# Thread: The Happiest Man in America

1. They asked 0.0003% of the US population if they're happy. That can't be accurate.

2. All waves lead to Rome?

3. To happiness maybe.

4. Originally Posted by Vasilisa
Gallup’s answer: he’s a tall, Asian-American, observant Jew who is at least 65 and married, has children, lives in Hawaii, runs his own business and has a household income of more than \$120,000 a year. A few phone calls later and ...
Well, this statistical analysis reaches a rather obvious conclusion: since those traits tend to be highly uncorrelated (tall+ asian american + observant jew + lives in hawaii? Their conditional probability must be almost zero), extremely happy people are extremely rare. Lies, damn lies and statistics, indeed

5. Originally Posted by FDG
Well, this statistical analysis reaches a rather obvious conclusion: since those traits tend to be highly uncorrelated (tall+ asian american + observant jew + lives in hawaii? Their conditional probability must be almost zero), extremely happy people are extremely rare. Lies, damn lies and statistics, indeed
But... if you are a tall asian american now you know the formula to use to maximise your potential happiness!

6. Originally Posted by guesswho
They asked 0.0003% of the US population if they're happy. That can't be accurate.
http://en.wikipedia.org/wiki/Sample_size

As long as you sample only a small percentage of the population, population size doesn't influence accuracy. A sample of 4000 people out of 1000000 is not noticeably more accurate than a sample of 4000 people out of 1000000000.

If your sample is 4000 out of 6000 than this obviously doesn't apply. This is due to it being impossible to have values like 100% happy if only 50% of people are happy. It's only when hard limits start to be set by the sampling that it gains accuracy due to population size.

So, Massachusetts doesn't stand out much except on overall and health insurance. Big surprise that mandatory health insurance puts you on top of health insurance percentage.

7. Originally Posted by guesswho
They asked 0.0003% of the US population if they're happy. That can't be accurate.
Yes, I daresay there is huge uncertainty in the data.

The part I object to though, is the idea that a transient questionnaire can validly measure ones happiness over time.

BlueGray - it is true that precision is not the same as accuracy. But 'significant' confidence interval cutoffs themselves are simply arbitrary. If you sample only 4000 people from a very large group, the chance of the data being unrepresentative is always higher than if you use a larger sample.

8. My place isn't so bad . Or actually, the 4 districts that encompass my city are doing pretty good. I think doing this in different cities/county borders would at least be a slightly better idea instead of through districts. Huge swatches of land like Montana and North Dakota don't really say much.

What I found interesting are the statics on inadequate food, inadequate shelter, diabetes, obesity, high cholesterol, smoking and exercise. Like how the scores here for diabetes and obesity can easily relate to the one done with cdc.

It is nice to see how people view their own area though. Kind of nice knowing that some districts view their work environment, job satisfaction, and community improvement on such nice standards.

9. Another interesting/slightly depressing US map

10. Originally Posted by Vasilisa
Another interesting/slightly depressing US map