So Just What did You Expect, Again?

Via a “share” from a Facebook friend’s page, we have this gem over at AlterNet.org.  I will observe that this is not the first article at that site that my friend has shared, on FB or otherwise.  I find the self-consciously cute name they’ve chosen for their site to be more than a bit ironic.  You see, it’s a play on the world “alternate” from which we are to deduce “alternative,” from which we are to conclude that this site is purveying news and opinion that’s somehow edgy, “alternative,” out there, or otherwise not just one more dead-fish organization going with the flow of the stream.  Except it is; what you’ll find there is pretty standard left-extremist claptrap. Like, for example, the linked article. We are all racists now, it seems.  As evidence for “subconscious racial bias” arising from “the most enduring, corrosive racial stereotype in America: the black-as-criminal mindset,” we have the observation:

“The archetype is so prevalent that the majority of whites and African Americans agreed with the statement “blacks are aggressive or violent” in a national survey.  In support of these findings, other research indicates that the public generally associates violent street crime with African Americans. Other nationwide research has shown that the public perceives that blacks are involved in a greater percentage of violent crime than official statistics indicate they actually are.”

There are two links in the just-quoted text at AlterNet.org, both to an article from 2007 over at the Journal of Contemporary Criminal Justice.  For the first link, no specific numbers are given in the linked JCCJ article, only the characterization “clear majority” of both groups.  As to the lattermost assertion, I could not find in the linked article on the public perception of percentage or prevalence of black crime relative to actual proportion of all crimes committed by blacks.  What I did find are some specific numbers of all crime committed by blacks relative to their statistical share of the gross population (remember that this article is now seven years old, and the research it’s based on presumably older still, so these numbers may well be out to lunch in one or more respects by now).  From the JCCJ:

Blacks are indeed involved in a disproportional amount of crime in general and violent crime in particular.  In fact, for violent crimes such as robbery and homicide, there have been times when Blacks were arrested in absolute numbers that surpassed those of Whites.  In more recent years, however, although Blacks did not surpass the actual number of Whites in nationwide arrests, their presence in these statistics has been greater than their representation in the general public. For example, although Blacks compose approximately 13% of the U.S. population, in 2002 they accounted for 38% of arrests for violent crimes and nearly 30% of arrests for property crimes. Juvenile arrest statistics indicate that during the same year, Black youth accounted for approximately 43% of arrests for violent crimes and 27% of arrests for property crimes. Researchers have suggested that crime committed by African Americans may be especially salient not only because it exceeds what would be expected based on the racial composition of the country but also perhaps because the violent crimes that tend to be most fearsome are the ones that are most disproportionately perpetrated by Black males.

Let’s see.  Black males constitute roughly 6.5% of the total population (half of 13% is female, right?).  Given that the overwhelming proportion of crime in general, and physically violent crime in particular, is committed by males, period, we can assume that somewhere north of 33% of violent crimes were accounted for by that 6.5%.  Which is also to say that black males (and the JCCJ article is predominantly about black males) account for violent crimes at over five times what you would expect if crime statistics were evenly distributed across all demographic groups (male/female, age, ethnicity, origin, etc.), and almost five times the rate for property crimes.  Remember that crime in general and violent crime in particular is not age-neutral; it skews strongly towards youth.  The numbers for black youth are even more alarming.  I haven’t seen an age-pyramid for black youth, but since birth rates trend negatively with increasing wealth, I’m going to assume that black youth accounts for something north of 13% of all youth.  Let’s assume 18% of all youth is black, making 9% of all youth both male and black.  That 9% of youth accounts for violent youthful offense at a rate 4.78 times their “statistical expectation,” and exactly triple the rate for property crimes.

