in light of the spreadsheet, and I'm just going to move past it.
Very important: he left 5 cities out of the spreadsheet (confirmed in a pdf with his notes that lists only 8 cities). This means that the values for Chicago, Baltimore, Lincoln, Ashevelle, and Urbana are unsupported by data. This is Bad. Throw away his ratios for those cities.
Further, the values for Chandler are based on only 13 arrests. I calculated p-values (recall: you want them to be <0.05, meaning there's a 5% or less probability the finding was due to chance). The p-value for European:Black was ~0.7. meaning there's a roughly 70% probability it was due to chance. Interpretation: "beyond the realm of wishful thinking." The value for European:Hispanic was ~0.2. Perhaps not beyond the realm of wishful thinking, but still solidly in the land of hooey. Okay, throw out Chandler.
Well.
This leaves us with seven cities, one of which is New York.
Remember what I said about sample bias, when the sample is dominated by one component? Well, New York had ~13,000 arrests for murder. The next closest was LA, with ~2,100. After that came DC with 565. The next four don't even add up to half of DC. This situation is called undercoverage.
When you do a statistical analysis, you want your sample to represent the population as a whole. If it's dominated by one group, this is generally not the case. This is why the FBI database is so wonderful. It has data from all 50 states plus Puerto Rico and DC (see for example table 4 in the 2019 section of the website).*
Conclusion: the entirety of chapter 4 and anything derived from it (e.g. chapter 6) are bogosity.
I just listened to a conversation between Coleman Young and Charles Murray on Murray’s new book on Coleman’s podcast. It was respectful but Coleman was quite persistent in challenging the implications of Murray’s work if it were to go mainstream—I would recommend people listen to it.
One conclusion I came away with is that Murray’s book is a response to the increasingly dominant anti racist narrative of CRT. His concern seems to be that embracing a position that promotes that America is systematically racist, that whites as a group are essentially racist, and that black peoples’ shortcomings, whether it is higher crime rates or lower intellectual accomplishments are a function of white racism—I.e., it’s not our fault—can only lead to a destructive backlash from whites and possibly other nonblack citizens. This backlash would lead to all citizens seeing themselves not as individual citizens who judge each other based on their talents and accomplishments but as members of racial and ethnic identity groups.
This would be the end of the American experiment. Murray wants to fight against the charge of systematic racism by arguing the empirically observed disparities between blacks, whites and, other groups are not the result of racism but primarily the outgrowth of differences in cognitive ability. He is less clear about whether higher violent crime rates among blacks are the result of cognitive differences or some other factor—e.g., cultural values. (As Val’s post below indicates Murray’s calculations on black-white and Hispanic-white crime differences are flawed because of his use of small sample sizes taken at the city level.)
I share Murray’s concerns about anti racism and identity politics threat to our Republic. However, it seems to me it’s a bit of a cop out to say that the only way to fight these destructive forces is to say that blacks’ disparities are the result of their cognitive shortcomings. This means there is no closing the gaps either through public policy or by groups pursuing a self development strategy. It strikes me that Murray’s position if embraced widely would lead to exactly the outcome he fears. It would underscore the idea in society that blacks are an inferior identity group that must be controlled.
I spent three years teaching science and math to high schoolers in Botswana in the 1980s. I had the full range of kids from brilliant to dullard. By and large, the kids that did well and could think analytically, came from families that had more exposure to technology and the outside world, or, whose parents were clearly pushing their kids to excel. There were exceptions, but it appeared clear to me that culture and expectations play huge roles in the average kid's intellectual development.
I spent much of yesterday trying to sift through data on crime statistics and ended up feeling dazed and confused by the variety of data sources and the difficulty of sifting through it all. Even the two sources Murray cites in his book are what I’d call databases of databases, and rooting through can feel overwhelming.
But then, today … I found the FBI crime statistics, and I felt dazed and confused for different reasons.
The FBI takes in data from thousands of agencies large and small, and assembles it all into one very tidy, very accessible pile. Arrests for serious crimes are separated into 31 categories that are broken down by year, race, and ethnicity.
For example, the 2019 data came from over 12,000 agencies representing ~250 million people, or about 75% of the total US population. Why would you spend weeks or months digging through databases of databases, only to end up with information from 13 cities/18 million people, when you could have 13-14 times that number of nationwide statistics, all in nice tables? You wouldn’t be forced to cull information on one or two things, because there are lots of categories right there, just waiting for you to look at them.
Why didn’t Charles Murray use the FBI data? His explanation on page 137 is weak at best. He claims you need to have city-specific arrest data paired with city-specific demographic data. But if a dataset represents 75% of the nation, you can use demographic data for the nation as a whole. This approach, combined with the large number of categories, makes for a much more generalizable conclusion than data from Fayetteville and New York and a relative handful of other places.
And what did I learn?
Summary: things do not break down into his categories (most violent first):
1. Black
2. Hispanic
3. White
4. Asian
True: Asian people did have fewer arrests overall for most categories (MOST). Beyond that, what I found was exactly what I’d expect from any large dataset; namely, there are some patterns, but also lots of variation from one category to the next. It’s not as simple as Murray implies.
And here’s my biggest challenge with this: I’ll lay it out findings and a discussion as succinctly as I can in the next comments, but if you want to understand what I’m writing, you’ll have to pay attention to some detail. The flip side is that Facing Reality can be summarized in one sentence: Blacks and Hispanics have lower IQs and commit more violent crimes. This is a major problem faced by people who struggle with misinformation: reality is usually nuanced and detailed, but misinformation and distortions can be shiny and pretty to the person spinning them. Such is the case with Facing Reality.
Summary of findings about murder arrests from FBI database:
I calculated murder arrests per 100,000, in 2018 just like Murray, and found the following:
Overall: 2.7 murders per 100,000
Black Americans: ...10.9 per 100K;
Hispanic Americans: ...2.4 “
White Americans: ...2.0 “
……………………………………FBI……..Murray
Black:White ratio:………. 5.5…….. 21.0
Hispanic:White ratio: ….1.2…….. 4.9
I calculated a comparison of two rates and got p-values <0.001, with values <0.05 indicating statistical significance. So it’s true that the value for Black Americans is much higher. But the value from the big FBI database isn’t even one quarter of Murray’s value, and the ratio for Hispanics is close to 1.
(The numbers for 2019 are B:White = 5.0; H:White = 2.0; for 2016, they were 5.3 and 1.1, respectively)
And now the background:
I believe that Charles Murray made a terrible mistake. It’s the kind of mistake that’s plagued drug development and given us expensive epic fails in clinical trials: a small sample size in your preliminary work can show false trends. And Murray’s data is just that: “preliminary.”
Recall that he didn’t say how many total murder arrests he was working with. I suspect his sample size was pretty small. This is because he said that some cites had 6 or fewer arrests over the entirety of the database and others had “at least 11.” The FBI had 8,957 arrests (from 12,000+ agencies) with known race in 2018 alone, and knew Hispanic/non-Hispanic ethnicity in 7,050 of them.
I can tell you with confidence that conclusions from small datasets can be horribly misleading. When you can count your data points on your fingers and toes, as is the case for at least some of his cities, as few as ONE new outlying data point can make a big change in your numbers. This means that Murray’s ratios for four cities are essentially meaningless, leaving 10 cities for Black Americans and 4 for Hispanics. And at least city one has only 11 data points.
So not only is Murray making generalizations about the behavior of millions of people based on arrests for one crime — murder — he’s doing so on scanty data.
Murray is intelligent and highly accomplished, but he’s not a scientist or someone skilled in statistical analysis, and this really shows in the book. He’s very good at making a case with what he’s got, but he doesn’t seem to see that what he’s got isn’t enough to make generalizations. If this book had been a paper submitted for peer review, the reviewers would have been all over it regarding total lack of statistical analysis, poor or no citation of sources, and lack of statements about sample size and the limitations of a small sample size and biased samples (means that if one or two of his cities had many more murders than the others, they would have had an outsized influence on his results…but he didn’t tell us, so we don’t know).
