Joel Stein Throws Numbers Around

I read several blogs written by mothers of children who have food allergies.
The food allergies are incidental to the purpose of their blogs for all of them.
Nevertheless, I have seen each one need to defend their concerns for their allergic children several times. Most recently, my blog-friend Rational Jenn wrote about the juxtaposition of a (thankfully) false alarm with her son, and a mocking diatribe by LA Times columnist Joel Stein suggesting that allergies to nuts are a figment of the Yuppie-Parent-Who-Wants-Attention's imagination. (You can read Rational Jenn's blog entry I'm Not An Attention-Seeking Psycho by clicking the linked title).

I followed the link Jenn provided to the Joel Stein piece, entitled Nut Allergies--A Yuppie Invention, and I read the column and the comments. According to some of the comments, the Stein column is supposed be witty and humorous. I failed to see either characteristic, but then I believe that humor and wit require more skill than was demonstrated in the writing I found. In fact, the Stein piece did a lot to help me understand why newspapers around the country are crying for a government bail-out. Stein cannot write well enough to clearly explain the facts he was throwing around, and the so-called humor was on the level of the TV sit-com that combines a continuous stream of put-downs and foul language with a laugh track and calls it funny. (This is why I have helped the Feds and the press go into crisis mode because I have not yet modified my TV for HD reception . . . but that's another blog entry).

Since I fail to even begin to see what is funny about Stein's column, I will instead put on my scientist hat and discuss his uncritical and uncomprehending way of throwing scientific language and statistics at the reader. Being an Educational Anarchist, I also discussed how to read and critique this kind of nonsense with my 15 year-old budding scientist, because even really bad writing can be an educational experience. So here we go . . .

NOTE: I will use quite a few biological terms below that are second-nature to me. Since they might not be for you, I will provide definitions at the bottom of the post.

Stein says: "genes don't mutate fast enough to have caused an 18% increase in childhood food allergies between 1997 and 2007 . . ."

There are two problems with this statement.

First, Stein is attributing an increase in childhood food allergies to mutations only, and he is thereby assuming that all food allergies are caused by one gene. In so doing, he is also applying the unstated assumption that no food allergies are polygenic (in which the phenotype expressed is created by a number of genes) or that there are epigenetic factors (the phenotype is created by gene-by-environment interactions). He is also assuming that all food allergy traits found in the population are attributable to genes alone (the heritability is 1), meaning that there is no environmental influence on the gene expression whatsoever. Given what we know about allergic reactions, and the number of proteins involved, as well as the variety of food allergies diagnosed each year, he is almost certainly wrong about every one of these assumptions. Finally, the statement indicates that Stein also assumes that there is no difference in gene frequencies among different populations, an assumption he states again, as we shall see. This statement, as Stein makes it indicates that Stein is ignorant of genetics and of evolutionary theory.

Secondly, Stein not only betrays his ignorance of genetics and evolution here, but he also demonstrates that he does not think critically about the numbers he throws around. Statistics do represent something real, and in bandying them about, it makes sense to think about what they mean. Let's look at the numbers he uses and how he uses them. I will repeat the statement he made to refresh our memories:
". . .genes don't mutate fast enough to have caused an 18% increase in childhood food allergies . . ."

First off, Stein does not provide the mutation rate for the genes in question. This may be because he does not know what genes are mutating, and if there is more than one gene that is mutating, and if so, how many genes he is talking about, and which genes are responsible for each food allergy, and if each gene is solely responsible for that particular food allergy (see above). But even if we give him all of this, his statement implies that an 18% increase in 10 years is a large number. However, to understand whether or not this is true, Stein should have asked himself this simple question: "What is the original number that has been increased by 18% ?" If the number of childhood food allergies is a small to begin with, and if we multiple that number by 18/100 (18%), we would get a very small number that represents the increase.

For example, although I do not know the number of total food allergies in the population, I do know that the number of peanut allergies diagnosed is about 1% of the population in the United States. That's a very small number. 18% of 1% (here you are multiplying two fractions) is smaller still. It is 0.018. This means that the actual increase is 0.118 per 100 individuals. That number is quite small, and since you can expect variance in the number of peanut allergies diagnosed from year to year, such an increase is not even statistically significant.

