$29 billion reasons to lie

[AMBrennan]

I just started reading the titular book by Justin Smith, and I was wondering if you noticed the error on the back? Don't cheat.

On the back of my issue of $29 billion reasons to lie about cholesterol it says:

Everywhere we look we are told that high cholesterol levels cause heart disease
[...]
However, the reality is that most people do not need to lower their cholesterol. In fact:

• Most people die of heart disease with low cholesterol than high cholesterol
• Most people who have a heart attack have an average cholesterol level

Which sounds really bad for the cholesterol hypothesis, right?

Well, take a closer look because there's a confounding variable that distorts the picture: The distribution of cholesterol levels.

For example, let's assume that the base risk of heart attacks is 10%, high cholesterol doubles it, and that 20% of the population has high cholesterol, you'd see the following in 100 people:

• 80 people with low or normal cholesterol, of which 8 will have a heart attack
• 20 people with high cholesterol, of which 4 will have a heart attack

Result: Most people with heart attacks (2/3) have low or normal cholesterol but, by assumption, high cholesterol doubles the risk.

Simply put, the statistic he quotes is unsuitable for proving his point. He may well be right about cholesterol, but you can't tell without additional information, making it a rather poor argument. It doesn't bode well that neither he, his editor nor anyone else involved in the production of the book noticed this.

 

[phoenix]

And just as your quoted statement is misleading so will be any results from your poll.
I've just voted twice on this poll, once yes once no, a fault with the software no doubt, but a fault that has been know for several years.

 

[librarising]

AMBrennan wrote
Simply put, the statistic he quotes is unsuitable for proving his point. He may well be right about cholesterol, but you can't tell without additional information, making it a rather poor argument. It doesn't bode well that neither he, his editor nor anyone else involved in the production of the book noticed this.

That's interesting, but that's all it is.

I would never go to the back of a book to find an 'argument.' That's what you read the book for.

Suppose someone said all pedestrian/car injuries and deaths were caused by someone driving a car. To reduce this, reduce the number of cars (high cholesterol) on the road.
Just as high a percentage of pedestrians (medium/low cholesterol) will be involved, and those involved got no protection by not driving themselves.

Perhaps high cholesterol is not the problem, but something else causing the high cholesterol.
For me, high cholesterol is neither good nor bad. High triglycerides/low HDL is bad.

I simply can't accept your 'poor argument' premise.
When people are told one thing, and the back of the book points out some additional facts that might question that belief, that's smart advertising.

Geoff

 

[borofergie]

And just as your quoted statement is misleading so will be any results from your poll.
I've just voted twice on this poll, once yes once no, a fault with the software no doubt, but a fault that has been know for several years.

I think that the American Electoral system works along the same lines...

 

[borofergie]

Don't judge a book by its cover?