So, yesterday I talked about Numbers and how they figure into my personal definition of myself. It was a discussion of how numbers create some of the meaning in my life.
Today, I want to give an example of how people use numbers in a completely different manner. In specific, I want to show how people use numbers as a means of creating arguments/supporting arguments and how they try to draw upon the “power” that numbers often possess to many audiences.
This power, not surprisingly, derives from the association that numbers have with quantification and the idea that quantification is a kind of “objective” process, and therefore, true. While this association between quantification, objectivity, and truth may seem obvious in a world filled with high-tech electronics, information technology, and data-driven search engines at our beck and call–one should not take such claims for granted, and should instead understand that this current state of belief in numbers is just as much of a construct of history as the technologies that we associate with it. A good treatment of such ideas can be found in Theodore Porter’s book, Trust in Numbers: The Pursuit of Objectivity in Science and Public Life. I came across this book during my prelims (which were focused on, the History of the Enlightenment, The History of the Applied Social Sciences, and The History of Technology), and it has some cool stuff in it.
It is academic prose though–so be forewarned..
Now.. back to the story. Rather than talk in the abstract about the power of numbers, what I really want to focus on is how this kind of power was applied to tell a certain kind of story and to push readers to a certain conclusion.
This story, my dear readers, began a couple weeks after the 2012 American Presidential Election, and it came about as I was reading my favorite political blog–hell, my favorite blog overall–namely Andrew Sullivan’s The Daily Dish.
In my daily perusal on Nov. 26th, I came across the following article–titled: The Politics of Fertility.
Now–in this blog post, Andrew & crew linked to another online article–by Lauren Sandler–that made the argument that the state-by-state election results could best be predicted by the fertility rate in a state–and that the correlation argued that larger family size/fertility rates–implying a kind of support for “family values“–meant more support for Republicans, while lower fertility rates meant more support for Democrats.
In fact, the title of Lauren Sandler’s original blog post was, “Tell Me a State’s Fertility Rate, and I’ll Tell you how it voted” and in that article, the basic claim is demonstrated in a graphic and a table that claim to show that states with fertility rates over 70 voted Republican, whereas states with fertility rates under 60 voted for Obama. One of the graphics is shown above. She followed up this graphic with a table that came from Governmental CDC data.
Now–before we go any further, I think that it’s worth doing a bit of thinking about the numbers that are assembled in these graphics. Remember, the claim is that these numbers make it easy to understand the reality around us–and that they simplify this reality in a way because they supposedly show us an underlying pattern that was obscured by all the messy details of reality.
Does this claim hold up to reality?
If you stop and think about the meaning of these numbers–if you try to understand the numbers and analyze where they come from, then the argument starts to break down pretty quickly. Some points to think about:
1. What about Hawaii? If you go look at the initial Map, one sees that Hawaii had a fertility rate over 70, and yet it voted for Obama. So the “over 70 rule” is not hard by any means.
2. If you look more specifically at the states that voted for Obama, the correlation between low fertility rates and support for Obama is all across the board. For example, New Hampshire has a really low fertility rate, and yet it voted for Obama by only a 5.8% margin. New York–with a much higher fertility rate–voted for Obama by 26%. Additionally, while Vermont and New Hampshire’s fertility rates were nearly equal, Vermont voted for Obama by nearly 36%–or over 6x the difference of New Hampshire. Similarly, you can see that New York and Florida had nearly the same fertility rate, and yet Florida barely went for Obama 50.0% to 49.1%, whereas in New York, it was 62.6% to 36%.
3. If you actually then look at the tables, you can easily notice two problematic issues in the column marked “total fertility rate over 70”: first, that they include two Republican states whose fertility rates are under 70 (Texas and Wyoming)–even if just barely so–but they do not include Hawaii in there, even though it’s fertility is over 70–because that would disrupt part of the heading of the column–namely that “all romney states.”
Thinking further on fertility rates, one might remember having heard somewhere that minority groups tend to have higher fertility rates than whites and that this is the reason that the white majority in this country is diminishing. With this in mind, one might then also note that this election clearly showed that minorities voted OVERWHELMINGLY for Obama.
So–you have an article stating that high fertility rates mean votes for Republicans–and yet those groups that have the highest fertility rates vote strongly for Obama?
Something really odd is going on here.
Specifically, what it points to is that the notion that a state’s fertility rate is really a bogus measure to use–as the causation between having a high birth/fertility rate in a subgroup does not correspond to greater support for Republican ideals at all. Actually, a close reading of the data (see pages 7 especially) leads to exactly the opposite conclusion–namely that the majority of groups with a high birth rate–at least ethnically–were much bigger supporters of Obama than of Romney.
So–the numbers do have power–but the problem here was that someone was trying to use the power of numbers to make a particular argument and claim that wasn’t actually supported by the numbers. Instead, they were misusing the numbers because they didn’t spend the time to grok the context or real meaning of them.
Now this might be a fine place to end the story, but in reality, there was one very big twist left in this tale. As I noted at the outset, originally I came across this article by reading Andrew Sullivan’s blog, where–without much comment–he represents the claims made by the Sandler blog post. Importantly, he and his crew included the map graphic that she had used–and I’ll include a screenshot of it here:
Now look very carefully at the map on this post. Compare it to this map:
Do you notice a difference?
Something has changed.
The President’s birth state has turned Republican Red.
How did that happen?
Overall, I’m still puzzling over the exact process that occurred between the blog post by Sandler–that has all the errors that I’ve noted, but which doesn’t mis-color Hawaii–and the final Daily Dish presentation where it appears that someone physically altered the image to correspond with the original blogpost’s argument, even though that meant lying about reality.
Numbers have power because they can be used in helpful ways to teach us about the external world.
But that accuracy is not a given.
People can lie with numbers–and they do.
In the end, the best strategy is to always make sure you understand the meaning of the numbers–to push further so that you can make the numbers make sense–otherwise, you may just be allowing someone to lie to you in yet another way…