Newsweek has a poll out showing
[John] Kerry … leads the pack of Democratic contenders among registered voters as the candidate who would have a better chance of beating President George W. Bush if the election were held today. A Kerry-Bush match-up would have Kerry up by 49 percent to Bush’s 46 percent. A Clark and Bush match-up would be a close race, with Bush at 48 percent and Clark at 47 percent. Bush would have an edge over Edwards (49% to 46%). Yet, with a plus or minus margin of error, these match-ups result in a statistical dead heat. And the President would beat Dean (50% to 45%) and Sen. Joe Lieberman (49% to 45%).
Newsweek buries the margin of error at the very bottom (+/- 3%) It does not say what question it asked to get that result. The results released also do not give any information on the political makeup of who was asked and whether the poll surveyed likely voters or “1,006 adults aged 18 and older on January 22-23, 2004,” which is what the press release says.
If they did use the latter instead of the former, the results are meaningless except to help John Kerry in New Hampshire.
Given the fact that no other poll has reflected Newsweek’s poll, I doubt Newsweek’s poll is accurate. In fact, I’d say it is highly misleading and not very journalistic.