Certain death? Black-White death dispersions

New research report, after rumination.

Knowing the exact moment of death is a common fantasy. How would it change your life? Here’s a concrete example: when I got a usually-incurable form of cancer, and the oncologist told me the median survival for my condition was 10 to 20 years, I treated myself to the notion that at least I wasn’t going to the dentist anymore (6 years later, with no detectable cancer, I’m almost ready to give up another precious hour to dentistry).

I assume most people don’t want to die at a young age, but is that because it makes life shorter or because it makes them think about death sooner? When a child discovers a fear of death, isn’t it tempting to say, “don’t worry: you’re not going to die for a long, long time”? The reasonable certainty of long life changes a lot about how we think and interact (one of the many reasons you can’t understand modernity without knowing some basic demography). I wrote in that cancer post, “Nothing aggravates the modern identity like incalculable risk.” I don’t know that’s literally true, but I’m sure there’s some connection between incalculability and aggravation.

Consider people who have to decide whether to get tested for the genetic mutation that causes Huntington’s disease. It’s incurable and strikes in what should be “mid”-life. Among people with a family history of Huntington’s disease, Amy Harmon reported in the New York Times, the younger generation increasingly wants to know:

More informed about the genetics of the disease than any previous generation, they are convinced that they would rather know how many healthy years they have left than wake up one day to find the illness upon them.

The subject of Harmon’s story set to calculating (among other things) whether she’d finish paying off her student loans before her first symptoms appeared.

The personal is demographic

So what is the difference between two populations, one of which has a greater variance in age at death than the other? (In practice, greater variance usually means more early deaths, and the risk of a super long life probably isn’t as disturbing as fear of early death.) Researchers call the prevalence of early death — as distinct from a lower average age at death — “life disparity,” and it probably has a corrosive effect on social life:

Reducing early-life disparities helps people plan their less-uncertain lifetimes. A higher likelihood of surviving to old age makes savings more worthwhile, raises the value of individual and public investments in education and training, and increases the prevalence of long-term relationships. Hence, healthy longevity is a prime driver of a country’s wealth and well-being. While some degree of income inequality might create incentives to work harder, premature deaths bring little benefit and impose major costs. (source)

That’s why reducing life disparity may be as important socially as increasing life expectancy (the two are highly, but not perfectly, correlated).

New research

Consider a new paper in Demography by Glenn Firebaugh and colleagues, “Why Lifespans Are More Variable Among Blacks Than Among Whites in the United States.”

I previously reported on the greater life disparity and lower life expectancy among Blacks than among Whites. Here is Firebaugh et al’s representation of the pattern (the distribution of 100,000 deaths for each group):

Black deaths are earlier, on average, but also more dispersed. The innovation of the paper is that they decompose the difference in dispersion according to the causes of death and the timing of death for each cause. The difference in death timing results from some combination of three patterns. Here’s their figure explaining that (to which I added colors and descriptions, as practice for teaching myself to use an illustration program — click to enlarge):

The overall difference in death timing can result from the same causes of death, with different variance in timing for each around the same mean (spread); different causes of death, but with the same age pattern of death for each cause (allocation); and the same causes of death, but different average age at death for each (timing). Above I said greater variability in life expectancy usually means more early deaths, but with specific causes that’s not necessarily the case. For example, one group might have most of its accidental deaths at young ages, while another has them more spread over the life course.

Overall, the spread effect matters most. They conclude that even if Blacks and Whites died from the same causes, 87% of the difference in death timing would persist because of the greater variance in age at death for every major cause. There are differences in causes, but those mostly offset. Especially dramatic are greater variance in the timing of heart disease (especially for women), cancer, and asthma (presumably more early deaths), The offsetting causes are higher Black rates of homicide (for men) and HIV/AIDS deaths, versus high rates of suicide and accidental deaths among White men (especially drug overdoses).

The higher variance in causes of death seems consistent with problems of disease prevention and disparities in treatment access and quality. (I’m not expert on this stuff, so please don’t take it exclusively from me — read the paywalled paper or check with the authors if you want to pursue this.)

