3

I'd like to assign undergraduate students with little to no math experience an article, short part of a book, or even a blog post about causality and counterfactual logic that is easy to understand.

It seems that most articles and books are too advanced for undergraduate students unfamiliar with formal logic or statistics.

firebird17139
  • 129
  • 2
  • 9
  • 3
    Look for books and articles by Don Rubin and Judea Pearl. – Michael R. Chernick Mar 25 '18 at 20:23
  • 1
    I think both are excellent writers but I'm not sure whether they have anything for more popular audiences. Do you have something specific in mind? – firebird17139 Mar 25 '18 at 21:34
  • 2
    I think Chapt. 2 & 3 from Morgan & Winship's "Counterfactuals And Causal Inference" is great for what you want. They offer a good introduction to what the authors denote as part on: "*Counterfactuals, Potential Outcomes and Causal Graphs*". These opening chapters assume *very* little Maths. The draft from Hernán & Robins' "[Causal Inference](https://www.hsph.harvard.edu/miguel-hernan/causal-inference-book/)" is very good too if you want to pick and choose specific topics but the first chapters take a while to get the counterfactuals in motion, it is better for the "graph" aspect of things. – usεr11852 Mar 25 '18 at 22:19
  • 1
    Also *maybe* (depending on what you expect from the students) you want to consider the articles on [Counterfactual Theories of Causation](https://plato.stanford.edu/entries/causation-counterfactual/) and [Probabilistic Causation](https://plato.stanford.edu/entries/causation-counterfactual/) from the Stanford Encyclopedia of Philosophy. The Maths/Stats requirements vary a lot within the text but on the plus side all students will have immediate access and it is quite throughout. – usεr11852 Mar 25 '18 at 22:27
  • To be clear, there is no way to get people to understand logical or probabilistic models of causation without at least a little knowledge of logic or probability, respectively. So presenting some amount of background material may be inevitable. – Kodiologist Mar 26 '18 at 19:59
  • 1
    You might be right, @Kodiologist. I do like the Stanford Encyclopedia entry. The historical approach to understanding causality is a really good idea! – firebird17139 Mar 27 '18 at 20:29

3 Answers3

5

Two longer recommendations that might fit the bill (depending on exactly what no math experience means):

  1. Causal Inference in Statistics: A Primer by Judea Pearl,‎ Madelyn Glymour, and ‎ Nicholas P. Jewell - a short book that covers the basic problem, with the first chapter covering basic probability concepts
  2. Scott Cunningham's Causal Inference: A Mixtape - a draft of a book with lots of great empirical examples done in Stata. Probably less demanding than the former.
dimitriy
  • 31,081
  • 5
  • 63
  • 138
  • 1
    Perhaps attacking this issue from a historical/philosophical approach is a more appropriate method. Pearl and Cunningham's books are fantastic introductions to the logic of causal inference though, but I assume no math experience at all with these students to be honest. One could take sections of the books and make them into a useful presentation of causality concepts. – firebird17139 Mar 27 '18 at 20:33
3

Pearl's The Book of Why is coming out soon, it might be an interesting read for students without background in math but that want to get started in causality. The Epilogue in Causality (The Art and Science of Cause and Effect) has a nice philosophical overview of the topic and is also a good read.

Carlos Cinelli
  • 10,500
  • 5
  • 42
  • 77
2

I recommend the chapter "Causal Inference" of Larry Wasserman's All of Statistics. It's 13 pages long, it has a few exercises, and it covers the counterfactual approach to causal modeling in a way that only requires basic knowledge of mathematical statistics (e.g., conditional probability). As a bonus, the following chapter covers the DAG approach to causal models, if you want to throw that in.

Kodiologist
  • 19,063
  • 2
  • 36
  • 68
  • 2
    I like Wasserman but I'd like to, as much as possible, avoid any math whatsoever (this isn't technically a stats class) and I think DAGs get complicated quickly. I'd like the students to understand the intuition behind causality before getting into modeling. – firebird17139 Mar 27 '18 at 20:27