Facing Reality doesn’t push flat-out misinformation along the lines of “vaccines cause autism.” It’s more that it distorts along the lines of something like “vaccines have mercury, which is toxic.” Sure mercury is toxic, but the amount of mercury in vaccines is only toxic to bacteria and fungi, which was the whole point of adding thiomersal. And besides, thiomersal is converted ethylmercury, which doesn’t accumulate in your body. But the original sentence IS true: mercury is toxic. It just doesn’t bring in any detail. Instead, it quietly encourages people to fill in gaps along the lines of, thiomersal = toxic vaccines = brain damage in me or my kid.
Murray does essentially the same thing, though he may have made a genuine mistake. He makes some true-ish statements, and readers who didn’t dig through the FBI database may fill in the gaps, along the lines of “Black Americans are 21 times more likely to kill people!”
Readers shouldn’t have to dig through an FBI database to know that the data doesn’t support this idea. Murray should have done that, and he didn’t. He just made a weak-to-poor justification for not using it. I’m not accusing him of malfeasance or RACISM! Again, it’s possible the problem is a genuine mistake based on lack of knowledge. What bothers me is the aura of expertise.
Research is complicated. It involves a lot of nuance, a lot of detail, and a lot of apparent contradictions — and sometimes other factors. Figuring it all out takes a lot of thinking and a lot of time (think about how long it took to sort out the whole sun-earth thing). The methods you have to use as a responsible scientist make this process very difficult.
And herein is a big red flag about Facing Reality: it’s a quick and easy read (Murray even says so on page VII). He took a very complex subject (crime) and told us we can reduce it to result arrests for murder in a handful of cities. He drew vast conclusions from this narrow set of data — and he gave almost no background information. He also ignored other forms of crime almost completely, as though they don’t count. More on that tomorrow or Monday if I have the energy.
The worst thing here is that when an idea gets a) repeated for a long time and b) politicized, people choose up sides and dig in (as was the case with the earth and the sun). This makes us even less likely to read a detailed counterargument to conventional wisdom (and more likely to get angry). We need to get past this. Somehow.
And if we don’t talk about WHY problems develop, we will never be able to solve them. That is another fatal weakness in Facing Reality.
I am tired now and will write even more next time. Next up will be some summaries showing other findings from the FBI database. Preview: what I found is different from what Facing Reality implies in Chapter 4.
I appreciate Dr. McWhorter's deep reading of Dr. Murray's latest book. He bent over backwards to be fair-minded. Not sure Murray deserves it. I wouldn't call Murray a brilliant thinker. He says things no one else says, and that counts for originality and courage in some quarters. His book might seem substantial up against books by, say, Robin D'Angelo or Ibram Kendi, he certainly has gathered more data, but his conclusions are, if not racist, than poorly considered.
To me the right answer is do nothing since all that data in aggregate does not inform on the individual. It is the individual that is doing the job, research, and analyzing. Being successful at a any of those is about much, much more than raw intelligence, it is about character traits like drive, work ethic, team work, and many more. Intelligence is just one element and we all know very "sharp" people that are lazy, lack focus and as such can't do the 'analytical' work either. We must focus on the individual and treat each as unique. Sure collect data to find gatekeeper that are grading on account of race, which simply is not a factor in the ABILITY to be successful, and remove that gatekeeper from such responsibilities.
Someone, perhaps it was Murray himself, has said that, as societies gain in general practical fairness--that is, the less invidiously discriminatory they become --the more discrepancies shall appear in these expectations for equitable _outcomes_. That is, in other words, inherent differences in talents and abilities, given greater scope by fairer practices as prejudicial practices are eliminated, shall _increase_ rather than decrease the occasion for casual observers' perceived _inequality_ in precisely these expectations for more "even distributions" in _outcomes_-based “equity”.
The same things were said of the Jews, had the test existed it would have been said of the Irish.
Binet was testing for learning disabilities for the French school system. When he saw where the Eugenics school was taking his test he mysteriously dies, perhaps of a broken heart…but possibly suicide.
Binet was Catholic. At the time suicides we’re damned and could not be buried in blessed ground, as we say “his death was doubtful “.
The test is FALSE, its perpetrators FRAUDS.
And Intelligence is not Virtue.
Intelligence will allow you to succeed at TASKS. To perform DUTIES requires Virtue- Courage. Something no one would accuse America’s Elites of, but they might see it in Black Men.
So you can put down the axe and let go of the Rack Professor Procrustes McWhorter - you’ve been had.
