Bits of Books - Books by Title


Why Experts Keep Failing Us - And How To Know When Not To Trust Them

David Freedman

Cardiologists recognized that people with irregular heartbeats were far more likely to die within 12 days of having a heart attack, so when drugs to combat this arrived in 1980's, they quickly became standard practice. And yes, patient's hearts beat more regularly on the drug, but they were also 3 times more likely to die.

Reason was that doctors were looking at a proxy measurement, or marker. They knew that what they really wanted to measure was survival, but that takes long and complex studies. So, having observed an apparent link between irregular heartbeats and death, it was reasonable to assume that if they controlled that, it would keep heart patients alive.

These surrogate measurements tend to lead researchers astray. Cancer drugs like Avastin made tumors shrink, but without adding survival time or other benefits. What it did add was blood clots and bowel perforations, among other side effects. Cholesterol-lowering drugs Vytorin and Zetia show no indication of lowering heart-disease or stroke risk on average.

Experts can make the right measurements but on the wrong people - people in studies may be particularly health conscious or unusually ill. Participants are often paid, which over-represents poor people and/or students.

Meta studies can be misleading because they don't take into account studies discarded because they didn't show the desired result. A 2008 study revealed that 23 out of 74 anti-depressant drug trials weren't published, and all but one of those showed the drugs to be no more effective than a placebo.

More books on Mind

People like being told that there are things they can do to improve their situation. You like to think, that as far as possible, your life is under control. Dale Carnegie's How To Win Friends and Influence People laid out a simple formula - smile and listen - which would supposedly make anyone likable.

When we can personally relate to something, we believe much easier. So political campaigns today are woven around simple stories and narratives that make issues seem black-and-white: "Obama pals around with terrorists" stories.

But we are complex creatures in a complex world, so why should we expect simple answers to any questions?

And the problems that lead us to turn to experts - how to be healthy, wealthy and happy - are extraordinarily complex. If there were simple answers, we would have found them long ago, and we would have no need for experts.

This gives us a clue as to what expert advice we should be listening to - it will be complicated and hedged in with ifs and buts, and it won't be easy to act on. But the less well-informed or qualified 'experts' offer one-size-fits-all, simple answers are probably going to be less complete.

Groups are horribly ineffectual. They are dominated not by people who are right, but by the most belligerent, persuasive and forceful. Even those with rudimentary political skills can get a small alliance going, and doubters don't want to be the odd man out. The more people in a group, the more likely to have someone who will provide fake data to back up his chosen policy. And because group responsibility, people less concerned about errors.

"Hitchcock Effect" (after a short film in the TV series) in which a guy gets a series of detailed predictions, all of which prove correct, to the point where he is happy to trust the guy with all his savings. But it turns out that the mailer had started out sending varied predictions to a large group of people, then focused each subsequent mailout only on the people who had received predictions which happened to prove correct.

Every year there are at least 500 books published that promise to share the secret of how to miraculously turn your company into a highly-profitable firm. As if a book you picked up at the airport can solve the business problems that no-one else has been able to solve so far.

Almost any change introduced by management - even something as meaningless as changing the lighting slightly - tends to improve the output of workers being studied ... TEMPORARILY. A Harvard Business Review study looked at 40 years of management fads (defined as ideas that went from sudden prominence back to obscurity) and found common denominators - 'simple', falsely encouraging' and 'one size fits all'.

There is really only one strategy for a company to win big, and that is to take big risks. And of course the downside is that you are much more likely to lose big. But the books and articles are written about the successes, and so risky strategies look great. And the studies ignore all the failed businesses which employed the same risky strategy.

Even if you could minutely dissect a successful business and determine what made it successful, there is too much randomness in the world of business to figure out how to be like them.

Several websites try to combine search engines with human advice - NozyJoe, Wikia, Squidoo, Mahalo, Sproose, ChaCha, Knol and Delicious.

Author went looking for advice on how to transfer music from an iTunes format to a non-iPod device. Many sites either said it was impossible or gave incorrect advice. You can burn the protected tunes onto an audio CD (not an mp3 CD) and then rip them back from the CD into mp3 files, or you can get software that will automate this process.

Problem with advice websites is that there is a shortage of wise people with spare time, and a surfeit of less wise people eager to contribute. Compounded by scaling problems - starts out small community of active users, most of whom are involved. But then as grows, get lots of freeloaders and cranks/trolls. So it gets harder and harder to plough though it all to find what you need.

Reverse crowd sourcing. Guy with brain-damaged child encouraged others to upload xrays and medical records to give more insight to conventional experts.

Books by Title

Books by Author

Books by Topic

Bits of Books To Impress