Sciences and nons(ci)ences

By Emanuel Derman
August 5, 2011

I have spent too much time on Twitter*.

Nevertheless, one of the things I did notice is how many of the economic tweets are arguments about statistics. One in particular, pointed at by @FelixSalmon and @JustinWolfers, is an article that stresses:

… statistics is a rhetorical practice.  The goal is not just to convey information but rather to change minds.

The author of that extract, Jeff Ely, is a professor of Economics at Northwestern. I wonder whether professors of statistics would agree with his definition. In any event, it made me think about whether statistics is a science, or, indeed, what is a science**?

In my categorical scheme, science is about finding the rules or relations the world satisfies. I suppose I am assuming the existence of an objective world.

Therefore:

  • I wouldn’t call mathematics a science. I’d call it a tool or a method. Gauss called it “the Queen of the Sciences”, a reference to the role of the queen in chess, I think.
  • I’d call physics, chemistry and biology, etc., the natural sciences, in accord with tradition.
  • I’d call economics and politics the moral disciplines because there will hardly ever/never be an objectively true answer to a question.
  • And I’d call statistics the heuristics of extrapolating data. I don’t mean that pejoratively.

In that case, when an individual ignores pure statistical evidence that comes from a collection of data, that is extrapolation heuristics, I don’t think you can fairly say they’re ignoring “science”.

I foresee someone pointing out that all real science is based on the statistical analysis of data too. That’s true. But it’s only science that’s being ignored if there’s a theory or model or explanation underlying it too, some deductive scheme based on structural principles.

________________

* I wonder if anyone has calculated the amount of GDP lost to the nation by people, not in the news business and still holding jobs, who obsessively look at tweets, their own, responses to their own, and others.

** A colleague of mine once told me that his father said that any field that attempts to make the word science part of its name (nutrition science, social science, domestic science) cannot be a science. When I was in graduate school, and my friends and I went off to work in the evening or on the weekend,we said we’re going to “do some physics”. The biologists I knew, under similar circumstances, said they were going to “do some science”. And they were, but I thought it was a strange and glamorous usage.

3 comments

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

I would say that if you ignore overwhelming statistical or evidence in favor of a scientific theory, and are unwilling to concede that you must posit an alternative theory fitting that evidence (and not contradicted by other evidence), then it is fair to say you’re ignoring the “scientific method,” and that it is what people may often mean when they say someone is ignoring “science.”

A good example are the few scientists who claim to be Christian fundamentalists. They may acknowledge the reality of micro-evolution but reject that macro-evolution resulted in the origin of species. They are rejecting scientific evidence without an alternative theory that holds up to the evidence.

(By the way, acc. to your second footnote, computer science is not a science. That is true. It’s really mathematics.)

Posted by MICRM | Report as abusive

My original post was motivated by the endless discussion on Twitter about whether we were likely to be heading into a double dip. All of the argument was statistical, without a reliable theory of what would cause a double dip. It seemed to me that calling people unscientific, when they disagreed with predictions based on a frequentist approach to the small number of previous double dips, was pointless. if people disagree with a more or less proven theory, that’s unscientific. But if they disagree with a theoryless regression, that’s a different story.

I agree with you about computer science: there’s only computer engineering and mathematics.
Probability (as opposed to statistics) I would count as mathematics.

According to classic philosophy and “classic” philosophers science is one of the methods of “cognition” (research, study). Another methods are philosophy itself, divine revelation, mystic experience. Before age of modern (in renaissance, medieval, ancient times) these methods were considered as equal (moreover, sciencific method was considered defective, because it was based on empirical studies). But now we have opposite situation – science is considered as the only possible way of understanding the world. To say “unscientific” means to give the most negative assessment. For me unscientific means exactly what it means – out of area of science. It does not mean something negative, it only means that we should use other methods of research.

Science begins when we have ideal object (or model) of some phenomenon (ideal gas is not a real gas). This object must be absolutely unchangeable. Scientific object was the same 1000 years ago and will be the same next 10000 years. Scientific knowledge is a knowledge about unchangeable features of the ideal object. If you change the object then begin to study it from the very beginning. Science tries to find Laws of Nature, or the principles of functioning of the World.

According to this classic definition, economics, sociology, political science and all other “social” sciences are not sciences at all. They have no unchangeable object. Empirical data cannot be universally interpreted.

In my opinion mathematics is a language of relationships, not a science. Chemistry is more an engineering then pure science.

Statistics is a method of searching for regularities, not laws. Regularity cannot give you a picture of future behavior. Law can.

Sorry for long comment.

Posted by complexwriter | Report as abusive