I have been selectively breeding crops for several decades. Facing data showing the identical ranking of different varieties of a crop in different environments and different years, it would be clear that something is being inherited differently by the different varieties. We seem to be discussing something parallel here. However, I also share John's gut feeling that culture, not genes is involved.
This is no contradiction. Plants transmit many things to their progeny and "genes" (DNA from the plant cell nucleus) is certainly important but is not the only category. Plants transmit cellular contents, whole organelles (e.g. chloroplasts), symbiotic fungi or bacteria, pathogens, nutrients, RNA and epigenetic "marks" (e.g. gene silencing) to progeny. Humans transmit these and, usually, culture, language, religion, property, caste, family names, music, money, debt, etc. Many of these forms of inheritance are much less predictable than genes and the thing transmitted may have much, much less fidelity than a DNA sequence. But collectively these inheritances can greatly affect an individual's success and even interact with the expression of one's genetic endowment.
I agree with John that it is essential to go beyond diagnosing that "something is inherited". Some folks argue that what "the thing" inherited is institutions and political system (e.g., institutional racism). McWhorter and Loury seem to be arguing that institutional and personal racism is plastic and has mutated to much less virulent forms in recent generations. They seem to find the contribution of these factors to be insufficient to explain the persistence of heritable differences between races.
On the other hand, it is really hard for me to imagine any human evolutionary trajectory where intelligence wouldn't have had extremely high fitness value. And even if there had been a slight difference in gene frequency, would it have been preserved after all of the gene flow between previously isolated groups that has happened in the Americas?
John's "victim syndrome" could certainly be heritable and could be present at significantly different frequencies in different populations--and influence test-taking phenotypes. But so could something as banal (and potentially treatable) as gut microbiome. I've often wondered if the idea of free will is inherited and is highly influential in childhood cognitive development--whether or not free will is "true" in a metaphysical sense.
The root problem is that demographic data are simply not accurate enough to support firm conclusions. Any first-order conclusions based on even mildly inaccurate data cannot be trusted and if someone is making higher order conclusions then they are essentially practicing astrology.
Here are two references (of many that exist) documenting errors of 15-30% in ethnic group in one case and diagnosis in another. Inaccurate data is great for coffee-table discussions on the "meaning" but should have no place in rigorous scientific discussions.
That means any large-scale "results" are not scientific. We need to accept this as a fact and deal with that, instead of let our desire to get results cause us to be delusional.
It seems silly to me to believe that our races make us inheritently different. Our culture and our ancestry does make us different. To the extent Black people are on average less intelligent than white people. It would be reasonable to believe poor people are less intelligent than well off people. Since more black people live in poverty it is reasonable to believe that that would have an effect on their ability to advance. Likewise though just as Asian and Jewish cultures push their kids to learn more and excel academically, there is a subsection of the Black culture that shows little interest in academic achievement. With the violence issue, yes with Black Americans they are more likely to be violent that cant be denied, but i would argue that is a relatively new phenomena brought on by the crack epidemic in the 80s and 90s. The ripples from the crack epidemic can no doubt still be felt in black neighborhoods. These are cultural issues, i think it is important to be honest the issues facing the Black culture so they can improve. But these issues are not genetic there is no reason to think one race is superior to another.
I wonder whether John would let Glenn know that I can pay 5$ for a subscription, but not 7$. Thanks.
You make some excellent points about the undesirability of a world using population to determine outcomes. BTW, in the 60s and early 70s there was an understanding that only a limited number of blacks could play for and be on the basketball court at any time. The introduction of race as a classification and analytical frame leads directly to assumptions about capability based on race. If we used other classification schemes we might come to very different conclusions. There is also the problem with trying to explain complex, multi variable human and social outcomes using a single characteristic or factor.
John, I'm writing a second responsive post after listing to a couple of your related and recent conversations (YouTube'd) with Glenn Loury. By the way, please keep posting these longer form conversations, as this present national focus on racism is ripe with layer and nuance.
Murray's latest work, to your point, compellingly presented that the numbers say relative to population-level IQ. The good and hyper-rational news is the vast majority of all four populations (Murray's grouping) are of sufficient and similar IQ such that mental agility (what IQ measures) is NOT a deciding factor for a successful life or community. Other factors such as early childhood education, two-parent present families, delaying children until after sufficient education, property-secure neighborhoods -- other social hygiene factors matter as much or more than mental agility. That is the space where most of the human population lives, works, and plays.
