When you see that trend in the data surely you should say "that's interesting, I wonder if it's accurate and hence something to do with being Hispanic" rather than "By what devilry does my data suggest that Hispanics live longer?".
For example, what if my data showed that less affluent people do better in school, or have better health, than more affluent people.
Would that surprise you, and would a healthy response be to check your data again before chasing after this conclusion?
I think generally, yes, because it conflicts with our notions (from earlier studies) about how socioeconomic status affects success in schools and in health.
And that's likely at the core of the 'paradox', that hispanic communities are relatively less affluent than a median citizen, and that lower socioeconomic status is normally linked to negative, rather than positive, affects on health. The surprise isn't really concerning.
I mean I get your point, it's important to remain open-minded and ask open-minded questions, as opposed to questions which carry lots of preconceived ideas about the world. But in this case it doesn't seem all that concerning to me.
This is a paradox especially if your prior is that longevity and race should be unrelated.
Considering that they live shorter in Mexico and Latin America, and that they are less afluent that the mean in the US, it makes sense to expect that.
That's a totally reasonable assumption given that we know: 1) Hispanics on average have lower SES; 2) longevity is correlated with SES.