I went for matched pairs because I could justify it more- we've learnt the definition as being matched on characteristics the researcher feels are important, so not necessarily age, but perhaps social class, or area of the country, as they'd affect your habits as well as your age.
Now, if only I'd explained that decently in the exam :L
But yeah, I'm pretty sure they'll allow at least the majority of the marks for either of the two we've mentioned with a justification, especially if it's split as such here!
I don't see how you could justify repeated measures, since they can't be both an adult and a teenager, but the other two are equally justifiable, so I'd imagine at least some marks will be rewarded.
I put independent measures & then changed my mind because I thought that I could justify matched pairs better!