Since "happiness" by definition is subjective, how do you determine what the proxies to happiness are? How do you correlate something to "happiness" (to determ

Since "happiness" by definition is subjective, how do you determine what the proxies to happiness are? How do you correlate something to "happiness" (to determine its a proxy), if you can't measure happiness directly? I mean, I agree that there's some significant problems in surveying a population, but I'm not aware of any other way to measure something subjective like happiness. I'd love to read the study you're referencing if you can find it.