The statement quoted above also discusses all food allergies and the title of the article refers to nut allergies alone, but in the text Stein switches between these two and allergies to peanuts as if these three terms were interchangable. This verbal sleight of hand reduces the clarity of the writing further, and obfuscates the real meaning of the numbers Stein throws on the page.

Although much of the article is, in Stein's own words, "saying something outrageous to get attention," he does throw another sentence at the page that shows his abyssmal ignorance of population genetics:

"A study of Jews of similar demographics and genetics in Britain and Israel found that British kids were 10 times more likely to have peanut allergies than Israelis."

First, the clarity of the writing suffers in that the reader must assume that he means "British Jews" rather than the more general "British kids." But more importantly, although he states that these two populations have "similar genetics" (by which he likely means that both are Ashkenazi), they are still separate breeding populations. This means that the founding gene pool of each was not identical, and that changes in the gene pool since the divergence between them have taken different trajectories due to natural selection, migration into and out of the gene pool, mutation, and genetic drift. The two populations are therefore diverging genetically with each generation. Joel Stein clearly does not have even an introductory-level understanding of genetics, population genetics and evolution.

And of course, my arguments above are made assuming that the phenotypes for food allergies are the result of genes alone. It is much more certain that they are not, which means that many other factors can account for changes in the frequency of the phenotype in question. Genetic expression in the somatic cell can be influenced by a whole host of environmental conditions that themselves change at the cell level, the organismal level and the population level. And on top of that, the frequency of diagnosis can change dramatically due to better analysis of symptoms, more efficient testing, and more widespread public understanding of a condition. With respect to this last, it is very likely that disparities in access to diagnosis (health insurance coverage, regular physician visits) could account for the class differences seen in the numbers of diagnosed food allergies.

Overall, Stein's column shows his ignorance of science, his woeful innumeracy, and his inability to critically think about what his sources are telling him. He also demonstrates a remarkably mean ill-will towards upper-middle class parents. It could be spite or envy, but whatever it is, the meanness obscures the feeble attempts at humor.

I am glad my local, independent newspaper, The Albuquerque Journal, buys syndicated columnists across the political spectrum that reflect much higher writing standards and much more sophisticated senses of humor. I would certainly not pay to read Stein's drivel.

________________________________________________________________

Definitions:

genotype--the actual genes the individual carries on their DNA

phenotype--how those genes are expressed, or more commonly, what traits the individual displays due to the combination of genes inherited

gene expression--the transcription and translation of the genetic code in the cell, which results in the making of proteins; gene expression is influenced by environmental factors and ultimately results in the phenotype

polygenic--a trait influenced by more than one gene

epigenetics--the combination of genetic and environmental influences on the expression of genes

heritability--the proportion of phenotype in a population that is due to genetics. A heritability of 0 would mean that the trait is question is entirely environmentally determined, and a heritability of 1 means that the trait is entirely determined by genes. Most traits of interest in a population are somewhere in between these two extremes.

allele--an alternative form of a gene that is located at a specific position on a specific chromosome

gene frequency (a.k.a. allelic frequency)--the number of alternative forms of the same gene in a population. If the allelic frequency of a gene is changing in a population, we can say that the population is evolving with respect to that trait.

mutation--a change in a gene (alleles are the result of mutation)

genetic flow--the movement of genes into and out of a breeding population (this occurs by migration of individuals into and out of the population)

genetic drift--the accumulation of random genetic changes that change the gene pool of a population

natural selection--the process whereby heritable traits become more or less common in a population due to their influence on survival and reproduction (some traits may be undergoing selection in an environment whereas others are not); also natural selection acts on the traits--that is the phenotype NOTE: the level of at which natural selection occurs is the population, individuals do not evolve

biological evolution--a change in gene frequencies in a population over time

divergence (divergent evolution)--when one gene pool becomes two due to some physical separation, the combination of natural selection due to different environmental conditions and gene flow, genetic drift, and mutation, changes the gene frequencies in each population independent of the other