Home > Uncategorized > Causal mediation

Causal mediation

from Lars Syll

mediatorIn the real world, it’s my impression that almost all the mediation analyses that people actually fit in the social and medical sciences are misguided: lots of examples where the assumptions aren’t clear and where, in any case, coefficient estimates are hopelessly noisy and where confused people will over-interpret statistical significance …

So how to do it? I don’t think traditional path analysis or other multivariate methods of the throw-all-the-data-in-the-blender-and-let-God-sort-em-out variety will do the job. Instead we need some structure and some prior information.

Andrew Gelman

Most facts have many different, possible, alternative explanations, but we usually want to find — since all real explanation takes place relative to a set of alternatives — the best of all contrastive explanations.

So which is the best explanation?

Many scientists, influenced by statistical reasoning, think that the likeliest explanation is the best explanation. But the likelihood of X is not in itself a strong argument for thinking it explains Y. I would rather argue that what makes one explanation better than another are things like aiming for and finding powerful, deep, causal, features and mechanisms that we have warranted and justified reasons to believe in. Statistical — especially the variety based on a Bayesian epistemology — reasoning generally has no room for these kinds of explanatory considerations. The only thing that matters is the probabilistic relation between evidence and hypothesis. That is also one of the main reasons yours truly finds abduction — inference to the best explanation — a better description and account of what constitutes actual scientific reasoning and inferences.

In the social sciences … regression is used to discover relationships or to disentangle cause and effect. However, investigators have only vague ideas as to the relevant variables and their causal order; functional forms are chosen on the basis of convenience or familiarity; serious problems of measurement are often encountered.

Regression may offer useful ways of summarizing the data and making predictions … However, I see no cases in which regression equations, let alone the more complex methods, have succeeded as engines for discovering causal relationships.

David Freedman

Some statisticians and data scientists think that algorithmic formalisms somehow give them access to causality. That is, however, simply not true. Assuming ‘convenient’ things like faithfulness or stability is not to give proof. It’s to assume what has to be proven. Deductive-axiomatic methods used in statistics do not produce evidence for causal inferences. The real causality we are searching for is the one existing in the real world around us. If there is no warranted connection between axiomatically derived theorems and the real world, well, then we haven’t really obtained the causation we are looking for.

If contributions made by statisticians to the understanding of causation are to be taken over with advantage in any specific field of inquiry, then what is crucial is that the right relationship should exist between statistical and subject-matter concerns …

The expert in social mobility who says education cannot make it happen |  Education | The GuardianThe idea of causation as consequential manipulation is apt to research that can be undertaken primarily through experimental methods and, especially to ‘practical science’ where the central concern is indeed with ‘the consequences of performing particular acts’. The development of this idea in the context of medical and agricultural research is as understandable as the development of that of causation as robust dependence within applied econometrics. However, the extension of the manipulative approach into sociology would not appear promising, other than in rather special circumstances … The more fundamental difficulty is that​ under the — highly anthropocentric — principle of ‘no causation without manipulation’, the recognition that can be given to the action of individuals as having causal force is in fact peculiarly limited.

John H. Goldthorpe

  1. Gerald+Holtham
    June 2, 2022 at 4:55 pm

    Perhaps I’m wrong but I thought everyone agreed that the causal hypothesis comes first and is then tested on data. No data can demonstrate a causal hypothesis, only reveal it to be consistent with or inapplicable to the test data. If inapplicable it cannot be universally valid, though it may have applications elsewhere. If the hypothesis is consistent with the test data you are entitled to maintain the hypothesis. though its application outside the test sample remains a matter of conjecture. Your degree of confidence in it may well depend on the qualify of the story you tell about it but that is a subjective matter. One person’s “powerful deep causal mechanisms” are another person’s fanciful conjectures unless the data can adjudicate.
    Finding “relationships” by regression analysis is something quite different. That is not a matter of discovering a causal theory; it is merely looking for leading indicators, that is variables whose movements tend to precede others.as a way of forecasting variables in a system that you do not understand. A good regression equation will often forecast better than a mere guess but that doesn’t mean you have discovered a causal theory or social “law”. If there are people deluded enough to think so, I have not met them.
    One day I hope to understand “abduction”. I’ll come clean and admit I think it is just a word and does not signify anything real at all. Perhaps it refers to the jump of intuition that leads someone to frame a causal theory after looking at some facts. OK but as Popper said the intuition is essential for framing useful theories but it carries no guarantee or even evidence of validity.
    Finally I am surprised at Lars’ hostility to the Bayesian approach because it seems much more consistent with his general methodological position,, as I understand it, than frequentist statistics. Lars worries, for example, that statistics posits imaginary worlds of which reality is merely a sample. Whether or not that is a misinterpretation, it cannot apply to Bayesian statistics which explicitly considers the data as unique and determinate while all uncertainty resides in our characterisation of it.

    • Meta Capitalism
      June 2, 2022 at 11:34 pm

      Finding “relationships” by regression analysis is something quite different. That is not a matter of discovering a causal theory; it is merely looking for leading indicators, that is variables whose movements tend to precede others as a way of forecasting variables in a system that you do not understand. A good regression equation will often forecast better than a mere guess but that doesn’t mean you have discovered a causal theory or social “law”. If there are people deluded enough to think so, I have not met them. (Gerald Holtham, RWER, On Finding “relationships” in regression analysis aka “fishing”, 6/2/2022)
      .
      Econometricians who try to extract causal theories [finding “relationsihps”?] from data are said in the trade to be “fishing” and the term is normally pejorative. (Gerald Holtham, RWER, On “fishing” aka finding “relationships” in regression analysis, 1/27/2020)

      .
      If you never met any and presume there are none why in is “fishing” normally a “pejorative”? Doesn’t that presume there are such econometricians and you have (or know someone who has) met them? It certainly implies they exist and are frowned upon by wiser and more experienced econometricians. It seems the “I have never met them” is a tag line with no substance in light of these two statements.

  1. No trackbacks yet.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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: