Tricky Tricks with Statistics
Don’t Be Fooled by Conservative Hocus Pocus Calling Itself Statistics
I have theory that I call the Reflective Attribution theory. This theory stipulates that people tend to impute their own motives and attributions onto others. For instance, according to this theory one might predict that those people who are inclined to lie are also people who are do not trust others, because they assume that “because I lie, others lie.” The opposite is also the case. Honest People tend to be trusting of others.
So, I can’t help but apply this theory in my everyday life. The other day I posted the following data on my Facebook (click the image for the source).
This is an answer to the claim that President Obama is overseeing an unprecedented growth in the federal government. The underlying claim is, of course, that if Obama is allowed to continue this massive federal government increase we will soon be under the thumb of a monolithic and oppressive institution. The data above, showing the growth of government as a percentage of GDP compared to the growth of the private sector, belies this claim. Yes, when you control for state and local governments there is some growth in at the federal level. But unprecedented? If you want precedent just look at the preceding president (yes, I did that on purpose).
Which brings me to my theory. A friend of mine responded with a gotcha. An article that shows, beyond a conservative doubt, that President Obama has, in fact, grown the government. Over the years I’ve noticed that conservatives are very leery of statistics. For someone like me, who lives on statistics, this strikes me as odd. One person, years ago, told me that she does not trust statistics because “statistics lie.” Actually, statistics don’t lie. People lie. Sometimes they use statistics to do it. This is something that I see from both sides of the political spectrum, but if I am to be honest, I notice that the right side is the overwhelming leader in the quantity of “damn lies and statistics.”
The data presented by the article can be summed up in the following graph.
Yep. That’s a bigger government all right. It’s simple. It’s succinct. A nice straight line. I think conservatives like straight lines. However, it is not descriptive of the Obama Administration. In short, it’s a deceptive use of statistics. It’s true, mind you. This stat is not, technically, a lie. But the intent behind the stat is certainly to deceive.
The article cited the Bureau of Labor Statistics as its source. So went to the BLS website, looked up the data, and put together the same chart. Only this time I filled in the gaps between January of 2009 and April of 2012. I got the following (program glitch. I used January of 2009 and then the succeeding Aprils of Obama’s administration. The computer added the intermediary months on its own.):
As you can see, the overall data is the same, but filling in a few missing dates paints a very different picture. Yes, the Obama Administration has overseen a net increase in the size of government as measured by the number of employees. The bulk of that increase, however, peaked two years ago. Since then, President Obama has overseen a reduction in government employment. Embarrassing.
This is a common technique in claims makingshowing only the statistics that proves your point. A pundit can be technically honest about this and still lie. Another technique is to offer a stat by itself with no comparative data against which to measure it. Notice that the New York Times data above offers a comparison to other presidents. That’s called responsible reporting. Could the Times have been more comprehensive? Every work of research can be more comprehensive. Let’s assume that the New York Times was under practical constraints to do so.
To remedy this with regard to federal employment, I decided to offer a comparison between President Obama and President W. Bush. Now I recognize that my more extreme conservatives consider President Bush to be a Manchurian Progressive (in retrospect, of course. During his presidency he was practically the second coming). Regardless, I would wager that, given a choice between Presidents Bush and Obama, these same people would vote for Bush. Please correct me if I’m wrong. I have no data to support this claim. Regardless, the claim is that Obama’s expansion of government is “unprecedented” and employment is the variable offered to prove this.
Alas, it appears that comparative data also refutes the unprecedented growth claim.
Then, just for fun, I thought, ‘let’s compare Obama to the Conservative God’s right hand.’ Yep. That’s right. Ronald Freakin’ Reagan. How does President Obama’s expansion of Government compare to President Reagan’s during the same time period. This was a little trickier. The best stats I could find that was relatively easy to access was a measure of end of year civilian federal employment for President Reagan, so that is what I used for President Obama. I used December of the year preceding each administration as an estimate of where each president started. To wit, I compiled the following graph:
Well what do you know? Three years into his administration, the Gipper had the bigger government. True, he did oversee a decline in his first two years, but more than made up for it in his third. And this growth in government continued for Reagan right up to his farewell address.
Truth be told, this data can be looked at in a different way. One could suggest that if you look at the net rate of growth between President Reagan’s and President Obama in their first three years, Obama’s was larger. Well, yeah, but bigger is bigger in this issue and Reagan had the bigger government. Again, the growth of Obama’s government is far from unprecedented.
Of course, looking at employment and change as a percentage of GDP are only two of many ways to analyze the size of government. It so happens that this morning I was reading the Krugman blog and noticed this graph, which speaks to my topic. So I stole it.
So back to my original thesis about the Reflective Attribution. It’s no wonder that conservatives are so leery about statistics. Either they don’t understand statistics, and therefor assume that nobody understands them, or they assume that all claims makers are equally disingenuous with their use of data. This is an unfortunate state of affairs, for understanding data, regardless of the nature of that data, is crucial to developing an objective and rational understanding of the issues. As a sociologist, I know that very often what we believe to be true is, in fact, false. If we base our decisions on nothing more substantial than what we believe to be true
the road we are paving can only have one destination.