However, the intractable and stark reality about mental agility population-level statistics concludes the distribution -- the right-size high standard deviation outliers correlate with ethnic population membership. The "high right-side," extremely high mental agility groups will have, proportionally and in descending order, Asian, European, Latin, then African members. Therefore, in a just and fair and meritocratic community, high mental-agility professions will populate accordingly. To demand some other proportionality in the spirit of fairness or equity is not only irrational, it is societal suicide as it demands we refuse the better qualified and capable and reward the less qualified and incapable. It is an attempt to thwart our evolutionary imperative of natural selection and survival of the fittest. It ain't gonna work, and in the spirit of Jerod Diamond's "Guns, Germs, and Steel", it will be only a matter of time when an adjacent tribe who doesn't not play this foolish woke fiction will conquer us, sooner or later.
Like China.
Added: I forgot that Murray had posted a spreadsheet with his data, so mea culpa there. But this still doesn't excuse leaving the numbers out of the book, and his case looked even worse when I re-examined his data.
Note: As for the low sample size, when he said "6 or fewer," he wasn't referring to total numbers of arrests, but of arrests for individual groups. His explanation makes no sense to me in light of the spreadsheet, and I'm just going to move past it.
Very important: he left 5 cities out of the spreadsheet (confirmed in a pdf with his notes that lists only 8 cities). This means that the values for Chicago, Baltimore, Lincoln, Ashevelle, and Urbana are unsupported by data. This is Bad. Throw away his ratios for those cities.
Further, the values for Chandler are based on only 13 arrests. I calculated p-values (recall: you want them to be <0.05, meaning there's a 5% or less probability the finding was due to chance). The p-value for European:Black was ~0.7. meaning there's a roughly 70% probability it was due to chance. Interpretation: "beyond the realm of wishful thinking." The value for European:Hispanic was ~0.2. Perhaps not beyond the realm of wishful thinking, but still solidly in the land of hooey. Okay, throw out Chandler.
Well.
This leaves us with seven cities, one of which is New York.
Remember what I said about sample bias, when the sample is dominated by one component? Well, New York had ~13,000 arrests for murder. The next closest was LA, with ~2,100. After that came DC with 565. The next four don't even add up to half of DC. This situation is called undercoverage.
When you do a statistical analysis, you want your sample to represent the population as a whole. If it's dominated by one group, this is generally not the case. This is why the FBI database is so wonderful. It has data from all 50 states plus Puerto Rico and DC (see for example table 4 in the 2019 section of the website).*
Conclusion: the entirety of chapter 4 and anything derived from it (e.g. chapter 6) are bogosity.
*https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-2019/topic-pages/tables/table-4
I just listened to a conversation between Coleman Young and Charles Murray on Murray’s new book on Coleman’s podcast. It was respectful but Coleman was quite persistent in challenging the implications of Murray’s work if it were to go mainstream—I would recommend people listen to it.
One conclusion I came away with is that Murray’s book is a response to the increasingly dominant anti racist narrative of CRT. His concern seems to be that embracing a position that promotes that America is systematically racist, that whites as a group are essentially racist, and that black peoples’ shortcomings, whether it is higher crime rates or lower intellectual accomplishments are a function of white racism—I.e., it’s not our fault—can only lead to a destructive backlash from whites and possibly other nonblack citizens. This backlash would lead to all citizens seeing themselves not as individual citizens who judge each other based on their talents and accomplishments but as members of racial and ethnic identity groups.
This would be the end of the American experiment. Murray wants to fight against the charge of systematic racism by arguing the empirically observed disparities between blacks, whites and, other groups are not the result of racism but primarily the outgrowth of differences in cognitive ability. He is less clear about whether higher violent crime rates among blacks are the result of cognitive differences or some other factor—e.g., cultural values. (As Val’s post below indicates Murray’s calculations on black-white and Hispanic-white crime differences are flawed because of his use of small sample sizes taken at the city level.)
