Quantitative vs. Qualitative Science

[IchorFigure]

I vaguely recall somewhere in Rand's non-fiction writing that she reffered to one of the "soft sciences" as a qualitative science. It may have been in ITOE. Speaking about either philosophy or economics as a science that is qualitative, as opposed to quantitative, like physics. This is something I've generally agreed with but lately I've been thinking about it more and I don't think my understanding of such differences is as solid as I thought.

Is this a useful distinction of *categories* of science? Not just as a designation of a study, but as the category of a science as a whole. And if so what do you think are the essential differences that seperate the two types?

[TheEgoist]

I think Rothbard's analysis is interesting and fits in with Objectivism, even given Rand's distaste for praxeology (which I think was a mistake on her part)

http://www.rothbard.it/essays/praxeology-as-the-method.pdf

[IchorFigure]

Hmm thanks. I've heard of his Praxeology with mixed opinions but not read any of its content.

But I think my idea of what would make a science "qualitative" or "quantitative" was probably not defined well and even floating. For example would physics be an example of a quantitative sciene, where psychology is qualitative?

[Grames]

Is this a useful distinction of *categories* of science? Not just as a designation of a study, but as the category of a science as a whole. And if so what do you think are the essential differences that seperate the two types?

Rand wrote about the qualitative vs. quantitative issue in ITOE chapter 4 "Concepts of Consciousness", where she took up the question among others "can you measure love?"

In regard to the concepts pertaining to evaluation ("value," "emotion," "feeling," "desire," etc.), the hierarchy involved is of a different kind and requires an entirely different type of measurement. It is a type applicable only to the psychological process of evaluation, and may be designated as "teleological measurement."

Measurement is the identification of a relationship—a quantitative relationship established by means of a standard that serves as a unit. Teleological measurement deals, not with cardinal, but with ordinal numbers—and the standard serves to establish a graded relationship of means to end.

The kind of measurement involved is the essential difference.

[Reidy]

Might this be the distinction between qualitative and quantitative comparisons in ITOE?

[Dante]

Is this a useful distinction of *categories* of science? Not just as a designation of a study, but as the category of a science as a whole. And if so what do you think are the essential differences that seperate the two types?

It is a mistake to attempt to classify an entire scientific field as either quantitative or qualitative. Some questions within any given field can be addressed quantitatively, and others are by nature not quantifiable. Physics might represent an extreme case where almost every relevant question is quantifiable, but in economics or psychology, for example, there are questions which are quantitative in nature and questions which aren't.

I think Rothbard's analysis is interesting and fits in with Objectivism, even given Rand's distaste for praxeology (which I think was a mistake on her part)

http://www.rothbard....-the-method.pdf

So, I've been having some of these thoughts churning for a while throughout going to PhD school for economics while also having respect for Austrian economics, and I'm gonna choose this point to just put some of them down. This isn't picking on you or this particular link so much, just stuff I've been chewing on for a while. Here goes.

Rothbard is mistaken to dismiss wholesale the methodology that he outlines in four steps as "positivism" in that piece. There are plenty of stylized facts which we observe with regularity but which are not well understood, and proposing and testing explanations of these facts is a worthwhile enterprise.

His rejection is based on his argument that Step 3, the testing step, is impossible, with the example of price inflation. Here he is simply mistaken; the entire goal of statistics and econometrics is to isolate and test variables in an environment where direct experimentation is not possible. Of course, it is true that every econometric methodology and equation has assumptions imbedded in it, and the strength of such an analysis ultimately rests on heuristic arguments about the plausibility of those assumptions (usually, that the dependent variable is uncorrelated with the error terms in an econometric equation). There have been some incredibly embarrassing empirical papers published relying on assumptions that can be shown to be wildly inappropriate. This should caution the economist when interpreting the results of such methods, but it is not an argument for the wholesale rejection of this kind of testing. Sometimes, a very good case can be made that the assumptions are accurate, and the econometric analysis that results is therefore definitive.