Are these differences in death timing enough to create differences in social life and outlook, or health-related behavior, between these two groups? I don’t know, but it’s worth considering.

Why Don’t Parents Name Their Daughters Mary Anymore?

Originally published on TheAtlantic.com.

Understanding the rapid decline of what was once America’s most popular name

Each year I mark the continued calamitous decline of Mary as a girls’ name in the United States. Not to be over-dramatic, but in the recorded history of names, nothing this catastrophic has ever happened before. Mary was the most common name given to girls every year from the beginning of record-keeping (at least back to 1800) through 1961 (except for a six-year dip to #2, behind Linda).

And then it happened. In 2011, according to the latest report from the Social Security Administration (SSA), Mary fell three more places, to 112th. In absolute numbers, the number of girls given the name Mary at birth has fallen 94 percent since 1961. Here is the trend:

The modernization theory of name trends, advanced most famously by the sociologist Stanley Lieberson, sees the rise of individualism in modern naming practices. “As the role of the extended family, religious rules, and other institutional pressures declines,” he wrote, “choices are increasingly free to be matters of taste.” Mary—both a traditional American name and a symbol religious Christianity—embodies this trend.

The other day I did a search of newspaper birth announcements for the name “Mary,” and turned up a lot of grandmothers named Mary. Here, from a recent day’s birth announcement page from Rock Hill, South Carolina, are three Marys in the grandparent generation, in three different families announcing the births of Mazie, Ja’Nae, and Asani. I diagrammed the family names:

Other generational sequences in recent announcements also mirror the history of common names: Mary–Jennifer–Madelyn; Mary–Ashley–Emily; Mary–Cora–Elizabeth.

In the tradition of treating statistical trends as horse races, I imagine that there is one person named Mary, who is constantly falling behind: first behind Linda, then Lisa, Jennifer, Ashley, Jessica, and so on, all the way to Isabella and now Sophia.

But that’s not how it happens—it just looks that way because of the amazing regularity in human behavior, which produces an orderly succession of names. Incredibly, out of 1.7 million girls’ names recorded by the SSA in 2011, I was able to predict to within 87 how many would be named Mary. By simply taking the number born in 2010 and subtracting the 5-year average decline, I predicted 2,584 would be born; the actual number was 2,671 (an error of 3.3 percent).

Somehow, out of the millions of individual decisions parents make, they produce steady trends like this. (If you’re as amazed as I am, consider a career in sociology! If not, please bear with me.)

So what does the Mary trend mean? First, it’s the growing cultural value of individuality, which leads to increasing diversity. People value names that are uncommon. When Mary last held the number-one spot, in 1961, there were 47,655 girls given that name. Now, out of about the same number of total births, the number-one name (Sophia) was given only 21,695 times. Conformity to tradition has been replaced by conformity to individuality. Being number one for so long ruined Mary for this era.

Second, America’s Christian family standard-bearers are not standing up for Mary anymore. It’s not just that there may be fewer devout Christians, it’s that even they don’t want to sacrifice individuality for a (sorry, it’s not my opinion) boring name like Mary. In 2011 there were more than twice as many Nevaehs (“Heaven” spelled backwards) born as there were Marys. (If there is anything more specific going on within Christianity, please fill me in.)

I’m not here to give advice to people who want to bring back the “traditional family.” But if I were I would recommend putting your names where your tradition is—and producing some more Marys.

There are precedents for bringing names back. My simple linear prediction method fails once in a while, when a name’s trend turns around. The greatest example is probably Emma. Emma was at number three when the SSA records begin, in 1880. She fell almost down to #500 by the 1970s. But after a decade of uncertainty she began a fantastic run, finally reaching number one in 2008.

I don’t know (yet) what makes a name turn around like that. Why Emma and not Mildred or Bertha, both former top-10s who fell into oblivion? But if any name has a chance for a similar resurgence, it might be Mary, at least as long as Christianity keeps hanging around.