This doesn't weaken the author's argument — if anything, it adds some support. By eliminating one possible way the polls could have been biased, this makes the polls seem more representative of people graduating university generally, regardless of their discipline.

This strengthens the author’s argument. By explaining that people with a university education tend to have been influenced by their professors, who are opposed to reducing government social services, this answer choice just offers an explanation for why the author's conclusion might be true. Note that this answer choice even bridges one of the assumptions in the author's argument, by giving us a reason to think that people with a university education generally, not just graduating students, will be opposed to reducing social services.

This answer choice is tricky, because we have to infer from the fact that more retired people without a university education want to reduce social services that fewer of them want to keep social services, compared to retired people with a university education. In other words, retired people with a university education are more likely to favor keeping social services than retired people without a university education.

This actually strengthens the author’s argument. By focusing on people later in life, rather than only graduating or incoming university students, this adds support to the claim that people with a university education, in general, are more likely to favor keeping or increasing social services than people without a university education.

This weakens the argument. It tells us that while graduating students are disproportionately against reducing social services, people polled five years of graduation are disproportionately in favor of reducing social services. It seems like people's views on social services might change after they graduate from college. Thus, it isn't necessarily true that all people with a university education will be more likely than average to be in favor of retaining or increasing social services.

Irrelevant. We're not interested in the strength of people's opinions on this topic, just whether they favor reducing social services or not, and the effect of a university education on their views. Notice that we can't assume that people with stronger opinions are more likely to answer polls on this topic, and so the polls are biased. This answer choice just tells us about the opinions of people who had been selected for the poll, but doesn't tell us anything about how they were selected for inclusion. It could have been a random selection, and this answer choice could still be true.