A poll of Philadelphia voters found mayoral candidates Sam Katz and John Street in a statistical dead heat. Another poll found John Street with a 23-percentage-point lead.

As pollsters, we tried to learn for ourselves how these polls could produce such disparate results. Our efforts were rebuffed by the candidates' pollsters.

No response to our inquiry was forthcoming from Katz's pollster, John McLauglin, and Ron Lester, Street's pollster, responded, "All information we have is confidential."

Each campaign is trying to influence the mayoral election by releasing selectively favorable polling information that inflates their candidate's standing.

Now, candidates have an absolute right to use proprietary data to guide their campaigns. Surveys provide invaluable information on which to base campaign decisions. But when candidates attempt to get their polling data to the public through the media, they become obligated to respond to inquiries regarding their polling methods.

Campaigns are compelled to inflate their candidate's prospects to improve fund-raising and generate media coverage. In short, no money, no news coverage, no campaign.

But does that make such practices acceptable? Should voters care that campaigns are stretching the truth or fiddling with survey methods to put their candidates in a favorable light? Yes. Here's why:

Manipulation has lasting effects on our political system. Willfully misleading the public alienates voters and non-voters alike, adding to increased cynicism about the political process. Cynicism is a major cause of the endemic distrust of political leaders.

Manipulation of polling data increases skepticism about polling and the contribution polls make to democracy. Surveys can be a powerful tool for providing a voice to citizens' interests. But if surveys are just another tool to twist reality, confidence in the promise of surveys will decline.

Both media and campaign polls should supply this information to the public:

A copy of the complete questionnaire.

A description of the calling procedures and calling outcomes (number of refusals, completes, ineligible households and so on).

A description of the sampling methods.

If a campaign says no to this modest proposal, journalists should refuse to report its survey results.