I really wish articles written about studies like this, where all they found was a association between stuff, and not a causal relationship, would explicitly say so. Perhaps through a little box near the end of the article stating, in simple terms, "the authors found that X and Y are associated. This does not necessarily mean that doing X will cause Y" I know the people who read the actual study in journal, entitled…
And in this case, you could easily imagine the arrow of causation pointing in the other way. If you have the good fortune of being generally healthy, and make enough income to have access to good healthcare, it's likely that you will live to 85. But these same factors also make you a likely optimist.
_That finding was independent of other factors thought to influence life's length — such as "socioeconomic status, health conditions, depression, social integration, and health behaviors"_
Literally the second paragraph on the article.
Read the article before posting comments, folks! :)