I share Murray’s concerns about anti racism and identity politics threat to our Republic. However, it seems to me it’s a bit of a cop out to say that the only way to fight these destructive forces is to say that blacks’ disparities are the result of their cognitive shortcomings. This means there is no closing the gaps either through public policy or by groups pursuing a self development strategy. It strikes me that Murray’s position if embraced widely would lead to exactly the outcome he fears. It would underscore the idea in society that blacks are an inferior identity group that must be controlled.
I spent three years teaching science and math to high schoolers in Botswana in the 1980s. I had the full range of kids from brilliant to dullard. By and large, the kids that did well and could think analytically, came from families that had more exposure to technology and the outside world, or, whose parents were clearly pushing their kids to excel. There were exceptions, but it appeared clear to me that culture and expectations play huge roles in the average kid's intellectual development.
I spent much of yesterday trying to sift through data on crime statistics and ended up feeling dazed and confused by the variety of data sources and the difficulty of sifting through it all. Even the two sources Murray cites in his book are what I’d call databases of databases, and rooting through can feel overwhelming.
But then, today … I found the FBI crime statistics, and I felt dazed and confused for different reasons.
The FBI takes in data from thousands of agencies large and small, and assembles it all into one very tidy, very accessible pile. Arrests for serious crimes are separated into 31 categories that are broken down by year, race, and ethnicity.
For example, the 2019 data came from over 12,000 agencies representing ~250 million people, or about 75% of the total US population. Why would you spend weeks or months digging through databases of databases, only to end up with information from 13 cities/18 million people, when you could have 13-14 times that number of nationwide statistics, all in nice tables? You wouldn’t be forced to cull information on one or two things, because there are lots of categories right there, just waiting for you to look at them.
Why didn’t Charles Murray use the FBI data? His explanation on page 137 is weak at best. He claims you need to have city-specific arrest data paired with city-specific demographic data. But if a dataset represents 75% of the nation, you can use demographic data for the nation as a whole. This approach, combined with the large number of categories, makes for a much more generalizable conclusion than data from Fayetteville and New York and a relative handful of other places.
And what did I learn?
Summary: things do not break down into his categories (most violent first):
1. Black
2. Hispanic
3. White
4. Asian
True: Asian people did have fewer arrests overall for most categories (MOST). Beyond that, what I found was exactly what I’d expect from any large dataset; namely, there are some patterns, but also lots of variation from one category to the next. It’s not as simple as Murray implies.
And here’s my biggest challenge with this: I’ll lay it out findings and a discussion as succinctly as I can in the next comments, but if you want to understand what I’m writing, you’ll have to pay attention to some detail. The flip side is that Facing Reality can be summarized in one sentence: Blacks and Hispanics have lower IQs and commit more violent crimes. This is a major problem faced by people who struggle with misinformation: reality is usually nuanced and detailed, but misinformation and distortions can be shiny and pretty to the person spinning them. Such is the case with Facing Reality.
Summary of findings about murder arrests from FBI database:
I calculated murder arrests per 100,000, in 2018 just like Murray, and found the following:
Overall: 2.7 murders per 100,000
Black Americans: ...10.9 per 100K;
Hispanic Americans: ...2.4 “
White Americans: ...2.0 “
……………………………………FBI……..Murray
Black:White ratio:………. 5.5…….. 21.0
Hispanic:White ratio: ….1.2…….. 4.9
I calculated a comparison of two rates and got p-values <0.001, with values <0.05 indicating statistical significance. So it’s true that the value for Black Americans is much higher. But the value from the big FBI database isn’t even one quarter of Murray’s value, and the ratio for Hispanics is close to 1.
(The numbers for 2019 are B:White = 5.0; H:White = 2.0; for 2016, they were 5.3 and 1.1, respectively)
And now the background:
I believe that Charles Murray made a terrible mistake. It’s the kind of mistake that’s plagued drug development and given us expensive epic fails in clinical trials: a small sample size in your preliminary work can show false trends. And Murray’s data is just that: “preliminary.”
Recall that he didn’t say how many total murder arrests he was working with. I suspect his sample size was pretty small. This is because he said that some cites had 6 or fewer arrests over the entirety of the database and others had “at least 11.” The FBI had 8,957 arrests (from 12,000+ agencies) with known race in 2018 alone, and knew Hispanic/non-Hispanic ethnicity in 7,050 of them.
