Early this month, a Bloomberg campaign representative dialed a Manhattan phone number.

When the voter agreed to answer a few questions, the campaign representative launched into a series of questions on his feelings about Mayor Giuliani’s performance after September 11 and whether the former mayor’s support would make him more likely to vote for Mayor Bloomberg.

The campaign representative also asked whether the voter intended to vote for Mr. Bloomberg, if the voter was more likely to vote for the mayor given his success in reducing crime, and what the most important issue was in New York City today.

The final question was a request for the voter’s updated e-mail address.

The answers to the questions were all plugged into the Bloomberg campaign’s voter list – which, after a reported investment to date of more than $5.4 million, is the most expensive and comprehensive list ever established by a candidate in a New York City election.

“It’s one of the things we do to put together our database and have it available when we need it,” a senior Bloomberg campaign adviser, Bill Cunningham, said of the phone calls. “Once we have data, we can use it in any number of ways.”

Mr. Cunningham said the mayor’s reelection campaign wants to communicate to voters accomplishments such as cutting crime, adding housing, creating jobs, and improving schools.

“To do that,” he said, “we need the capability of talking to people about what they care about.”

A New York political consultant who creates voter lists for candidates, Jerry Skurnik, said almost all campaigns buy formatted voter information from the Board of Elections. That includes voters’ names, addresses, party enrollment, voting history, date of birth, and gender.

Many candidates also pay his company, Prime New York, to enhance the data, adding information such as phone numbers and voter ethnicity, Mr. Skurnik said. The data, while not 100% accurate, give campaigns an idea of the likely profile of the New Yorkers they are trying to target.

In the 2004 presidential election, the Democrats and Republicans enhanced their voter lists with consumer data as well, including such information as what cars the voters drive, how many television sets they own, whether they have children, and which magazines they subscribe to.

But Mr. Skurnik said that what the Bloomberg campaign is doing – paying a market-research firm, Penn, Schoen & Berland Associates, to compile personal information about individual voters, rather than or in addition to merging information from public and commercial lists to the Board of Elections’ voter list – is unique.

“No local candidate that I know has ever done that before, because no one has ever had the resources that Bloomberg does,” he said.

Closer to the election, the Bloomberg campaign will be able to call up specific interests and the feelings of specific voters and target their message accordingly.

For example, a voter who answered “no” to the question about whether Mr. Giuliani’s support would boost his opinion of Mr. Bloomberg would not receive mailings, phone messages, or e-mail correspondence from the former mayor.

A voter who said the most important issue was the environment might get a mailing about Mr. Bloomberg’s environmental record, or information about his winning the endorsement of the League of Conservation Voters, while a voter who said the most important issue in New York City today was education might receive campaign materials detailing Mr. Bloomberg’s schools initiatives and his success in boosting test scores of students in the public schools.

Not only will the thorough list allow the campaign to target messages to the precise interests of particular voters, it will also allow the campaign to avoid turning off certain voters – which could lead them to vote for Mr. Bloomberg’s opponent – and bothering, say, an 85-year-old retiree with information about the school system.

Mr. Cunningham said the campaign’s voter list would be more comprehensive this year than it was in 2001 because this year Mr. Bloomberg has a different goal.

“Last time, it was a matter of introducing Mike Bloomberg to New Yorkers,” he said. “Now it’s a matter of introducing his record. It’s extensive, so we have to figure out the best way to put it in front of people. … We just believe the more you know, the better you are able to communicate.”

The director of the Center for the Study of American Politics at Yale University, Alan Gerber, said the technique, known as “micro-targeting,” would probably give Mr. Bloomberg an advantage.

“Not every campaign has the benefit of knowing if the person in the house feels favorably to Rudy Giuliani,” he said by way of example.

While there’s no scientific evidence on the marginal return of an additional column of data on each member of the electorate, it seems clear that voter research is a good use of money, Mr. Gerber said. He said if Mr. Bloomberg spends a total of $70 million on his campaign, spending $10 million on research rather than direct communication with voters seems a reasonable expenditure. In his last campaign, Mr. Bloomberg spent $74 million.

“It really doesn’t have to improve the quality of your communications that much for that money to be money well spent,” Mr. Gerber said. “It’s a huge amount of money, but in proportion to the overall campaign activity by Bloomberg, we’re talking much less than 10%.”

An authority on politics and elections at the American Enterprise Institute in Washington, D.C., Norman Ornstein, said that until 20 years ago most national campaigns did not attempt to compile in-depth voter lists. He said last year, President Bush and Senator Kerry created sophisticated databases that allowed them to mobilize voters and target individual voters instead of reaching the larger mass of voters. He said it seems like the strategy stimulated a sharp increase in voter turnout.

Still, he said, the kind of voter lists that Mr. Bush and Mr. Kerry used in swing states is very rare at the state and city levels, mostly because of the price tag. He said no one knows exactly how much the Republicans and Democrats spent on voter-list development in the 2004 election, but he put it at somewhere between $30 million and $50 million on each side. That’s at most nine times what Mr. Bloomberg has already reported spending on developing lists in New York alone.

Not everyone was convinced that spending big on “voter list development” would help the mayor win reelection.

A spokesman for one of Mr. Bloomberg’s Democratic opponents, Anson Kaye, said his candidate, Rep. Anthony Weiner of Queens, would have a “smart and vigorous voter outreach operation,” but when asked if the Weiner campaign would spend anywhere near $5 million on a voter list, Mr. Kaye laughed.