Lies, Damned Lies, and Statistics in Abortion Rhetoric

Based on an invitation from Kari Lundgren of Religious Rhetorics, this paper is as presented on May 29, 2010, at the 14th Rhetoric Society of America Biennial Conference in Minneapolis, MN, entitled Rhetoric: Concord and Controversy. I was asked to speak for about 15 minutes as part of a panel on rhetoric in the abortion debate. My citations in this paper are in the form of web links.

From a statistical viewpoint, we could talk about many things, such as the choices made by the APA. To keep to time, I'm going to focus on just two areas. The second will be how we perceive large quantities, or as attributed to Stalin, "A single death is a tragedy; a million deaths is a statistic." But first, I want to talk about our level of trust in individual numbers, and how much we should believe them. This idea comes from the title, often attributed to Mark Twain: "There are three types of lies: lies, damned lies, and statistics."

Interestingly, both those quotes are misattributed. Twain was not the first with that quote, and Russian historians have no record of Stalin speaking the million quote. Like finding the original quote author, the abortion debate is complicated. It has plenty of mis-attributions, the lies, and selective information, the damned lies and statistics. Let's look at some recent examples, beginning with a personal anecdote.

Four weeks ago (May 2, 2010), after submitting my grades. I went to California and San Juan Capistrano. After a day of wandering around the historical mission, fruitlessly searching for swallows, I attended Mass at that Mission Basilica. At the end of Mass, a gentleman approached the lectern to speak about the Knights of Columbus Ultrasound Initiative. The speaker said that the percentage of mothers who do not choose abortion went from 20% to 62%. The audience was impressed, and afterwards, plenty of people dropped money into his collection basket.

Intrigued, I searched for more information on that claim. The Knights had additional information on their website, including this quote: Reports indicate that up to 90% of women considering an abortion choose to have their baby after seeing an ultrasound image. They hear their baby's heartbeat, they see their baby's head and fingers. They know it is a child, not a "choice."
Though I heard about this number recently, it's been around for a while. In December 2000, Michelle Malkin wrote an opinion piece on the idea. As part of it, she cites a seemingly powerful statistic: "Dorothy Wallis of the Care Pregnancy Clinic in Baton Rouge, La., reports that 98 percent of women who have ultrasounds choose to carry to term."

Taken together, these sound strong. Unfortunately, neither hold up to strict scientific review. Let's take the Knights first. Neither the Mass speaker nor the Knights' webpage included a proper citation. I couldn't find such powerful general results. This week's New York Times cited a peer-reviewed article on the subject, which I could find online, from 2009. In this study, two abortion clinics in British Columbia found that 73 percent of patients wanted to see an image if offered the chance. Eighty-four percent of the 254 women who viewed sonograms said it did not make the experience more difficult. None of the women changed her mind about having the abortion after having seen the ultrasound. Zero percent is a lot less than 62, or 90. There's a contradiction here. Evidence favors the Canadian authors, with their published peer-reviewed results. At a minimum, the Knights should have given online citations for their work, not just "reports indicate." That implies scientific support, which hasn't been proved, an error.

But what about Ms. Malkin? Her quote is likely correct. Dorothy Wallis does lead a pregnancy clinic in Louisiana, providing support for her 98% figure. However, she forgot to mention one thing. Women near Baton Rouge know what's up with the Care Pregnancy Clinic. As their website states: "The Care Pregnancy Clinic is a refuge of hope and healing for abortion vulnerable mothers."
We have selection bias. Most women coming to the Care Pregnancy Clinic are very likely predisposed to carrying the pregnancy to term. There's no evidence of additional benefit from the ultrasound. Thus, while the 98% figure is technically correct, it means something different from what Ms. Malkin claims.

By the way, what's the ultrasound reality? Anecdotal evidence exists from reviewed, published interviews (in 1980) that seeing the potential life in an ultrasound causes some pregnant women to reconsider and change the words describing their fetus. Also, in 2002, Crisis Magazine published records from a Boston clinic that introduced ultrasound screenings and then had a reduction in the pregnancy termination rate. Taking into account missing data, a defendable estimate is that 10% of women who see a sonogram shift their views. That's non-trivial. It's just nowhere near what I heard.

