Chris Esposo's Reviews > Mindware: Tools for Smart Thinking

Mindware by Richard E. Nisbett
Rate this book
Clear rating

M 50x66
's review

liked it

A decent book on heuristics and frameworks one can leverage to make sound personal and professional judgements. The first 1/3 of the book focuses on psychological errors people tend to make when forced into ad-hoc decision making, most of this section was almost entirely forgettable and there's a lot of redundancy with other pop-psychology books, as well as behavioural economics work popularized of the past decade. There's also a bit on using cost-benefit analysis for decision-making, but unless someone is going to pull out a pen and paper to reconcile the accounting between the costs and the benefits of the situation, this is a poor recommendation for normal people.

The middle 1/3 deals with leveraging statistical tools to make decisions. The author makes an especial subject of the multiple linear regression equation (MLR). The author's, at length, lament of the inability to use MLR is elementary and misleading. There's actually a whole lot of literature and research on how one could tease out causalities while feature-engineering the inputs for MLRs, from engaging in a conditional/mediation analysis of the features to doing causal inference more recently, to which there's an excellent book by Judea Pearl that was just published for mass-audience on the subject. So although the author is correct that a naive deployment of MLR will have these problems, there are ready-made processes to address the issue of discovering causation in current understanding.

The author uses his simple version of the MLR as a straw man to argue that causality can best be discovered through studies made from randomized control trials (RCTs). Again, this isn't the last word on the subject, Judea Pearl's book, "The Book of Why" is an excellent and detailed exposition on why RCTs might not actually be the best solution, and certainly not the only one. A good thing to remember that the author does hammer in, in this section, is that the lack of observed correlation does not imply the lack of causation, which I think is another way for him saying that the lack of evidence does not prove the desired hypothesis.

The last 1/3 is mostly concerned with using formal logic tools to make sound decisions, including the classical syllogism and the dialectic. The author makes an interesting side trip here and comments that formal logic, in the western vein, was never developed in China, but instead, the Chinese of antiquity utilized dialectical reasoning to discover many natural and social phenomena. This statement is only partially true, according to one of Joseph Needham's volumes on the subject, a type of syllogism and a proto-predictive logic was developed in early China under the Mohist school of thought. This school lost court favour and were "removed", but "logic" or games of logic were not unknown to Chinese scholars. The author also makes the point that the Chinese were also aware of the Indian tradition of logic, but did not seem to leverage it much in their own works. The point here is to be familiar with both approaches. There also seems to be a similarity with this the syllogism/dialectic dichotomy identified here with the synthesis/analysis dichotomy identified by Kant on his short treatise on logic.

Overall, a good read, with many redundancies with pop-psychology books at the beginning of the text. It needed more "context" and detail in the statistical discussion, but would be great as the first book on clear thinking. Conditional recommend

Sign into Goodreads to see if any of your friends have read Mindware.
Sign In »

Reading Progress

Started Reading
October 9, 2017 – Finished Reading
January 9, 2019 – Shelved

No comments have been added yet.