Confirmatory research, not physics envy

As part of the roundtable on Clausewitz’s On War, Joseph Fouche makes this point regarding Nicholas Taleb’s The Black Swan:

One of Taleb’s main themes is the tendancy for specialists in any field to develop physics envy and attempt to reduce the horrifically complex phenomena that they are studying to a deterministic and mechanistic theory complete with complex equations. This doesn’t lead to a higher level of truth and accuracy. It leads to a higher level of systematic self-deception and delusion. It creates financial weapons of mass destruction such as an MBA armed with a spreadsheet and the belief that manipulating rows and columns bestows the ability to prophesy. Vain is the life of man.

Joseph’s mention of ‘physics envy’ is a common put-down against quantitative research by those who enjoy qualitative research. But there is another, better, way of framing the debate: exploratory research and confirmatory research.

Confirmatory research includes the tools that most researchers think of as “quatitative,” such as analysis of variance (ANOVA), multiple regression, chi-square tests, and so on. Confirmatory research is also what we might call Popperian science, after the famous philosopher of science Karl Popper. In confirmatory research, we begin by having a research question, we translate it into null and alternate hypothesis, we device our experimentals, and attempt to reject or fail to reject our null hypothesis, as the case may be. Confirmatory research can be thought of as the process of trying to infer the parameters of a population from the statistics of a sample.

Exploratory research, on the other hand, is an attempt to understand the world so that a sensible research question can be asked in the first place. Some exploratory techniques are also quantitative. Exploratory Factor Analysis (EFA) and exploratory Principle Components Analysis, for instance, are exploratory statistical techniques without any null hypotheses that allow researchers to examine the various rotations of the hidden factor structure in a data set. Other exploratory techniques may be called theoretical research. In this approach, a researcher goes through large body of research to tease out hidden themes and discover gaps in existing research. Lastly, the sort of qualitative research implied by the initial questions to this problem are exploratory qualitative research. The focus is on understanding a process, rather than estimating the parameers of a population from the statistics of a sample.

While not all quantitative resarch is confirmatory, all qualitative research is exploratory. Hence the intended gist of the questions in this problem. The difference between exploratory research and confirmatory research is that the former attempts to understand a process, while the latter seeks to estimate the parameters of a population from the statistics of a sample. The first of these is always true of qualitative research. The last of these is sometimes true of qualitative research. To phrase it more simply: confirmatory reseach (that is,some of quantitative research) is driven by theory, or should be; exploratory research (that is, all qualitative research, and the rest of quantitative research) drives theory.

5 thoughts on “Confirmatory research, not physics envy”

  1. We can see the pattern now:

    1) Quants drive out the Marxists by waving numbers around.
    2) Quants begin believing their own numbers.
    3) Quants take unbelievably large risks.
    4) Everthing blows up.
    5) Quant-fed capitalist excess get blamed.
    5) Marxists come roaring back based on their card carrying non-quantness.
    6) The state withers away.
    7) Profit!!!

  2. While we’re now approaching mass bank nationalization [1], it’s striking how non-Marxist the discussion has been (with certain exceptions [2]). The debate between free market economics and social welfare states is an important one,

    The only redoubt that the Marxists have is in the universities. Their fate depends on the funding decisions of granting agencies such as the National Science Foundation, the National Institute of Health, and so on.

    I like your follow-up post [3], by the way.

    [1] http://www.calculatedriskblog.com/2009/01/nationalization.html
    [2] http://economistsview.typepad.com/economistsview/2009/01/the-union-way-up.html
    [3] http://committeeofpublicsafety.wordpress.com/2009/01/25/week-links-in-the-chain-iv

  3. Thanks,

    Since the true depths of the GFC are still to come, the Empire striking back is still in its infancy. I’ve already seen many left of center figures like Rep. Henry Waxman (D-CA) who now have real teeth claim that the entire crisis was generated by the triumph of free market ideology and its discontents. At minimum, we’re seeing a Keynesian resurgence, for good or ill. At maximum we’ll see a fossil Marxist resurgence, if only on the coasts and oversees. I think a neo-Fascist (though not explicitly called that by its adherents) resurgence, mixing features of the free market with authoritarian and populism, will be more common than classical Leftism. However, you have figures like Chavez and Evo Morales who swallowed the leftist Koolaid whole so I might be wrong.

  4. “While not all quantitative resarch is confirmatory, all qualitative research is exploratory.” (tdaxp)

    It can be very scary when qualitative research is used as explanatory. In my academic department we have weekly talks given by someone outside the department. A speaker came in to give a presentation on her latest research. The research was centered on “proving” how the “cycle of poverty” theory was wrong [1]. She said the “cycle of poverty” was based on “stereotypes.”

    So she got together with some local teachers and had them list some stereotypes (not enough reading in the home, no books, parents on drugs, violent disruptive atmosphere, lack of interest in kids). She then told these teachers to go to the homes of the kids and find examples of where these “stereotypes” were untrue.

    The teachers came back to her with examples proving these “stereotypes” to be untrue (some kids had books, some parents read stuff, not all parents on drugs). This proved the “cycle of poverty” theory was untrue, and the real reason these kids weren’t doing well in school was due to “low expectations” and “lack of opportunity.”

    During the Q&A I asked if she gathered any information on the home-lives of the kids who DID do well in school? I though for sure it would be interesting to compare the two groups of kids. She said “no” becuase she “didn’t want to make it seem as if there was a ‘right’ home-life and ‘wrong’ home-life.”

    Now I should point out that I don’t really care for the “cycle of poverty” theory myself (doesn’t consider genetic factors). But I shiver at the thought of this kind of research driving any kind of policy in this country? 🙁

    [1]http://en.wikipedia.org/wiki/Cycle_of_poverty

  5. Seerov gives an excellent example of the sort of Marxist qualitative nonsense that dominates too much of academia. Seerov does a great job demolishing it. It was because of Obama’s association with such a thinker than I was worried how research-grant decisions would be made under his administration. [1]

    J.F. makes the excellent point that the great financial crisis may lead to the rise of demagogues. The decline of oil prices naturally hurts those like Morales, Chavez, Putin, and so on. The real danger would be the rise of protectionism (under whatever ideology) in America, Europe, or China.

    [1] http://www.tdaxp.com/archive/2008/05/27/william-ayers-actually-dangerous.html

Leave a Reply

Your email address will not be published. Required fields are marked *