I can tell you with confidence that conclusions from small datasets can be horribly misleading. When you can count your data points on your fingers and toes, as is the case for at least some of his cities, as few as ONE new outlying data point can make a big change in your numbers. This means that Murray’s ratios for four cities are essentially meaningless, leaving 10 cities for Black Americans and 4 for Hispanics. And at least city one has only 11 data points.
So not only is Murray making generalizations about the behavior of millions of people based on arrests for one crime — murder — he’s doing so on scanty data.
Murray is intelligent and highly accomplished, but he’s not a scientist or someone skilled in statistical analysis, and this really shows in the book. He’s very good at making a case with what he’s got, but he doesn’t seem to see that what he’s got isn’t enough to make generalizations. If this book had been a paper submitted for peer review, the reviewers would have been all over it regarding total lack of statistical analysis, poor or no citation of sources, and lack of statements about sample size and the limitations of a small sample size and biased samples (means that if one or two of his cities had many more murders than the others, they would have had an outsized influence on his results…but he didn’t tell us, so we don’t know).
FBI databases: https://ucr.fbi.gov/crime-in-the-u.s
Regarding distortions and misinformation
Facing Reality doesn’t push flat-out misinformation along the lines of “vaccines cause autism.” It’s more that it distorts along the lines of something like “vaccines have mercury, which is toxic.” Sure mercury is toxic, but the amount of mercury in vaccines is only toxic to bacteria and fungi, which was the whole point of adding thiomersal. And besides, thiomersal is converted ethylmercury, which doesn’t accumulate in your body. But the original sentence IS true: mercury is toxic. It just doesn’t bring in any detail. Instead, it quietly encourages people to fill in gaps along the lines of, thiomersal = toxic vaccines = brain damage in me or my kid.
Murray does essentially the same thing, though he may have made a genuine mistake. He makes some true-ish statements, and readers who didn’t dig through the FBI database may fill in the gaps, along the lines of “Black Americans are 21 times more likely to kill people!”
Readers shouldn’t have to dig through an FBI database to know that the data doesn’t support this idea. Murray should have done that, and he didn’t. He just made a weak-to-poor justification for not using it. I’m not accusing him of malfeasance or RACISM! Again, it’s possible the problem is a genuine mistake based on lack of knowledge. What bothers me is the aura of expertise.
Research is complicated. It involves a lot of nuance, a lot of detail, and a lot of apparent contradictions — and sometimes other factors. Figuring it all out takes a lot of thinking and a lot of time (think about how long it took to sort out the whole sun-earth thing). The methods you have to use as a responsible scientist make this process very difficult.
And herein is a big red flag about Facing Reality: it’s a quick and easy read (Murray even says so on page VII). He took a very complex subject (crime) and told us we can reduce it to result arrests for murder in a handful of cities. He drew vast conclusions from this narrow set of data — and he gave almost no background information. He also ignored other forms of crime almost completely, as though they don’t count. More on that tomorrow or Monday if I have the energy.
The worst thing here is that when an idea gets a) repeated for a long time and b) politicized, people choose up sides and dig in (as was the case with the earth and the sun). This makes us even less likely to read a detailed counterargument to conventional wisdom (and more likely to get angry). We need to get past this. Somehow.
And if we don’t talk about WHY problems develop, we will never be able to solve them. That is another fatal weakness in Facing Reality.
I am tired now and will write even more next time. Next up will be some summaries showing other findings from the FBI database. Preview: what I found is different from what Facing Reality implies in Chapter 4.
I appreciate Dr. McWhorter's deep reading of Dr. Murray's latest book. He bent over backwards to be fair-minded. Not sure Murray deserves it. I wouldn't call Murray a brilliant thinker. He says things no one else says, and that counts for originality and courage in some quarters. His book might seem substantial up against books by, say, Robin D'Angelo or Ibram Kendi, he certainly has gathered more data, but his conclusions are, if not racist, than poorly considered.
