Do any benefits of alcohol outweigh the risks?

19 May, 2021 at 23:02 | Posted in Statistics & Econometrics | 3 Comments

.

Identifying the data generating process sure is important if we want to be able to understand data. Finding a correlation between drinking alcohol and mortality does not explain anything, and certainly does not mean that you have been able to identify a causal relation between variables. Regressing on covariates is not enough. There are tons of alternative explanations for the (alleged) causal relationship, and as long as you haven’t been able to convincingly block them all, you haven’t really succeeded with your identification strategy.

3 Comments

  1. Our ancestors survived and thrived over many thousands of years because they learned about numerous vital causal relationships. For example, they learned from accidental experiences and repeated trial and error experiments that various types of berries are poisonous (eg yew, ivy, holly) causing stomach cramps and even death, that eating psilocybin (aka magic mushrooms) causes psychedelic experiences, and that various forms of cyanide are extremely toxic causing coma and rapid death.

    Our ancestors made these discoveries without without any modern understanding of the scientific mechanisms involved. They did not have the faintest idea about what Prof. Syll calls a “data generating process” might be, unless this phrase refers to myths and magic.
    The validity of inductive inference is amply demonstrated by the continued existence of homo sapiens up to the present day..

    These historical facts utterly demolish Prof. Syll’s claim that “finding a correlation between drinking alcohol and mortality…does not mean that you have been able to identify a causal relation between variables.”
    This claim is only true in the trivial sense that simple correlations cannot expose the detailed mechanisms of causal relationships or the complex inter-relationships between several simultaneous causal relationships.
    Even today our understanding science and the “data generating process” is very limited so we are very far from “really succeeding”.

    Nevertheless, today, as in ancient times, correlations do indicate the probability of causal relationships. Of course, as with all inductive inference, complete certainty is never possible.

    Comment by Kingsley Lewis— 20 May, 2021 #

    • Are you saying humans continue to drink alcohol because they are ignoring the correlation, or because they have discovered other causal relationships that science has not?

      Comment by rsm— 20 May, 2021 #

      • Is the risk so high that a reasonable risk manager would rule out moderate amounts of alcohol?

        Comment by Nanikore— 20 May, 2021 #


Sorry, the comment form is closed at this time.

Blog at WordPress.com.
Entries and Comments feeds.