Reprinted from The Wall Street Journal, October 5, 1972
WASHINGTON - A man with Trotsky-style eyeglasses and a bushy mustache is sipping
bourbon from a coffee cup and manipulating the tape recorder that's playing
"Jesus Christ Superstar." His employees, mostly long-haired young
people, are sitting on the floor, eating potato chips. A girl with a felt-tipped
pen is circling the hole in a male colleague's shoe. Somebody throws a paper
wad.
What's going on here?
What's going on here is an ordinary Friday afternoon staff meeting at Independent Research Associates, Inc., a political polling and analysis concern. The meeting looks as casual as a pot party, but the talk is all business: deadlines, computer printouts and plane schedules. For 33-year-old William R. Hamilton, the man with the mustache, runs one of the most active firms in this small but competitive field, and what he and the people on the floor do in the next few weeks will influence the course of elections throughout the U.S.
Mr. Hamilton is one of a small number of private pollsters - as contrasted to public pollsters such as Lou Harris and George Gallup. The private pollsters work directly, and confidentially, for political candidates who want to know as precisely as possible what the electorate thinks so they can devise an effective campaign strategy.
Among Mr. Hamilton's clients this year are John D. Rockefeller, IV, the West Virginian Democrat who's running for governor; Virginia Democratic Sen. William Spong, up for reelection; Indiana Democratic Rep. John Brademas; Congressional candidates in Florida, Ohio, Mississippi, and New Jersey; a candidate for parliament in Canada and a gubernatorial candidate in Puerto Rico.
Successful Hamilton customers have included Democratic Sens. Harrison Williams of New Jersey, Vance Hartke of Indiana, and Lawton Chiles of Florida, all of whom won in 1970. But Mr. Hamilton's clients don't always finish first, of course. He polled early this year for Sen. Edmund Muskie's ill-starred presidential primary campaigns. Another client, Georgia Sen. David Gambrell, was defeated in a Democratic primary run-off election in August.
What pollsters do, how they do it and why are matters of national debate every election year, and this year it's no different. The other day Democratic presidential nominee George McGovern, trailing badly in the polls, announced that polls are "rubbish" and are "made up in the back rooms of the polling organizations." A House subcommittee, meanwhile, is considering legislation to require polling firms to disclose their methods.
The debate centers on the public polls, mainly because their findings are published widely and regularly. But all polling firms, public and private, use generally similar sampling and interviewing techniques. In view of the controversy that surrounds these men who figure so heavily in American political campaigns, it's worthwhile to look closely at the style and techniques of one of them.
Perhaps the first thing to understand about Mr. Hamilton is that he can be scrupulously objective when he's wording a questionnaire or picking a block in suburban Atlanta in which to conduct interviews. And it's those two types of decisions, he says, that determine more than anything else whether a poll is accurate.
But the second thing to understand is that he has strong political opinions, which he keeps about as well-hidden as the four-letter words that lace his conversations. "I believe in the two-party system," he says, "but I just think the Democrats are better than the Republicans. I'm a wild-eyed populist. I like to work for somebody who wants to knock off one of the stodgy old guys."
So his firm works almost exclusively for liberal Democrats, although sometimes it will accept a liberal or moderate Republican who is challenging a staunch conservative.
A strong ideological bent may seem surprising for a man whose job is to measure public political opinion objectively. But a number of polling firms are identified with one party or the other. The political division of Market Opinion Research in Detroit, for example, polls for Republicans. Hart Research Associates here, a chief competitor of the Hamilton firm, is known as a Democratic outfit. And the pollsters insist such ties don't harm their work.
Mr. Hamilton says he can eschew ideology altogether when he's reporting data to a client and making campaign recommendations. He emphasizes further, that he and his staff are highly motivated to do a good job because they work for candidates who suit the Hamilton philosophy.
Peter Hart, a competing pollster adds: "I only want to work for a guy who I want to see elected." Mr. Hart says that having personal political feelings "doesn't make you any less objective with numbers. The science (of polling) remains the same. What it comes to is this: A good pollster can make an awful lot of difference in a campaign, so the question is, who do you try to help?"