To me the right answer is do nothing since all that data in aggregate does not inform on the individual. It is the individual that is doing the job, research, and analyzing. Being successful at a any of those is about much, much more than raw intelligence, it is about character traits like drive, work ethic, team work, and many more. Intelligence is just one element and we all know very "sharp" people that are lazy, lack focus and as such can't do the 'analytical' work either. We must focus on the individual and treat each as unique. Sure collect data to find gatekeeper that are grading on account of race, which simply is not a factor in the ABILITY to be successful, and remove that gatekeeper from such responsibilities.
(corrected revised post)
for some background reading : Charles Murray answers criticisms of "The Bell Curve".
Commentary magazine: May, 1995 :
https://www.commentarymagazine.com/articles/charles-murray/the-bell-curve-and-its-critics/
Someone, perhaps it was Murray himself, has said that, as societies gain in general practical fairness--that is, the less invidiously discriminatory they become --the more discrepancies shall appear in these expectations for equitable _outcomes_. That is, in other words, inherent differences in talents and abilities, given greater scope by fairer practices as prejudicial practices are eliminated, shall _increase_ rather than decrease the occasion for casual observers' perceived _inequality_ in precisely these expectations for more "even distributions" in _outcomes_-based “equity”.
The same things were said of the Jews, had the test existed it would have been said of the Irish.
Binet was testing for learning disabilities for the French school system. When he saw where the Eugenics school was taking his test he mysteriously dies, perhaps of a broken heart…but possibly suicide.
Binet was Catholic. At the time suicides we’re damned and could not be buried in blessed ground, as we say “his death was doubtful “.
The test is FALSE, its perpetrators FRAUDS.
And Intelligence is not Virtue.
Intelligence will allow you to succeed at TASKS. To perform DUTIES requires Virtue- Courage. Something no one would accuse America’s Elites of, but they might see it in Black Men.
So you can put down the axe and let go of the Rack Professor Procrustes McWhorter - you’ve been had.
Its all fraud. As is Murray’s bogeymen.
I have been selectively breeding crops for several decades. Facing data showing the identical ranking of different varieties of a crop in different environments and different years, it would be clear that something is being inherited differently by the different varieties. We seem to be discussing something parallel here. However, I also share John's gut feeling that culture, not genes is involved.
This is no contradiction. Plants transmit many things to their progeny and "genes" (DNA from the plant cell nucleus) is certainly important but is not the only category. Plants transmit cellular contents, whole organelles (e.g. chloroplasts), symbiotic fungi or bacteria, pathogens, nutrients, RNA and epigenetic "marks" (e.g. gene silencing) to progeny. Humans transmit these and, usually, culture, language, religion, property, caste, family names, music, money, debt, etc. Many of these forms of inheritance are much less predictable than genes and the thing transmitted may have much, much less fidelity than a DNA sequence. But collectively these inheritances can greatly affect an individual's success and even interact with the expression of one's genetic endowment.
I agree with John that it is essential to go beyond diagnosing that "something is inherited". Some folks argue that what "the thing" inherited is institutions and political system (e.g., institutional racism). McWhorter and Loury seem to be arguing that institutional and personal racism is plastic and has mutated to much less virulent forms in recent generations. They seem to find the contribution of these factors to be insufficient to explain the persistence of heritable differences between races.
On the other hand, it is really hard for me to imagine any human evolutionary trajectory where intelligence wouldn't have had extremely high fitness value. And even if there had been a slight difference in gene frequency, would it have been preserved after all of the gene flow between previously isolated groups that has happened in the Americas?
John's "victim syndrome" could certainly be heritable and could be present at significantly different frequencies in different populations--and influence test-taking phenotypes. But so could something as banal (and potentially treatable) as gut microbiome. I've often wondered if the idea of free will is inherited and is highly influential in childhood cognitive development--whether or not free will is "true" in a metaphysical sense.
The root problem is that demographic data are simply not accurate enough to support firm conclusions. Any first-order conclusions based on even mildly inaccurate data cannot be trusted and if someone is making higher order conclusions then they are essentially practicing astrology.
Here are two references (of many that exist) documenting errors of 15-30% in ethnic group in one case and diagnosis in another. Inaccurate data is great for coffee-table discussions on the "meaning" but should have no place in rigorous scientific discussions.
That means any large-scale "results" are not scientific. We need to accept this as a fact and deal with that, instead of let our desire to get results cause us to be delusional.