Incidentally, his reference to the "well-known social science analogue of the Heisenberg uncertainty principle" is indeed well-known in economic circles as the Lucas critique, and no economics paper that fails to confront it will be taken seriously in modern academia.

I share Rothbard's regard for the methodology of praxeology; making deductions from true facts is a methodologically sound way of forming general principles about economics. However, if this were all economics was, it would be extremely limited in what it could tell us about the real world. Often such a process of deduction leads to two theoretical effects which work in opposite directions, and without turning to the data at that point, one is left unable to make any final conclusions. Oftentimes, the data allows us to conclude in such a situation that one effect is typically much more significant than the other, leaving us ahead of where we were before data analysis.

It always strikes me as odd that Austrian economists such as Mises and Rothbard reject any sort of attempt to quantify relationships between economic variables, given their regard for the ability of entrepreneurs to navigate the marketplace. Rothbard, for instance, says in that article:

Among the many factors helping to determine the demand and the supply of butter, for example, are the valuations placed by each consumer on butter relative to all other products available, the availability of substitutes, the climate in the butter-producing areas, technological methods of producing buffer (and margarine), the price of cattle feed, the supply of money in the country, the existence of prosperity or recession in the economy, and the public’s expectations of the trend of general prices. Every one of these factors is subject to continuing and unpredictable change. Even if one mammoth equation could be discovered to “explain” all recorded prices of butter for the past 50 years, there is no guarantee, and not even the likelihood, that the equation would have anything to do with next month’s price.

Of course, he is correct that everything on which such a numerical relationship depends could change by tomorrow. And yet, the methodology he rejects is exactly what entrepreneurs do when they attempt to predict market demand in order to decide how much to produce, what price to set, etc. What does he think butter manufacturers rely on when they decide how much to produce? Rothbard and Mises consistently defend the ability of entrepreneurs to develop methods of meeting market demand as well as is possible; how does he think they do it? They do it by relying on the relative stability of economic relationships that have been shown to be pretty stable in the past. Of course, these relationships are based on consumer valuations at the core, and they could change drastically tomorrow. Sometimes they do, and entrepreneurs go bankrupt, and economists relying on these relationships say things that are way off. However, most of the time these relationships do hold up.

In summary, good economists do recognize that when they find an empirical regularity and explain it, they haven't done the same thing that a physicist has done when he models a natural law. Empirical regularities in economics depend on individual behavior, and that can and does change constantly. However, given this fact, it is still possible to glean something useful from quantitative relationships. If this were not possible, entrepreneurs would be just as lost as economists in producing for anticipated future demands, and the whole free market system, which relies on entrepreneurs to make good predictions (as Mises, Rothbard, and Kirzner constantly remind us) would be dysfunctional.

Edited March 29, 2012 by Dante

[Dormin111]

Interesting examination. I think it is worth mentioning that empirical economics tends to breed a quasi-psychic hubris in its most devoted practitioners that leads them to believe they can use the government to do good for the world because their models are "scientifically proven." In you opinion, is there any valid macroeconomic use for empircal economics or does its nature make it only useful for the microeconomic scale?

[softwareNerd]

... ... empirical economics ... ...

Is the term "empirical economics" used to mean controlled social experiments? For instance, where some student-subjects show they value gaining a dollar more than they dis-value losing a dollar... things like that? Or does it mean something else?

[Dormin111]

By "empricial economics" I mean what Rothbard called "positivism" or the application of hard science methodology to economics.

[Dante]

In you opinion, is there any valid macroeconomic use for empircal economics or does its nature make it only useful for the microeconomic scale?

Certainly, there are a number of valid macroeconomic uses for empirical economics. Money is inherently a macroeconomic phenomenon, and there are a number of empirical regularities concerning the behavior and effects of various monetary phenomena. For example, monetary shocks take about 5 quarters to affect output, and 6 quarters to affect inflation. These are purely economic regularities, and yet they are important to any investor or economist wishing to forecast the likely effects of monetary policy.