These studies are always funny. So what this is trying to tell us is: Volunteering boosts your health, while the data only provides us with "volunteers are generally in better health". Now this study "tried" to adjust to that by some filtering, but this is still a pretty big leap. Essentially what they would need to do is this:
Have one group of people who "actively" is kind to people, i.e. they actually act on their desired.
Then compare the results to a control group who "wants" to be kind to people, ideally has a proven track record of being kind to people, but are not allowed to act on that during the duration of study. Then of course they also need to still be as satisfied and happy as before, otherwise they turn into a biased control group.
Then if there are no differences, and compared to another control group of "normal" humans, there is a statistically significant improvement in health, then and only then, they may be on to something... Otherwise this is just another case of survivor bias.
This is all just pointless. At least here it's for a good cause, but I always am amazed by what kind of conclusions people derive from the well established: Causation implies correlation... Erm, NOT.
Have one group of people who "actively" is kind to people, i.e. they actually act on their desired.
Then compare the results to a control group who "wants" to be kind to people, ideally has a proven track record of being kind to people, but are not allowed to act on that during the duration of study. Then of course they also need to still be as satisfied and happy as before, otherwise they turn into a biased control group.
Then if there are no differences, and compared to another control group of "normal" humans, there is a statistically significant improvement in health, then and only then, they may be on to something... Otherwise this is just another case of survivor bias.
This is all just pointless. At least here it's for a good cause, but I always am amazed by what kind of conclusions people derive from the well established: Causation implies correlation... Erm, NOT.