Lies, Damned Lies, and Statistics

By Dr. Jeff Mirus ( bio - articles - email ) | Nov 05, 2008

I spent a very long time today working up a blog entry which purported to prove from government statistics that one adult woman out of every three has not had an abortion. I still have a deep suspicion of that figure, which has been used repeatedly by both pro-abortion and pro-life organizations. The pro-life groups use it to show how widespread the problem is; the pro-aborts just want us to accept abortion as normal.

So I took a few hours to work up the total number of women who had passed through one or more of their child-bearing years since Roe v. Wade in 1973, and I checked the numbers of abortions during the same years as reported both by the Alan Guttmacher Institute and the Centers for Disease Control. I also tried to find out how many women have had more than one abortion, a question which, based on my examination of the data, seems often to be ignored. One study reported that 47% of all women seeking abortions had had at least one abortion previously, and obviously some women have many abortions, but it is hard to find sound, extensive data on that question.

My labors led me to conclude that an accurate assessment would place the number of women who have had abortions at between 1 in 3.5 (worst case) and 1 in 10 (best case). It would be too much to say that I was elated, but I found my conclusions heartening.

But then I noticed a funny thing. If I construed the fertile years to be between 15 and 50, my methodology returned far better numbers (from the pro-life viewpoint) than if I took a more conservative span of years from 15 to 44. Thinking about this problem, I soon realized that no ultimate conclusion can be reached without something more than the number of abortions (which must be under-reported anyway) and the number of women in a generalized fertility range. The actual fertility window is critical in every respect, as one can recognize immediately if one were to assume that all women are fertile for exactly one year, and so the number of abortions each year would have to be measured against the number of women who were exactly that age—which would prove (wrongly, because of the assumption) that a significant majority of women had abortions. In short, to nail this down, a better methodology is needed.

Some researchers have tried to achieve this by either studying far more data or by extrapolating from small amounts of data drawn from better-known specific populations. The going is very difficult, and I suspect most people have adopted rules of thumb that are more or less dubious.

I may be thought a fool, but I do not wish to remove all doubt. Therefore, instead of heartening statistical news, I can offer only a larger and perennially valid Christian point: People are not statistics, and should never be treated—or even suspected—as if they were.