Mind Scalpel

(Non-humorous screed alert!)

Alright, this one was so damn subtle it almost slipped by me. But not subtle enough!

Currently, there is in at least one East Coast market a radio ad campaign for “Wireless Amber Alerts.” One ad features someone who is obviously supposed to be a police dispatcher making a radio call for an Amber Alert — the abducted child is said to be something like a “six-year-old African American girl.”

The suspect?

A “Caucasian male, 30-35 years old, blue shirt, black jeans, 180 pounds…”.

I’m in favor of the Amber Alert system, and I think the addition of the Wireless Amber Alert text messaging system is a great idea — in fact, I went to the web site to sign up for it.

Then I heard the above radio ad again, a few times.

And I thought, wtf?

Why, exactly, was this particular racial combination (African-American victim; Caucasian suspect/offender) chosen to be the example for the Wireless Amber Alert ad? How often, based on actual statistical evidence, is that particular racial combination of the suspect/offender and victim present?

Instinctively, I felt that this was likely to be an extremely uncommon combination, since:

1. Most juvenile abductions are conducted by a family member, usually a parent;
2. Interracial marriages, particularly marriages between a white husband and a black mother, have been extremely uncommon (so a situation where a white father was abducting his black daughter would also be extremely uncommon); and
3. If someone wanted to abduct a juvenile who was not a family member, it would seem likely that the abductor would choose a juvenile of his own race, to minimize the chance of detection.

A crude, rudimentary statistical analysis (which I’ll outline below, to avoid boring you here), relying on a few publicly-available (though admittedly somewhat aged) reports of United States census data on interracial marriages and Department of Justice studies on juvenile abductions, supported (1) and (2) above. (3) is based solely on logic and inference, as I could not find direct statistics on the frequency of different racial combinations of suspects and victims in juvenile abductions.

I got especially curious because, when I obtained an on-line recording of the Wireless Amber Alert radio spot, the on-line version did not mention the race of the offender. It did mention the race of the victim, but the offender was merely described as having “blue shirt, black jeans.”

That’s it.

Now, I wondered for just a second if my memory was faulty. It isn’t. I’ve since heard the spot several times.

So someone made an active choice either to include the suspect/offender’s race in the radio spot, or to exclude the suspect/offender’s race from the internet spot.

Which is it, and why? Were they worried about the reaction the portrayal of a black victim/black perpetrator would provoke (”you’re demonizing blacks!”)? A white victim/white perpetrator (”you’re ignoring black victims!”)? After all, in a 1999 study the Department of Justice itself estimated that about 25% of victims of all juvenile abductions were black — while about 53% of victims of juvenile abductions were white. So if someone wanted to choose the race of a “typical” abductee, wouldn’t they naturally choose a white victim?

And if, for the sake of argument, they chose to have a black victim portrayed (either randomly, or for some ulterior motive), what is the likelihood that the perpetrator would be white?

Well, if it’s a familial abduction (which the DoJ estimated were most often perpetrated by a parent (53% father, 25% mother)), the chances of it being a white father doing it are really low — only about 2.3% of marriages involving a black wife involved a white husband.

I could not locate statistics on the race of the perpetrators of non-familial abductions (either because that information was not gathered, or it was gathered but not published) — but again, if perpetrators are at all rational, one would think they’d probably want to minimize their obviousness by choosing a victim of their same race — and only about 22% of all juvenile abductions are non-familial abductions (in other words, familial abductions are 78% of all juvenile abductions), and only about 42% of those non-familial abductions involved a black child, so even if every single one of those non-familial abductions of black children were perpetrated by a white male, that would still amount to only around 9% of all juvenile abductions.

So why is it that the ads “just happened” to settle on an atypical racial combination that had a white suspect and a black victim? Hmmmm….

– Mind Scalpel

Addendum (Sources and Analysis)

Family versus Non-family Abductions

According to the Department of Justice’s October 2002 report on the 1999 National Incidence Studies of Missing, Abducted, Runaway and Thrownaway Children (”NISMART”) titled National Estimates of Missing Children: An Overview, there were an estimated 58,200 non-familial juvenile abductions and 203,900 familial juvenile abductions during the study period.

Familial Juvenile Abductions

According to the U.S. Department of Justice’s October 2002 report titled “Children Abducted by Family Members: National Estimates and Characteristics,” about 12% of familial juvenile abductions were of black children (see Table 3 — the report has a disclaimer noting that that estimate is unreliable because there were too few sample cases to be reliable — but totaling up the other “reliable” numbers leaves the maximum percentage of black child victims being 20% of familial abductions, so we can use that percentage to be conservative).

