Correlation & Causation; the Bad & the Beautiful

The majority of autistic children are diagnosed between the ages of 18 months and 3 years old.*

Children receive many immunizations in this same period of time.

Therefore, it has often been concluded that “vaccines cause autism.”

But this conclusion is a case, and a “rather commonly abused” one, of correlation being mistaken for causation.

Mark C. Chu-Carroll of Science Blogs in a January 25th post on Good Math, Bad Math provides a clear and thoughtful explanation of these terms:

• Correlation…..
• …..”is a linear relationship between two random variables.”
• ……”means that as one value changes, another variable changes in the same way.”

An example of correlation that Chu-Carroll uses is to measure shoe and also the heights of a group of men. As measurements generally show that foot size is in proportion to one’s maximum adult size, he notes that ” if I did a mapping between the size of feet and height in adult men, I would find a strong correlation between height and foot size.” He refers to an actual study in which the length of people’s ring fingers was compared to “the size of the vocabulary that people use in the course of a normal day” (based on a hypothesis regarding the hormones that a fetus is exposed to in the womb) as an example in which there was “no correlation to speak of.”

Causation, on the other hand is not “really a mathematical notion.” Rather:

• Causation
• ……”means that when one value changes, it causes the other to change.”

Chu-Carroll then considers the immunization-autism correlation:

… people see the correlation between receiving immunizations and the diagnosis of autism, and assume that that means that the immunizations cause autism. But in fact, there is no causal linkage. The causal factor in both cases is age [my emphasis]: there is a particular age when a child’s intellectual development reaches a stage when autism becomes obvious; and there is a particular age when certain vaccinations are traditionally given. It just happens that they’re roughly the same age.

Nonetheless, Chu-Carroll is careful to note that “there’s a catch” to this neat division between correlation and causation, as “correlation does strongly suggest a causal relationship.” In the case of such a “strongly” suggested “causal relationship,” Chu-Carroll emphasizes that further examination of the data are needed, and that one should keep one’s eyes open for some additional third factor (such as age in the immunization-autism correlation). He further points out that, in the face of proponents of one theory or another “waving around statistics”:

To show causation, you need to show a mechanism for the cause, and demonstrate that mechanism experimentally. So when someone shows you a correlation, what you should do is look for a plausible causal mechanism, and see if there’s any experimental data to support it. Without a demonstrable causal mechanism, you can’t be sure that there’s a causal relationship - it’s just a correlation.

I find Chu-Carroll’s explanations of these concepts of correlation and causation quite illuminating in regard to one of the thorniest issues in Autismland, the vaccine-autism issue and its corollaries, the mercury/thimerasol-autism issue and (now that epidemiological studies show that the numbers of autism diagnoses continues to increase even years after thimerasol was removed from childhood vaccines), the mercury/environmental toxins/autism issue. That not everyone in Autismland is considering so carefully what is correlation and what looks like causation—-that there is, rather, a tendency to see one thing (child exhibiting autistic “symptoms”) and recall another thing (dates of immunizations) and add these up to equal “vaccines cause autism”—has been egregiously apparent in much online writing about autism in the month of January. Read David Kirby on there is no autism epidemic or two recently posted Huffington Post pieces by autism mothers (and numermous other stories) which all but state that mercury poisoning causes autism, and that call on us to stop wasting time presenting views of autism that “soften” what it is and to go on an all-out, and well-publicized search—whatever the correlations—for the cause. By omitting the crucial step of looking for a third factor, proponents of these theories make connections that seem obvious, but that are not completely substantiated, nor true.

Indeed, I note a certain correlation, somewhat measurable, among the above-cited posts and their use of terms such as epidemic.

The Kirby post referred to was the one that mentioned children suffering from “rivers of diarrhea,” et alia. A few days later, autism mother Kim Stagliano wrote a post about all things kaka in life with autism. While her Stupid Autism Tricks post did not make explicit reference to this topic, more than a few commenters agreed with her inveighing against some recent “softening” views of autism by mentioning numerous bodily substances that ended up in various places around the commenters’ houses from which they would have preferred not to have had to scrape them off. Just a few hours ago, yet another autism mother-authored post appeared that detailed the messy and the kaka as “what autism really is,” vs. the (literally) sanitized version on The View’s recent autism show and various other studies, books, etc..

David Kirby is not, of course, an autism parent, but still it seems to have become part of the genre of “real autism writing” to include concrete descriptions of kaka (not a euphemism) into one’s writing about autism, especially when one is writing about a very “real” phenomenon related to autism, the “epidemic of autism.

Is there a correlation here between increased mention of kaka and mention of an “epidemic” of autism? Further: Might this be one of those cases in which, as Chu-Carroll notes, correlation is causation? In which constant, perseverating, focus on the “bad things” (this is the literal translation of the classical Greek word kaka) of life with autism results—causes—some to think that there is an “epidemic” of autism, as if to underscore or prove just how bad things are?

And if you don’t think that all of that computes, consider: What if we focused, each time we wrote about autism and especially our autistic children, just on what they do, however small and mundane: Standing on the porch to look for us when we come home from work, gnawing on chicken bones, sitting on a New Jersey Transit train, running hyper circles through the house late at night while warbling sounds at the top of his lungs. What if we tried to take even some of those “bad things” and just see them not as bad, not even as good, but just what is—what our child is—and see how simply rewarding life can be. Because we have this life with our child, whom we love.

Perhaps if we tried this, even the notion of an “epidemic” of autism would not seem so scarey, not evoke so much alarm, so that—even as research continues into finding the causes of autism—we can alike continue to live with both the kaka and the kalagathia, the beautiful and the good, of autism.

*[12.40pm, 1 Feb 07: A study published in 2006 found that the average age of diagnosis in children in metropolitan Atlanta was 5 years 1 month overall and 4 1/2 years for children who were more severe (see this CDC press release from May 2006); information from New Jersey's Division of Family Health Services: Special Child Health and Early Intervention Services notes that "national average age of diagnosis is between the ages of 3 and 4, even though a reliable diagnosis can be made by age 2." Thanks to Club 166 for this information.]