Not necessary, because the argument is based on the lack of impact on the “nearly all” voters who have already made a decision. Whether or not columnists might influence the readers who haven’t already made a decision is irrelevant, because we already know that the vast majority of people have made a decision.

Not necessary, because the author’s position is just that the columnists are rarely successul in persuading readers to vote a particular way. This doesn’t imply that any readers are caused to vote a different way. The readers might simply be unpersuaded to change and simply stick with the candidate they were already going to vote for.

Not necessary, because the argument doesn’t rely on any claims about people who “regularly read columns.” We know that nearly all people have already decided who to vote for by the time a column comes out. This is true regardless of whether those people are regular readers or not. And although the author assumes that these people are unlikely to change their minds because of the column, that doesn’t imply that the author thinks it’s because the people are regular readers who already agree with the columnist.

Not necessary, because the author never compares “regular readers” of columnists to people who “seldom read” such columns.

Necessary, because if it were not true — if it’s not rare for people to be persuaded to change their minds about who to vote for after they’ve already made a decision — then the fact most people have already made a decision before a column is published does not establish that the column is unlikely to change their minds. In other words, the negation of (E) shows that it might not be rare for columns to persaude people to vote for certain candidates, even if those people came to the column already decided.