Polling is the basic service the Hamilton firm provides, and to poll for a typical congressional race Mr. Hamilton charges about $5,000. For that fee, the firm interviews in detail a sample of 300 or 400 persons who are likely to vote.
From census data, the pollster ascertains the composition of the district's population in terms of race, sex, age, and other factors. He then tries to build a sample of likely voters that reflects that breakdown. Each interviewer is sent to a specific area to interview a designated number of men and women in specified age groups. Interviewers are instructed, among other things, to skip the corner house on a block; the folks who live there may be a cut above their neighbors financially.
Basic goals are to find the issue of most concern to voters and to determine what they think about those issues. Persons interviewed are asked to name the biggest problems they believe face the country and to say what should be done about them.
Questions must be stated in neutral terms to avoid imparting a bias to the response. If a question calls for a choice between two responses, the interviewer should state both choices instead of just one. ("Do you feel that people in this area of Virginia are better or worse off than they were two or three years ago?")
The sequence of questions is important, too. Those asking about the client candidate should be scattered through the interview, rather than bunched together; that way it's harder for the respondent to guess who the client is. Mr. Hamilton says some people, once they know who the interviewer is working for, "tend to say nice things about the fellow just be nice to the lady asking the questions."
Such a poll ascertains the relative popularity of candidates and their current likely share of the electorate. But those aren't the crucial findings, Mr. Hamilton insists. Rather, he says, "the key thing is the relationships among the data and how it all can be used in the campaign. That's what the good analysts have to be able to spot."
Three or four weeks after the polling, the candidate receives a printout of the data, followed by a report of 30 or 40 pages analyzing the material and making recommendations on campaign strategy.
Mr. Hamilton won't discuss the recommendations he's making this year, but the sort of advice he gives can be seen in a 47 page poll report he wrote in 1970 for a Democratic Senator running for reelection.
That document was full of bad news for the candidate. The Senator is a liberal, it noted, but the majority of the state's voters are moderate or conservative. "Almost one-third say they have a generally unfavorable opinion of (the Senator) - a very high figure for any man holding office as prestigious as Senator."
Especially unsettling for the Senator was the news that he was regarded by "a significant minority . . . as a wheeler-dealer, and a slightly shady character. The fact that these characteristics are credible, even to a minority of voters, could prove to be highly dangerous in a closely contested election."
The report offered plenty of advice, including:
When it was all over, the Senator had won a narrow victory.
For nonincumbents, the poll findings about the popularity of the rival candidates are almost useless, Mr. Hamilton says. In such cases, the analysis focuses largely on the extent and nature of the "soft" portion of each candidates' expected vote share.
Mr. Hamilton loosely defines a "soft" voter as one who tells an interviewer he intends to vote for candidate Smith when specifically asked which man he favors, but elsewhere in the interview has something negative to say about Smith. Mr. Hamilton believes that going after these voters is one of the best ways a candidate in a close race can spend his time and money.
For an additional $1,500 or $2,000, the firm will conduct what it calls a "panelback" survey two or three months after the initial poll. As many as possible of the original respondents are resurveyed - this time by telephone and through a briefer questionnaire. The purposes are to get a later reading on the candidates' standings and to measure whether the client's strategy is working.
What a candidate does with a Hamilton poll once he's paid for it is up to him - almost. A front-runner may release it to the press, in an effort to create a bandwagon psychology. But attached to every Hamilton poll is a string:
" . . . We reserve the right - in the event any portion of this report is released to any public medium - to make public the entire report and methodology to clarify the meaning of the released portion." This reservation is aimed at cases in which a poll might show a candidate getting 30% of the vote and his opponent 15%, with 55% undecided.
"Suppose the fellow then tells a newspaper that he's just taken a poll that shows he's got a two-to-one lead," says Mr. Hamilton. "That statement is technically correct, but misleading as hell."
The firm never has had to call a candidate's bluff by releasing the complete results of a partially leaked poll, but Mr. Hamilton says he wouldn't hesitate to do so.
|
4201
Connecticut Avenue, NW, Suite 212 |