How Do We Know Anything?

In my tax policy seminar this week, we discussed the estate tax. I assigned readings that made various theoretical and empirical claims about the effects of the estate tax in the U.S., including the usual politically contentious questions about whether the estate tax breaks up family farms and businesses (clearly not) and whether it is "inefficient" in the standard sense of that term (not even close, in comparison to any other way to raise revenue). Even though the state of knowledge on those empirical questions is pretty clear, the readings included some assertions contesting those basic findings in the course of making moral and political arguments against the estate tax. This put some of the students at a loss. Given that they are law students and not statisticians, they did not feel competent to assess the empirical claims. As one student put it:
I’ve unconsciously set up a heuristic based on a belief adopted over time that any data put before me is either the result of faulty methodology, framed imprecisely to skew my perception, incomplete in its presentation, or fatally compromised by some other flaw or combination of flaws. Why? Because I know that I can Google around and find a study refuting everything the one I’m reading says, written by someone who is apparently qualified to weigh in on the subject. I don’t know how to figure out who is right and who is full of it, so I simply don’t know how to think about the issue beyond what my gut tells me.
As a social scientist and an educator, this statement scared me. It made me realize just how ill-equipped people are to assess the empirical claims that they hear every day. Still, I asked the students how they knew anything at all, including things that ultimately rely on facts or knowledge beyond their own experience and expertise. Do they believe that Global Warming is happening? Do they believe that smoking causes cancer? Clearly, none of the students has the knowledge or training necessary to assess conflicting claims on these topics, yet they were certain that they knew the answers. (Yes to both, for those of you keeping score at home.) And there are certainly people who deny each of these things, people who hold advanced degrees with which a layperson might be impressed. Why do the students believe each of these things yet doubt who is right about the empirical claims with respect to the estate tax, when the professional consensus on them is also quite strong?

My questions were somewhat unfair, because I was really asking my students to discuss epistemology in fifteen minutes. Our discussion, unsurprisingly, did not break new ground in philosophical discourse, but the conversation was interesting and lively. In the end, I explained that the reason I wanted them to discuss this larger question is that I view a big part of my job as trying to identify when some contentious questions really should not be contentious. We can disagree about whether, say, the tax code should be more progressive, but we should be able to have some way to determine objectively just how progressive it currently is. (I generally resist the overuse of the positive/normative distinction, but it is at least a reasonably useful way to characterize the difference between fact and opinion.) If people are not capable of seeing through specious empirical claims, what can we do to make it clear when the "two sides to every issue" framework makes it appear that there is more doubt than really exists?

The short answer, of course, is that this is not possible -- or, at least, there is no generally applicable formula that allows us to know when reasonable minds can no longer disagree. What scares me even more is that there seems to be no connection between the degree of certainty that we have about an empirical result and the degree of public confusion about it. Empirical work on the death penalty overwhelmingly and repeatedly fails to find a deterrent effect, yet so long as the occasional study comes along to keep the deterrence claim alive, the question remains "controversial" and leaves people like my students saying, "How can I tell who is right?" (As Mike's discussion on Monday suggests, the frequently repeated false claims about the U.S. tort system are another example of this type of degraded debate.)

As if that were not bad enough, it now appears that issues that were once well settled can be thrown into doubt with no apparent justification. We have recently "learned" that the New Deal caused the Great Depression, not on the basis of any new evidence or theory, but simply because some people have decided that it is now advantageous to say so. Imagine my surprise to hear that hiring people to build and repair things is not stimulative, and that "in the history of mankind and womankind, government -- federal, state or local -- has never created one job. It's destroyed a lot of them."

The great philosophical questions will be with us always. Is it impossible to do something about the most egregious nonsense?

-- Posted by Neil H. Buchanan