The most striking thing from the book are the sections on confidence. By acting confident in your abilities and predictions, people's confidence in you goes way up. If you hedge your predictions, act uncertain at all, or consult reference material - their confidence in you goes down, and satisfaction in your performance also goes down.
So what's the problem? Well, the scariest part for me was talking about doctors. They set up an experiment where they filmed two actors playing a doctor and a patient.
The patient was asking the doctor about getting antibiotics prescribed before a dental procedure, which might or might be necessary and helpful. They presented one of three scenarios to viewers and asked them to rate the doctor and how satisfied they'd be with the doctor:
Scenario #1: The doctor prescribes the antibiotics confidently, with no hedging or uncertainty.
Scenario #2: The doctor says, "I'm not sure if these are really necessary... but what the heck, there's no harm" and prescribes the antibiotics.
Scenario #3: The doctor consults a reference medical book to get numbers and clinical results in front of the patient to be sure it's really necessary, and then prescribes.
What's scary is that the scenario #3 doctor was given the lowest ratings on performance, even lower than "what the heck, there's no harm" cavalier attitude. But the authors noted that in real life, it's the most skilled and thoughtful doctors who question their own judgment and are willing to check reference materials to make sure they're doing right by their patients.
The book provided similar examples in other fields, including the judicial system and academia. Confident people are believed much more often. Estimating the chances you're wrong, admitting you might not be correct, or consulting reference material lowers people's belief in you and their satisfaction in your performance.
This is a huge, nasty conflict of interest. If you want people to believe you, you should use raw confidence, you should refrain from hedging your predictions or admitting you might be wrong, and you shouldn't publicly consult reference material in a case where you might have incomplete information.
Of course, that's pretty much the opposite of what you should do if you're looking to maximize the chance of a correct outcome. Some reasonable self-scrutiny, estimating the odds your prediction is wrong, and checking reference materials are all likely to make you more correct.
So there's a big conflict of interest here - people will be most satisfied with your performance if you admit no chance of error and consult no outside materials. On the other hand, this makes you more likely to make block-headed mistakes. Discuss.