In the social sciences, it's usual to use an operational definition [1]. For many variables are constructs [2] and they cannot be observed directly. The author

In the social sciences, it's usual to use an operational definition [1]. For many variables are constructs [2] and they cannot be observed directly. The authors of the above paper seem to have used different operational definitions, and they all point in the same direction. So one can state with reasonable certainty that women's happiness declined in the last 30 years. A discussion about potential explanations for the phenomenon can hardly be called pointless. [1] http://en.wikipedia.org/wiki/Operational http://en.wikipedia.org/wiki/Operational [2] http://en.wikipedia.org/wiki/Construct_(philosophy_of_science) http://en.wikipedia.org/wiki/Construct_(philosophy_of_scienc...