Thus in point of fact just as a perception of statistical reality, the popular perception that blacks are more prone to commit violent crimes (or even property crimes) would coincide exactly with observable data.  I wish the JCCJ article had broken some of the numbers out in greater detail (alas! there are no hyperlinks in it).  I’d be interested to know what that “percentage of crimes committed by” figure looks like when you add to the question the qualifier “as experienced by members of specific groups.”  Thus, what percentage of violent crimes committed against blacks are committed by other blacks?  And the same question for whites (and East Asians, and South Asians, and Aboriginal Americans, for that matter).  From everything I’ve ever heard, the answer to that question, percentage of black victims of black crime, the number approaches depressingly close to 100%.  Small wonder that blacks might perceive each other to be prone to violence, when almost all the violence they experience is in fact at the hands of their own ethnic group.  And in fact the overwhelming percentage, from everything I’ve heard, of black violent crime in general is directed at other blacks.  Meaning that you’d expect white victims of black crime to be a smaller percentage than that 38%.  This would, again, match everything I’ve ever read, namely that all ethnic groups experience violence principally from members of their own group.  It wouldn’t surprise me at all if the proportion of white victims of black crime were something less than 13%.  So why might whites in general entertain that non-statistically-valid perception (as to themselves only)?  Unfortunately I don’t have hard numbers, but my understanding is that to the extent that whites experience violence from outside their own group, that violence comes nearly exclusively from blacks.  As a matter of logic that doesn’t make a whole lot of sense to extrapolate from that data point.  If I’m getting robbed at gun-point, or if a friend of mine has been robbed, what difference does it make what color skin the perp had?  But humans aren’t always the most logical creatures.  As a human child growing up on the playground I’ve probably been bitten by more humans than dogs, if you were to go back and count.  But the one bite of my life I still remember is when the German shepherd sank her teeth into my butt, way back in the early 1970s.

You won’t find much cogitation in that AlterNet.org article along the lines of the preceding paragraph.  What you will find are outright misrepresentations along the lines of, “Remember Zimmerman’s false syllogism?  A few blacks committed burglary, Trayvon was black, therefore Trayvon was a criminal.”  No.  What came out at the trial is that the housing development where Zimmerman lived had specifically been the subject of multiple break-ins, at least some of which George Zimmerman had observed, and at a minimum those which he’d observed had been committed by black males (my understanding is that to the extent the race and sex of the other perps were known at all, it was black and male).  Martin (you remember him; he was the one who was trying to splatter George Zimmerman’s brains onto the sidewalk) was observed by Zimmerman, wandering in the rain, pausing and looking into windows of housing units.  Martin may have been lost or disoriented, or just curious as to what sort of people lived in the place he was visiting.  But from Zimmerman’s perspective it looked like someone casing the joint.  And that’s how he reported it.  Unlike the 911 transcript fraudulently edited by the news networks, it wasn’t Zimmerman who brought up Martin’s skin color.  He didn’t mention skin color until he was specifically asked about it.  The “syllogism” claimed is simply bullshit.

In what she no doubts prides herself on as her demonstration, our author starts with the usual recital of America’s foundations in slavery, and the post-slavery history of violence against blacks committed by whites, in the form of lynching.  [A couple of observations are here in order.  For starts, given the explicitly racist practice of most law enforcement until the 1960s, you have to assume that for most of American history the vast majority of violent crime against blacks, committed by anyone, never made it into the official numbers.  They were sub-humans, so who cared if they were robbed, beaten, murdered, stabbed, raped, etc?  A lynching gets attention; knifing someone in a bar fight over a woman, not so much.  Secondly, given how geographically concentrated the black population was until post-1910, you have to assume that black-on-white crime was vanishingly rare.]

As Gentle Reader might suppose, there are pretty detailed data on lynchings by year, and in fact by race as well.  Here’s a tabulation maintained by the Tuskeegee Institute, for 1882 through 1968.  Not that it matters a hill of beans for this discussion, but you could have won some money off me betting that the number of white victims would have exceeded the number of black victims for any year at all . . . and yet for the first four years that’s exactly what happened.  Look at the total for both races for the 83 years from 1882 until the last recorded, in 1964: 4,742, of whom 3,445 or 72.649% were black.  To put some perspective on it:  That’s only 500 more total victims that the number of race-unknown homicide offenders in 2010 alone (links to FBI and Census Bureau data below), and it’s less than the white offenders for 2010 and less than the black offenders for 2010.  It’s not quite 36.5% of the victims for 2010 alone.  To put some even more distressing perspective on that:  Across the entire Reconstruction and Jim Crow eras, not quite 73% of all lynching victims were black; in 2010 alone, 49.78% of all murder victims were black.  Even after a good 15 or more years of dropping violent crime statistics, we’ve got a problem that’s two-thirds as exclusively black as lynching.  Someone remind me again why this isn’t getting more play in the lamestream media.

