Selfishness can also mean whatever ones does for oneself in a rational context always includes what is non-harmful or beneficial for others as well. Doing things that don't harm others is good for us. Acting selfishly for our own benefit in a rational manner is best for all, one could say.
The caterpillar does indeed need to struggle out of the chrysalis, and other example. But that's not the general idea, although valid to suggest as applying. Anyhoot, I get your reservations, but I tend to agree with the overall message, as children from the same parents can sometimes mimic the selfishness and other times not be like those parents. The same can be said for the reverse with empathetic and supportive parents, but this is not to say to baby them like you seem to suggest. I can picture the general idea they are putting forth, despite there lacking a larger sample size or more verification. It's valid from my view of things. I don't think they, or I, tried to paint it as black and white, sorry if it came out that way by maybe not explicitly covering how these qualities aren't absolute, or exclusive.
That is the main thing I tried to get across. ;)
I do not think you tried to paint it in any way, but the original article did. It might be that I am just very biased in general towards statistics and especially the ones based on surveys. To shortly explain that:
People in a poll might not be honest. Just because someone says he is social does not mean he actually is.
Statistics can only discover correlation, but people tend to make causation conclusions.
Studies are often conducted to reach a certain conclusion, so there is a incentive to rig the surveys in slight ways (maybe even subconsciously) that help to reach said conclusion.
There is no reason to apologize to someone like me, who has severe trust issues when it comes to statistics (and my personal life :D). Thank you for your extensive replies.