I discuss a change from the past in which most Cuban-Americans, at least in Florida, were reliable voters for Republicans in Presidential elections.
How is this a change from the past, when this very poll, the very numbers that you're quoting, shows most Cuban-Americans in Florida currently supporting Romney over Clinton in a hypothetical presidential matchup?
Now, you can argue that the sample size on these subgroups are too small to mean anything, but that's not what you're saying. You seem to think that this poll shows Cuban-Americans in Florida supporting Democrats when that's the opposite of what it shows. It shows Romney doing better with Cuban-Americans in Florida than he does with either non-Cuban Hispanics or with whites. Is that really so confusing?