I think it’s pretty safe to assume that the number of black lynchers was zero, so you’ve got attribute all of those victims, white and black, to white perps exclusively.  But how many “offenders” were there?  I don’t think it’s appropriate to ask just how many people were actually involved in making the noose, tying the victim, looping the rope over the tree or lamp post, or whatever.  I think you have to attribute some moral guilt to to at least some of those who showed up, even if only out of curiosity.  I have no idea whether anyone has ever even attempted to figure out how many people attended these things.  How many of them drew a large crowd for a small town?  How many of them were just a couple or ten people in the dark of a night?  How many drew a crowd of thousands, as several well-known lynchings did?  So let’s just assume an “average crowd” of 750.  Gentle Reader is reminded how few places in the South during the years when most of the lynchings occurred (out of the 4,742 shown, 2,359 or 49.7469% had occurred by 1896, and 3,179 or 67.03922% had occurred by 1903 — whatever else it was, lynching as a widespread problem was overwhelmingly concentrated in the pre-World War I South, even though other states also knew it; for example, the lynching that prompted the poem “Strange Fruit” went down in Marion, Indiana, and even Minnesota can show at least one, of some circus hands) had other than minuscule populations.  So I don’t think 750 people is an unreasonably small number.  Applying that across 4,742 lynchings produces 3,556,500 “offenders,” and that’s if you consider all attendees equally guilty.  Now let’s ask ourselves how many tens of millions of people were living in the South during those 83 years.  I suppose a statistician could cipher that out, but I’d be amazed if the number was any less than two hundred-plus million.  The 1900 census data show 18,975,665 people living in the eleven states that had seceded, out of total population of 76,212,168; that’s 24.89847% of the gross.  In 1900 there were 115 lynchings.  Even if you assume an average crowd of 1,000 per, (and I think that’s a grotesque over-estimate) and even if you attribute all lynchings to those states, that gets you to 115,000 people or slightly over six-tenths of one percent of the gross population.  And yet we have tripe (on AlterNet.org, no less), such as the bilge I defenestrated here, in which the entire South is lumped into a single, seething, bloodthirsty mass.  Remind me again why this ahistorical bullshit is considered insightful analysis, and yet it’s conclusive evidence of racism! when popular perceptions of the prevalence of violence match observable statistics.

Just as an exercise, I spent some time looking for data on homicides, age, race, and total population.  I also looked for data on mass killings (most things I’ve run across define a “mass killing” as one where there are more than three victims in the same (e.g. Oklahoma City) or a closely-related sequence of killings (e.g. Virginia Tech)).  I also tried to tie the data I found to the same year, since things can change radically and very quickly.  Remember that 2007 data above?  Well, the one thing that’s been happening in the past six years is that violent crime of all kinds, and so far as I know, across all groups, has plummeted.  So even if a particular group X is “more disposed to violent crime than statistically predicted,” over the last six years they’ve got significantly less disposed to it.  In order not to spend more time than I have, I confined what I was looking at to homicide, since it’s the hardest to conceal and the most likely to be pretty fully reported.  I settled on the year 2010 so I could use the 2010 census data, available here.  The FBI has homicide data, both as to victims and perps, by age and race, here.  Finally, I couldn’t seem to find “official” numbers on mass killings, but Mother Jones has a tabulation on “US Mass Shootings, 1982-2012.”

A couple of cautionary notes about the data.  The census bureau reports, for racial self-identification, not only single-race responses but multiple-race responses.  The distinction can be significant when you’re talking about a gross population, as of Census Day 2010, of 308,745,538.  A total of 38,929,319 self-identified as black only.  That’s 12.60887% of the gross population.  But 42,020,743 self-identified as black-alone-or-in-combination.  That’s 13.610% of the population.  Unfortunately I could not find age distribution data for the 3,091,424 who self-identified as black-and-something-else, so I had to apply the percentage distributions of the age brackets reported to the larger number.  That requires some assumptions about birth patterns for which I have no support in the data I could find.  Secondly, there is a large difference in the FBI’s data between homicide victims (12,996) and homicide offenders (15,094).  On the other hand that’s logical because killing someone is a sufficiently egregious act that for a not insignificant number of perps it’s not something they’ve got the guts to do alone.  Additionally, while data is pretty comprehensive on the sex, race, and age of the victims (e.g., out of the 12,996, only 152 or just over 1% are shown as age-unknown), you’ve got to bear in mind that out of the 15,094 perps, 4,224 are shown as race-unknown; that’s 27.98% of the total.  There is enormous room for conclusions to move.  Just by way of extreme example, if you attribute all those unknowns to whites, you get 60.11% of homicides done by whites; if you attribute 13.61% of them (575) to blacks, you get 6,345 done by blacks, or 42.04% of the total.  In addition to Mother Jones’s data being non-verified (although they’ve got zero reason to understate any of it, with their known political/policy affinities), it’s only mass shootings, which is of course a subset of mass killings.  So it’s not complete (see Oklahoma City and its 168 dead); on the other hand, it’s jolly hard to kill more than one person with a knife, baseball bat, or claw hammer (by the way, although not relevant to our current discussion, Gentle Reader ought also bear in mind that blunt instruments are used to kill more people each year in the U.S. than firearms of all kinds), so any discrepancies are unlikely to be very large.

