Having read A Field Guide to Lies twice - once thoroughly, and once more in a more cursory fashion because it was so good (!) - I must say that the cognitive psychologist and neuroscientist author Daniel J. Levitin really hits the mark. In today's world, we are bombarded with so much misinformation, and that is ostensibly because there are so many people who are tripped up by common fallacies which Levitin brilliantly debunks and unravels. (So, misinformation is allowed to propagate as long as it has a willing audience.) But I digress: he covers the whole "bag of dirty tricks" in an advertiser's arsenal: the confusion of mean with median, unlabeled or uneven axes on graphs, correlation implying causation, and lack of any mention of a base rate in telling your audience about a percentage of some shocking figure, e.g. death rate from breast cancer with or without screening at a given stage.
On that last item, he writes at length about Bayesian statistics, i.e. the statistical analysis of conditional probability, or "probability of A given that B..." or vice-versa. This acknowledges the possibility of false positives or false negatives in any sort of clinical test, which may not be revealed or may be framed in a misleading way.
Another crucial distinction in the health sciences that Levitin explores and elaborates on is that of incidence vs. prevalence vs. mortality rates (of a given disease or illness). He explains that incidence is those who pick up a sickness at any given point in time, while prevalence measures those who already/currently have the sickness. (You can see how this was relevant with the recent COVID-19 pandemic, even though the book predates that!)
So you can also see how some health products companies or advertisers could compel you into making an impulse purchase, or even a bit of a panic.
In any case, for me the big take-away was how many of us use conditional probability / Bayesian learning without realizing it; but we are also prone to ignore it based on heuristic "snap" judgements. His writing really helps in stimulating this "blind spot" residing in so many of us, and so for that reason alone I highly recommend it as an engrossing read. I can also truly say that Bayesian learning and methods are a significant part of analytics and machine learning, so by that token it has great relevance to a professional community such as INFORMS.
------------------------------
Jason Oliver
Project Leader, Compliance Research and Data Science
CANADA REVENUE AGENCY
Ottawa ON
------------------------------