But my whole point was that if it's an empirical statement, then we shouldn't be offended by it.

I'm going to sidestep the talk of "offense" because I think it's sufficient to talk about whether a statement is morally right or wrong ("offensive" seems to be "morally wrong" with some extra baggage).

Two cases in which I might judge an empirical statement as morally wrong:

1) the statement is false, and yes, saying false things is usually considered morally wrong

2) the statement is true, but is used in a context where it will have negative repercussions - for example, telling your kid a huge amount of factually true statistics that cast a bad light upon a group you don't like (blacks, jews, women, etc.), or teaching a madman how to make explosives, etc.

In this case we're talking about the value a statement not in the abstract, but as life advice given from a father to his daughter. The important part isn't as much the truth of that particular piece of advice, but of what it allows us to infer about the general quality of the life advice given.