INFORMS Open Forum

The next INFORMS Book Club meeting will be held on Thursday, February 23rd at 1:00 ET. Have you started reading the book yet? If so, what do you think? Have an opinion you'd like to share with the group or a topic you'd like to discuss? Please post it here! And in case you haven't registered yet, here is the link:

https://us06web.zoom.us/meeting/register/tZ0ldeqqpj4qGtV-2In4G-NqQrv_57Codne2

Look forward to seeing everyone on February 23rd!

------------------------------
Jill Capello
Membership Associate
INFORMS
Catonsville MD
jcapello@informs.org
------------------------------

I read it about 5 years ago. It was an interesting experience, because in my area (operations management) no one was using that kind of approach, and I felt quite skeptical.  The author's style and tone was one of the reasons for it -- I am not used to such assertiveness in writing. However, soon after that I started working closer to an ITM department, and everyone was talking about "mediation effects" and "causal modeling", though I have not yet seen formulas "do-calculus".

One of my favorite parts of the book was the cause-of-fire example: "the fire wouldn't have happened if someone didn't light the match" vs. "the fire wouldn't have happened if the house didn't have oxygen in it". It illustrates nicely how ambiguous the idea of cause and effect is, especially when it comes to placing the blame.

------------------------------
Yulia Vorotyntseva
Saint Louis University
St Louis MO
------------------------------

Science had an interesting article recently about using statistics in medical murder cases to "prove" guilt.
https://www.science.org/content/article/unlucky-numbers-fighting-murder-convictions-rest-shoddy-stats
I'm not familiar enough with the type of statistics in the book to say if its approach could have been used or would have helped if it had. What the Science article did remind me of from the book (I also read it several years ago) was the difficulty in accounting for the many possible factors (including issues with the data and the statistical analysis) in trying to reach a conclusion or recommendation. And some of the examples in the article of factors that should have been included look "obvious" after someone brings them up later, but how do you make sure you identify and include the most important factors during the original analysis (or in advance), rather in hindsight (when time and resource pressures are much less)?

------------------------------
John Clifford
Alexandria, VA, United States
------------------------------

Original Message:
Sent: 01-25-2023 14:45
From: Yulia Vorotyntseva
Subject: INFORMS Book Club - Have you started reading "The Book of Why: The New Science of Cause and Effect"?

I read it about 5 years ago. It was an interesting experience, because in my area (operations management) no one was using that kind of approach, and I felt quite skeptical.  The author's style and tone was one of the reasons for it -- I am not used to such assertiveness in writing. However, soon after that I started working closer to an ITM department, and everyone was talking about "mediation effects" and "causal modeling", though I have not yet seen formulas "do-calculus".

One of my favorite parts of the book was the cause-of-fire example: "the fire wouldn't have happened if someone didn't light the match" vs. "the fire wouldn't have happened if the house didn't have oxygen in it". It illustrates nicely how ambiguous the idea of cause and effect is, especially when it comes to placing the blame.

------------------------------
Yulia Vorotyntseva
Saint Louis University
St Louis MO
------------------------------