Non-familial Juvenile Abductions

A study of non-familial juvenile abductions, interestingly, indicates that 42% of them were of black children — more than any other racial group. That might appear to explain the ad’s choice of the victim’s race - until you remember that there were 203,900 familial abductions but only 58,200 non-familial abductions — a ratio of 3.5 to 1 (i.e. only about 22% of all juvenile abductions are non-familial abductions). And by the way, the report notes that that 42% number is based on a sample that’s “too low to be reliable.”

Percentage of Juvenile Abductions with Black Victims

Totaling up the above (20% (remember, we’re being “generous”) of familial juvenile abductions being of black victims * 203,900 familial abductions) + (42% of non-familial juvenile abductions * 58,200 non-familial juvenile abductions) = (24,500 + 40,780) = 65,280 juvenile abductions with black victims. 65,280 / 262,100 total juvenile abductions = 25%. So if you were to choose the race of the victim at random, there’s only a 25% chance that the race of the victim would be black.

Interracial Marriage

According to the U.S. Government’s Census data from 1992, interracial couples accounted for 2.2% of all married couples. 21.2% of those interracial couples were a combination of “Black/White” (you with me so far? so that means .4664% of all married couples are “Black/White” couples). Of those “Black/White” couples, 33.7% of them consisted of a “White husband/Black wife.”

So, doing the math once again, if the frequency of interracial marriages haven’t changed substantially since those census data were gathered, 0.157% of all married couples consist of a “White husband/Black wife.”

As important, there are about 42 marriages with a black husband and a black wife for every interracial marriage with a white husband and a black wife. So the odds, where you start with the presumption of a black child being abducted by her father, of the father being white, are about 42 to 1 (percentage-wise, about 2.3%).

Once you’ve chosen the race of the victim, the next question is whether it’s a familial or a non-familial abduction. About 63% of abductions involving a black victim are familial. If the abduction in the ad was a familial abduction (and here I’m assuming it’s the father doing the abducting), there’s only a 2.3% chance that the father would be white.

There’s a 27% chance that the abduction would be a non-familial one. The available statistics do not make clear what percentage of non-familial abductions are interracial — the only racial statistics given are the race of the victims. But I ask you, if a perpetrator is intent on drawing as little attention to himself as possible, is he really going to choose a victim of a different race? Wouldn’t he think that the odds of detection would be higher? So it stands to reason that the substantial majority of non-familial abductions of black victims would be perpetrated by black offenders; this is not a racist observation — it also stands to reason that the substantial majority of non-familial abductions of white victims would be perpetrated by white offenders.

But let’s go crazy and assume that all non-familial abductions of black juveniles are perpetrated by white male suspects. That still means that only 27% of all abductions of black juveniles would be perpetrated by white suspects. That amounts to about 6.75% of all juvenile abductions.

Limitations and Disclaimers

I use the terms “black” and “white” rather than “African-American” and “Caucasian” for a number of reasons, including 1) those are the categories used in the DoJ reports and the census; and 2) they’re easier to type.

This analysis is subject to all the limitations of the studies from which it is derived; I refer you to the source documents linked above for those limitations.

Some of the above figures and calculations reflect rounding of decimals.

I have not presented other standard statistical tools or measures, including confidence intervals, standard deviations, etc. because they are 1) set forth in the source documents; and 2) IMHO unnecessary to make the general point.

There are a number of assumptions in the above analysis that could affect the precision of the conclusions. I have attempted to set forth the assumptions expressly in the analysis. Other limitations include the fact that neither the age nor the gender of the victim was factored in to the analysis. Also, obviously, the census data are old (from 1992 at the latest), as are the NISMART data (from 1999). Nevertheless, I’m not aware of more comprehensive data that would indicate that things have changed substantially since then for interracial marriage or juvenile abductions.

Another possibility is that the type of abduction actually featured in the radio ad is more likely to be made the subject of an Amber Alert (which begs the question of why that would be the case).

If you argue, without data, that white fathers kidnap their black daughters more frequently than black fathers kidnap black daughters, or that white perpetrators in general kidnap black children more frequently than the available data suggest, then I’d say you’ve pretty much missed the entire point of this post. But hey, if I’ve made a significant math mistake, let me know.

Trackback URI | Comments RSS