Let's look at another example that misleads, from the other side of the debate. Last year, Rachel Maddow reported on MSNBC that 87% of US Counties had no abortion provider. Her quote was an implication, that "Why bother making it illegal if you can just make it impossible to get?" This 87% number has become a rallying cry. Google returns plenty of hits, including the New York Times, Planned Parenthood New Jersey, and NARAL. It also returns the original source, the Guttmacher Institute. The Guttmacher Institute is generally considered an unbiased counter on the subject. This number is true.

We hear the big number, 87 percent of counties, and conclude that a large majority of women must have huge travel burdens. But is that true? Well, no. Each county does not have the same amount of people. The most populous county in the US is Los Angeles County, California, with about 9,800,000 people. The least populous county is Loving County, Texas, with 67. Does Loving County need an abortion provider?

Because of the unequal distribution of population, only about 35% of women actually live in a county without an abortion provider. The big, scary number just became a lot smaller. Furthermore, county lines are arbitrary, because counties also vary in size. My home state of Kentucky has 120 counties, despite being below average in area. Colorado, a much larger state, has 64. California has 58. The county is the wrong unit. A better approach might be to look for "significant travel", say more than one hour. The Guttmacher Institute estimated that about 25% of women had to travel at least 50 miles. While 25% is a non-trivial percentage, at the same time, it's much smaller from the 87% figure continually repeated. Like Ms. Malkin, Ms. Maddow was misleadingly selective.

Well, now that I've made you all extra critical (as if rhetoricians weren't there already!) let's think about abortion count. America is a big country, with many people and abortions. According to the generally accepted estimate from Guttmacher, there were about 1,200,000 abortions performed in the US in 2005. Since Roe v Wade was ruled in 1973, there have been more than 50 million induced pregnancy terminations. The anti-abortion movement often refers to these numbers. Take, for instance, a speech given by Kansas City Catholic Bishop Robert Finn to the 2009 Gospel of Life convention: "The constant magnitude of this crime against humanity is staggering. We must never get used to it. In the United States there are 4000 abortions every day. Compare that to the tragedy of September 11, or to any other war, or even to the genocidal Holocaust of 6 million Jews and many others under the Nazi regime. The count of abortions over the 36 years, since its legalization in January, 1973, is beyond 50 million human lives. These are just the reported abortions. There are more. There are many, many more worldwide. But keep reflecting on 4,000 killings a day of innocent babies."

In this address, Bishop Finn focuses on the daily count. Why so? Because the person who popularized "Stalin's" quote was right. Humans don't understand large numbers; we can't grasp them well. We need figures based on values within our daily life. For instance, look at how we talk about the oil spill in the Gulf of Mexico. Do we really understand 5,000 barrels per day, the original estimate, or the current estimate of 500,000 to a million gallons per day? Generally, no. That's why the spill is often described in terms of the Exxon Valdez, something in our collective memories. Or it's described in terms of filling gymnasiums. The spill is described relative to a state, like "bigger than Maryland and Delaware combined." (Note from August 2010: These estimates would be different now, of course.)

Similarly, Bishop Finn has to make comparisons. His speech tries to compare the abortion count to other things we find repulsive, the "tragedy of September 11" which had about 3000 deaths, and "the genocidal Holocaust of 6 million Jews." It's difficult, but he has to make that endeavor. Despite the attempts, this approach is not very effective. Though it frustrates quantitative people like me, humans take better to stories. It's the reason I began this talk in San Juan Capistrano, for instance. I find less reference to numbers from supporters of abortion; they tend to focus on individual stories and symbols like coat hangers.

As for this study, where does it conclude? Things look bleak, with lots of controversy. I gave three examples in this paper, from both abortion supporters and opponents. I didn't have to work very hard to find them. At the same time, the abortion debate is particularly immoral in its citations; every field has plenty of lies and statistics. This paper provided short-term lessons through the examples. The more important long-term lesson is to assist people with tools to analyze statistical rhetoric. That goal, statistical literacy, is one where both sides can be in concord.