Lies, Damned Lies, and Statistics


From Ancient Greece to the late 19th century, rhetoric played a central role in Western education in training orators, lawyers, counsellors, historians, statesmen, and poets. However the rise of empiricist and positivist thinking marginalized the role of rhetoric in 20th Century university education. Julie Reuben in “The Making of the Modern University: Intellectual Transformation and the Marginalization of Morality” writes about this change as follows:

“In the late nineteenth century intellectuals assumed that truth had spiritual, moral, and cognitive dimensions. By 1930, however, intellectuals had abandoned this broad conception of truth. They embraced, instead, a view of knowledge that drew a sharp distinction between “facts” and “values.” They associated cognitive truth with empirically verified knowledge and maintained that by this standard, moral values could not be validated as “true.” In the nomenclature of the twentieth century, only “science” constituted true knowledge.”

Once the positivist idea that knowledge consisted purely of facts and logic became dominant, persuasion became unnecessary. Anyone who knew the facts and applied logic would automatically come to the same conclusion. “Rhetoric” or persuasion was considered to be a means of deception by positivists – we could persuade people only by misrepresenting the facts or by abuse of logic. The foundations of statistics were constructed on the basis of positivist philosophy in the early twentieth century. Great emphasis was put on facts – represented by the numbers. Rhetoric (and values), represented by how the numbers are to be interpreted, was de-emphasized. This led to a tremendous rise in the importance of numbers, and their use as tools of persuasion. The rhetoric of the 20th Century was based on statistics, and data were used to present the facts, without any apparent subjectivity. As the popular saying goes, “you can’t argue with the numbers”.

By the middle of the 20th Century, logical positivism had a spectacular collapse. The idea that the objective and the subjective can be sharply separated was proven to be wrong. For a recent discussion of this, see Hilary Putnam on “The Collapse of Fact/Value Distinction”. Unfortunately, these developments in the philosophy of science have not yet reached the domains of data analysis, which continues to be based on positivist foundations. Rejecting positivism requires re-thinking the disciplines related to data analysis from the foundations. In this paper, we consider just one of the foundational concepts of statistics. The question we will explore is: What is the relationship between the numbers we use (the data) and external reality? The standard conception promoted in statistics is that numbers are FACTS. These are objective measures of external reality, which are the same for all observers. About these numbers there can be no dispute, as all people who go out and measure would come up with the same number. In particular, there is no element of subjectivity, and there are no value judgments, which are built into the numbers we use. Our main goal in this paper is to show that this is not true. Most of the numbers we use in statistical analysis are based on hidden value judgements as well as subjective decisions about relative important of different factors. It would be better to express these judgments openly, so that there could be discussion and debate. However, the positivist philosophy prohibits the use of values so current statistical methodology HIDES these subjective elements. As a result, students of statistics get the impression that statistical methods are entirely objective and data-based. We will show that this is not true, and explain how to uncover value judgments built into apparently objective forms of data analysis.

It is useful to understand statistics as a modern and deadly form of rhetoric. When values are hidden in numbers, it is hard for the audience to extract, analyze, discuss, and dispute them. This is why it has been correctly noted that “There are lies, damned lies, and statistics”. The most popular statistics text of the 20th century has the title “How to Lie with Statistics”. In this sequence of posts, we will analyze some aspects of how values are hidden inside apparently objective looking numbers.

  1. Asad, I wish you well in enlightening folk about the proper interpretation of statistics.But please be aware of the following view, which you might be advised to skirt around rather than tackle head-on, unless you really need to.

    The UK GDP and inflation rate estimates generally settle down to be as FACT-like and even TRUE to the appropriate norms as any empirical data gets. Where there is a genuine issue, which all good statisticians (and mathematicians, scientists etc) recognize and even emphasise (given half a chance) is in the INTERPRETATION of those facts. E.g., are they what we should be paying attention to?

    While many (including me) might debate such views at great length, given 1/10 of a chance, from the point of view of WEA pedagogy I think this a distraction from the more important question.

    Whether such statistics as GDP and inlfation are true or not, the key issue is surely what reliance we can put on them, and how else we might form a reasonable opinion about economies? Hopefully you are goping to enlighten us on this, at least. (Hope so!)

  2. The poisions of positivism which I absorbed during my education destroyed my ability to reason clearly, and the process of recovery was slow and painful. The problem is not in inflation statistics, but in the lack of realization of the massive element of subjectivity that goes into the manufacture of all statistics. If this was out there in the open, if we were able to acknowledge the Radical Uncertainty within which we operate, we would have a better chance of understanding the world and coping with it, and perhaps changing it for the better.

    • Econoclast said:

      Asad, thank you for sharing your painful experience. I had my own version. Nearly all my friends for decades were rationalists who constantly spouted the facts versus values stuff. Lonely place to be. One economist who had his head straight on all this was Gunnar Myrdahl.

    • Robert Locke said:

      It is, was, out there in the open, you just had been caught up in an educational system that denied it. I wrote, quoting others, about the shortcomings of econometrics and neoclassical economics when they were being introduced into the New Economic History in the 1960s. That effort failed as was acknowledged by its promoters, e.g. McCloskey, in his/her well know essay (1982) “The Rhetoric of Economics. At 88, (tomorrow) I am weary to have to go back an rehash what we have known all along.

  3. Nadia Hassan said:

    I have observed two common believes of statisticians:
    1. Data and statistics have no hidden value. It is the context that changes the interpretation of a number.
    2. The nature of data guides us about what measures we should use to analyze a fact.
    In above note, you have given a contrary argument to the first point “values are hidden in numbers”. Looking forward more elaborate discussion on these points in upcoming posts.

  4. The failures and confusions or pains relevant reflect just the failure of plutology, or the flaws of economics principles, and furtherly philosophical chaos. Criticism toward positivism, or objectivism unfortunately failed to cause its collapse, only retreat instead. Subjectivity emerged, but has not been integrated. The late 20th century witnessed philosophical silence, void and awkwardness. Now, the grand synthesis is subject to Algorithmic approach, which will lead to a clear methodology of social science, and economics, where positivism, or neutralism will still enjoy its relative usefulness as one of the tactics among various research tools. Nothing is totally objective, just as nothing is totally subjective, but tentatively distinguishing them is technically helpful, we boundedly-rational researchers have to distinguish them from time to time, and have to tolerate the leftover unclearness. Thanks!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: