One of the problems with the social sciences is that just because you know what something is, doesn’t mean that you can count it. The classic case of this is pornography: we all know what pornography is, but defining it precisely enough to legislate it or to measure it is extremely difficult. Which is why a Supreme Court justice once said about pornography that, while it may be hard to define, ”I know it when I see it.”
It’s not just pornography. Justice, democracy, freedom, oppression, tyranny, war, peace, violence, ethnicity… we know on some level what all of those things are, but they are extremely difficult to define precisely enough to measure. And yet measuring them is often critical to understanding how the world works.
To get around this problem, social scientists talk about two different kinds of definitions. The theoretical definition tries to capture the intuition behind a concept. In most cases, the theoretical definition is either uncontroversial, or is debated at length only by philosophers. As important as they are, most social scientists prefer not to get bogged down in those debates–which is one reason why philosophers tend to look down their noses at social scientists. Instead, social scientists are usually interested in the operational definition. The operational definition attempts to boil down the theoretical definition into something that can be measured.
To understand why operational definitions are important–and why they are so hotly contested–read today’s NYTimes.com, specifically Thomas Edsall’s discussion of the problem of counting the poor. We all have some idea of what poverty is. So a theoretical definition isn’t too hard: poverty is the difficulty or inability in acquiring basic necessities of life because of a lack of money. I’m sure some people will have minor qualms with that definition, and we can get into the details of that sometime if you want, but the basic gist is correct.
The problem, as Edsall notes, is that if you want to actually do something about poverty, you first need to know who is poor and why. In other words, you need to measure poverty–and that requires an operational definition. Edsall, in particular, notes that there are currently three commonly used, ways to measure poverty, and that all of them imply different things about who is poor and what we should do about it, as a society. He even notes that all three measures may be inadequate at actually capturing the underlying concept of what we mean when we say that someone is “poor”.
What Edsall misses–and what, I’m afraid, most lawmakers and even policy experts miss when discussing these things–is that the operational definition needs to be crafted in the context of a larger question. For instance, one of the operational definitions of poverty takes into account the amount of money that individuals receive in aid from the government. For anyone trying to measure the actual living conditions of actual people, this is a perfectly fine thing to do. But if the point of the definition is to help inform lawmakers about how to spend public funds, this definition can lead to the wrong conclusions.
So let’s say that this “assistance included” measure of poverty says that very few children are poor, while many adults are poor. A policy maker might look at that data and reasonably conclude that we need to shift assistance away from helping children, and towards helping adults. But, of course, the reason that fewer children are poor is that government provides more assistance to poor children than to poor adults. Shifting resources like that only means that the next time the statistics are measured, there will be more poor children and fewer poor adults. If the point of shifting resources was to actually deal with poverty and lower the overall poverty rate, such a shift in funds would fail miserably. Instead, law makers who are interested in dealing with poverty need to use a poverty measure that does not take into account government assistance, so that they can understand what causes people to need government assistance in the first place.
In short, the same theoretical concept might need multiple operational definitions–and it’s up to the people who use those definitions to understand them well enough to pick the right ones. There will never be a single, widely agreed upon operational definition of poverty. But that’s okay–as long as the lawmakers and policy makers understand the strengths and weaknesses of each one.