[Gorm-Casper]

To make a general statement like that based on a study purely on USA, seems wrong.

I used to live in Denmark where they (according studies I won't bother digging up now) have both the happiest people on the planet AND pay the highest percentage in tax.

[CWuestefeld]

To make a general statement like that based on a study purely on USA, seems wrong.

On the contrary. The author seems to be trying only to apply her conclusions to America, so it would be appropriate.

Also, bringing other variables into the regression (societal factors, etc.) would just muddy the waters. If we've got a sample that excludes some variables altogether, it would make sense to use it (with the caveat that we can't know what effect those other dimensions might have).

[ajross]

Uh, what? Ignoring evidence is never correct. The burden goes the other way around -- you need to prove the absence of uncorrected correlations if you want to do a good science. You can't just limit your study to a cherry-picked set and announce that "you're only trying to apply your conclusions to the studied subjects".

[drp]

States in this case is referring to US states, so I wouldn't say the statement is general.