T Nation

Atkins Diet Study

This is from http://www.livescience.com/humanbiology/070410_bad_atkins.html

Retards. They beat into you in a statistics class that a 95% confidence interval means that there is a 95% chance that the true mean is somewhere between x-y. That does NOT mean that there is a 95% certainty that they overlap, as the article implies. The Atkins could have a significance of p<.05 from 15-7, and the Ornish could be p<.05 from 8-2.

You forgot the gold

[quote]Thousands of years of data

Most humans for the past several millennia have eaten a diet largely comprising grains, seeds and vegetables (that is, carbohydrates) with a little meat. Country by country, populations become obese when they adopt an American diet high in animal fat and simple sugars.

An animal-based diet is costly to the consumer and the Earth. Rice and beans are cheaper than meat; and millions of acres of more land would be needed to grow the corn to feed the animals to make the meat for the baloney-and-Diet Coke diet.

At his best, Robert Atkins preached against the evils of simple carbohydrates, the sugars found in processed foods and drinks. Studies have shown the link between these kinds of foods and weight gain. A healthy diet could be as simple as avoiding packaged foods and using meat as a condiment, as our ancestors did.

As millions of Americans exercising and dieting earnestly will attest, losing weight is very difficult. The best advice for the young is to try your best not to put on extra weight.[/quote]

Maybe the article should be renamed “Whiny meat-hater shows that there it is possibility that low-carb diets aren’t for everyone”.

[quote]LoneLobo wrote:
Retards. They beat into you in a statistics class that a 95% confidence interval means that there is a 95% chance that the true mean is somewhere between x-y. That does NOT mean that there is a 95% certainty that they overlap, as the article implies. The Atkins could have a significance of p<.05 from 15-7, and the Ornish could be p<.05 from 8-2. [/quote]

When testing your null hypothesis that the difference equals zero, you may do so by looking at the confidence intervals, i.e. if they overlap than you cannot say that there is a significant difference between them at that confident (alpha) level.

Although I agree that the authors made it appear that there was a 95% chance that they overlap, which is incorrect.

Is very low fat similarly restrictive in the sense that “very low” implies when stated for carbohydrate levels?

Haven’t we figured out that lipids are an important part of our diet and involved in many hormonal pathways, cells walls, the brain and other vital bodily components and functions?

[quote]vroom wrote:
Haven’t we figured out that lipids are an important part of our diet and involved in many hormonal pathways, cells walls, the brain and other vital bodily components and functions?[/quote]

Sigh, apparently not…

[quote]LoneLobo wrote:
Retards. They beat into you in a statistics class that a 95% confidence interval means that there is a 95% chance that the true mean is somewhere between x-y. That does NOT mean that there is a 95% certainty that they overlap, as the article implies. The Atkins could have a significance of p<.05 from 15-7, and the Ornish could be p<.05 from 8-2. [/quote]

True. I don’t see how i missed that. But what I thought was funny was that they asked the creator of the diet. I’m sure his opinion was impartial.