With all that in front of us, let’s look at the data.  First raw numbers.  Of the 15,094 homicide perps, 5,770 were black (more ominously, of the 12,996 victims, 6,470 were black, a catastrophic 49.78%, and for the age brackets between 17 and 39, blacks made up over 50% of the victims in every stinkin’ one of them), which is 38.23% of the total.  Whites, by the way, accounted for 32.13% (please to remember the race-unknowns, Best Beloved).  The overwhelming majority of all perps for whom sex is known were male (ex: of the 20-24 perps, 2,546 total, 2,315 of them were male, 90.93%; the divide hovers around 90% male for every single age bracket).  So our first conclusion stares us in the face:  If you want to be afraid of someone killing you, be afraid of a generic male.  If you want to assume that someone is violent, assume it’s a male.

But everyone (except perhaps the people who write for, and read, AlterNet.org) realizes that homicide is not evenly distributed across age, either.  For both white and black, it’s massively concentrated in the ages 15-40.  Let’s look:  For whites, 66.23% of their total (4,849) is accounted for in the 17 to 39 age brackets; for blacks the number is very similar: 72.65%.  Second conclusion:  If you’re going to be frightened of a putative murderer, don’t imagine him with a whole lot of gray hair.  By the way, it appears that whites remain more violent longer than their black age cohort.  The last black age bracket showing more than 100 perps is 50-54 (129); whites keep killing a full decade longer, until the 60-64 bracket (112).

Knowing that we’re discussing a perception issue here, and further knowing that whatever we perceive to be the level of violence associated with either race (high, medium, low), our perception is going to have to be grounded in reality, to the extent it is at all, in the data for males ages 15-40 of both groups.  So let’s see how that shakes out.  Applying the age bracket proportions for black-only to the black-alone-or-in-combination figure produces a total of 7,670,023 males in the 15-39 age range, which works out to be 2.48425% of the 308,745,538 gross population.  Now let’s compare that to the percentage of homicide offenders in the 17-39 age range (the FBI’s next lower age bracket is 13-16, and among blacks they account for only 265 of the 15,094, so I feel comfortable ignoring them here).  The black 17-39 age range accounts for 4,192 of the 15,094 homicide offenders, or 27.77% of the total.  Let’s juxtapose that even closer:  2.48% of the population is accounting for 27.77% of the killers, more than ten times their “statistical expectancy.”

In the interest of comparison, using the same extrapolation of age brackets for white-only to white-in-combination produces 37,210,162 white males age 15-39, or 12.052% of the gross population.  White homicide offenders in the 17-39 age range account for 3,212 of the 15,094 offenders, or 21.280% of the total.  Again the side-by-side:  12.05% of the population is accounting for 21.28% of the killers, not quite twice their “statistical expectancy.”

So as to both racial groups, their young males are statistically over-represented among killers, with the degree of over-representation being about five times higher among blacks.  A further point of commonality is that among that 17-39 range, the bulk of the killers are concentrated in the 17-29 range and the deadliest single bracket is 20-24.  For both races you’ve still got guys committing murder into their 30s, but they’ve started to taper off (most likely because they’ve been caught and are in the criminal justice system somewhere).  A point of distinction, however, is how much of each group’s race-in-combination population total is represented by that age range.  Among blacks, 1.852% is represented by males ages 18-19; for whites the figure is 1.381%, a full 25.4% less.  For 20-year-olds the numbers are 0.8943% and 0.6887% respectively, 22.99% less for white males.  In that deadliest, 20-24 bracket, the percentages are 3.95632% for blacks and 3.31933% for whites, a drop of 16.1%.  In other words, in 2010 a greater proportion of blacks were concentrated in the age and sex group most likely to become homicide offenders.  The black population is both younger and in the highest-risk group more heavily male.  That’s going to skew the numbers somewhat.

And at this point we run out of the purely numeric and shade into the concept of the “reasonable.”  Is it unreasonable, when two groups are both prone to excessive behavior on any scale, but one group is five times further out on that scale, that popular perception — unscientific as it always will be — will still reflect that?  Would it be unreasonable for someone to conclude that, all else being equal, blacks make better athletes, when the four data points are proportions of black males versus white males in the NBA and the NFL?  You can debate all damned day long about the why it should be so, but to argue that it’s not so is just damned foolish.

