Kerry had the support of 52 percent of respondents, compared to 42 percent for Bush, with 3 percent backing Winsted native Ralph Nader and 3 percent undecided, according to the poll conducted by Maryland-based Research 2000.
Kerry has consistently led Bush in other statewide polls, including a Quinnipiac University survey conducted late last month that gave the Massachusetts Democrat a 6-point lead.
In this week's poll - commissioned by the JI, The Day of New London, and the Waterbury Republican-American - Democrats overwhelmingly supported Kerry and Republicans were equally enthusiastic about Bush.
Meanwhile, voters who identified themselves as independents split 47 percent for Kerry to 43 percent for Bush.
The respondents said Kerry could handle the economy and the war in Iraq better than Bush by similar margins - 48 percent to 41 percent on the first issue and 48 percent to 42 percent on the second.
Asked specifically for their opinion of Bush's handling of "the situation in Iraq," 56 percent said they disapproved, 42 percent said they approved, and 2 percent said they weren't sure.
Kerry's advantage evaporated, however, when respondents were asked which candidate could better protect the United States from another terrorist attack. Bush was favored by 49 percent, while 38 percent picked Kerry, and 13 percent said they were not sure.
The poll also revealed that while 56 percent of the women who participated backed Kerry, only 48 percent of the men did.
Bush had the support of 47 percent of the men, compared to just 37 percent of the women.
The poll gave Dodd a 66 percent to 27 percent lead over Orchulli, with 7 percent undecided.
Similar responses were recorded when the voters were asked their opinions of the Senate candidates, with 64 percent giving Dodd a favorable rating compared to 24 percent for Orchulli.
A majority, however, said they had no opinion about Orchulli, while only 14 percent had no opinion of Dodd.
Dodd enjoyed a slightly bigger lead over Orchulli in previous statewide polls, which gave the incumbent advantages of 44 and 45 points.
The poll was conducted by Research 2000 of Rockville, Md., between Tuesday and Thursday. Respondents were selected by the random distribution of the last four digits of their telephone numbers. A cross-section of exchanges was used to ensure an accurate reflection of the state, and quotas were assigned to reflect voter distribution by county.
The poll had a margin of error of plus or minus 4 percentage points.