But Miss AlterNet.org isn’t arguing that.  She’s not arguing that blacks overall or black males in particular are not statistically more prone to acts of violence than whites.  She’s not impugning the numbers; she impugning the perceptions.  She’s arguing that because the “most horrific” crimes are committed by whites, and we (as a country) don’t perceive whites in general to be disproportionately violent, that’s evidence that we’re all racists.  Let’s tee up Mother Jones, bearing in mind my caveats above.  Looking broadly at their spreadsheet, it rapidly becomes apparent that mass shootings are (i) predominantly a white phenomenon, (ii) overwhelmingly a white male phenomenon, and (iii) by and large a crazy white male phenomenon.  But let’s look at just the numbers, ma’am.  In 2010, out of 12,996 homicide victims, we’ve got . . . 9 killed in a mass shooting.  That’s not quite seven-one-hundredths of one percent of the total.  I went back and added up all the mass shooting fatalities (Mother Jones gives numbers of wounded as well, by the way) since 1993, added in the 168 of Oklahoma City (but excluded the 3,000+ of September 11) and came up with . . . 588.  Thus, if you go back a full 21 years, you get 4.5245% of one year’s total homicides in the form of crazy white males shooting the place up or blowing up entire buildings.

Let’s go back to our observations about reasonableness.  Sandy Koufax was one of the all-time greats.  Hall of Famer.  Is it reasonable to conclude from his success that Jews make great athletes?  Or how about Croatians?  There are some very good Croatian basketball players, including some as play successfully in the NBA, and they routinely field outstanding teams in international competition.  So we know they make some damned fine basketball players there.  But how reasonable is the conclusion that “Croatians make great athletes” relative to the conclusion “blacks make great athletes,” based solely upon the data point of how many of each are playing in the NBA?  Let’s see . . . the NBA is . .. gosh . . . I don’t know (Wikipedia.org to the rescue: according to them in 2011 the NBA was 78% black and 17% white), really heavily black.  In fact, five times 17% works out to 85%, not much more than 78%.  Huh.

Gentle Reader will tax me with a false equivalence.  The make-up of the NBA and the perception of crime statistics are not the same thing.  Well, yes and no.  Where they both are similar is that both (i) are highly unrepresentative snap-shots of very large population groups, and yet  (ii) are highly visible markers which are flung in our faces remorselessly, and further (iii) represent the extreme point on their respective behavioral spectrum.  To illustrate the first point:  There were 7,670,023 black males in 2010 in the 15-39 age range.  Out of most of that number (17-39) they produced all of 4,192 known homicide offenders, and even if you attribute all the race-unknowns to black males ages 17-39, you get every bit of 9,962 offenders, or 0.12988% of the total in that age range.  Just over one-eighth of a percent of all black males turned out to be killers that year.  Meaning almost ninety-nine and seven-eighths didn’t.  Bearing in mind that even in the 24-hour news cycle there’s exactly X amount of information that can be put out, how reasonable is it to expect that news of a homicide is not going to get pretty good billing?  Although I’ve not crunched the numbers on other crimes of violence or property crimes, is it reasonable or unreasonable to expect that an armed robbery is going to be covered rather than an apartment that got broken into and a computer and some stereo equipment got stolen?  Finally, if killing is the ultimate crime, you must recognize that playing in the NBA is the ultimate in basketball athleticism.

So while it is entirely statistically defensible to state the conclusion “blacks are more likely to be killers than whites,” it’s not only statistically not supportable but morally reprehensible to conclude “blacks are likely to be killers” because neither group is very likely at all to be killers.  Neither.  But remember Mlle AlterNet.org isn’t about facts; she’s about perceptions.  If most Americans get their information, to the extent they get any, from television, and if television portrays only the most egregious events as “news,” and if any particular group X is in fact, undeniable, count-it-up-and-do-the-math fact, vastly disproportionately over-represented in any particular egregious behavior, precisely how is it that you expect such perceptions not to be awakened?  If all I’m shown is X with a smattering of Y thrown in, upon what basis do you conclude that I’m wicked for concluding, “X”?

Yet Mlle AlterNet.org wants me to be a bigot for thinking, “You know, maybe X.”  This passes for thinking nowadays, it seems.  I notice that she is identified as a “legal analyst.”  Good for her, because she’s a lousy statistical analyst.