Medical Device Daily National Editor

Pre-election polling is one of the more inexact sciences around, and a new poll from Harris Interactive (Rochester, New York) perhaps underlines the highly "informed guessing" quality of this activity.

With that qualification, Harris's new poll indicates three things:

• That U.S. voters whose main concerns are national security and international issues may favor Sen. John McCain as the next U.S. president ...

• and that those who are more concerned about specific domestic issues, including healthcare and the environment, are more likely to favor Sen. Barack Obama ...

• or neither of the above holds true, since issue concerns don't always determine a voter's final choice, according to Harris.

These are the main conclusions, or non-conclusions, of the Harris poll that assesses the two presidential candidates' strengths – or at least "public perception" of them — on 16 issues.

And a poll by Harris, issued last week (Sept. 25), might suggest that domestic issues (currently most in the news) could drive a final choice. That poll gave the winning vote very narrowly to Obama over McCain, 47% to 46%, among "likely" voters (if held at that time).

But Harris candidly acknowledges that issue orientation may not be a decider once a person closes the voting booth curtain.

Harris says that its most recent poll reveals that a majority of those surveyed said they weren't very familiar with the candidates' differences on the issues. And it adds what we know from other polling: that voters often cast their ballots based on factors other than issues.

Though the new poll is timely, taken between Sept. 15 and Sept. 22 (online by more than 2,300 people, identified as "adults" but not as voters or likely to vote), it may not reflect any possible changes of opinion following the most recent uncertainties of the U.S. economy, again highlighting domestic concerns.

So here are the specific polling figures on issue preferences:

On healthcare, Sen. Obama is seen as "the better of the two candidates" by 16 points, and on the environment by an even greater lead, 22 points.

(Will recent fiscal uncertainties change this? McCain is promising a $5,000 tax credit for buying health insurance in the private sector, more of a "cash in the pocket" offer than Obama's promise of "universal" healthcare coverage.)

Other "point" leads for Obama are for education (+19), jobs and employment (+12), the economy (+10), gas prices (+8), energy policy (+7), and inflation (+5).

McCain was seen as the better of the two candidates in being able to handle defense, homeland security and keeping the U.S. safe from terrorism (by 18 points), Afghanistan (+9), the Middle East (+8), Russia (+8), Iran (+7), gun control (+5) and Iraq (+5).

Neither candidate has a significant advantage on trade or taxes, according to these poll results.

Perhaps most interesting of all, however, Harris says that "only between 23% and 36%" of those polled said that they understand the candidates' policy proposals "very well." And this figure probably over-states the actual certainty, since people polled generally are judged as providing an answer that they believe is more acceptable to the pollster.

Harris provides a variety of other qualifying statements suggesting that its poll doesn't offer a pristine crystal ball – but is something more like the reading of tea leaves or sheep – for predicting a November "winner."

It says that whether or not people are actually well-informed about a candidate's issues, they tend to have "strong feelings" about the ability of one candidate over another to handle an issue. And Harris says that voter choices are "both conscious and unconscious" and so not really made on the issues.

Often, according to Harris, "pollees" who say that their votes are issue-based are "rationalizing their preferences."

Then again, maybe Harris should tell us something we don't know.

What we know is early predictions of a close race, along with rather quick changes in polling trends.

The Sept. 25 Harris Poll said the election would be too close to call ("neck and neck with only six weeks to go," if held at that time, it said), giving a one percentage point lead to Obama over McCain, 47% to 46%.

But others tell us something different: In reporting its survey results at that time Harris noted an ABC/Washington Post survey, of about the same period, giving Obama a 9-point lead over McCain.

Exactly who is polled and who votes are other questions.

Harris said that the poll issued Sept. 25, in comparison to its earlier survey focusing on registered voters (which gave Obama a 2% lead over McCain), looked at "registered voters [defined] as registered voters who are highly motivated" to vote (1,562).

It said: "In previous elections there were sometimes substantial differences between the voting intentions of likely voters and registered voters, but in this poll there is no significant difference, except that likely voters are less likely to say they are not sure."

(Finally, as I write this, CNN has just issued its polling results: 49% for Obama, 44% for McCain, 7% undecided.)

Still another shift may, or may not, come this evening, with the debate between vice presidential candidates which may reinforce what Harris called the "Palin bounce," or prick the "Palin bubble" – or add further uncertainty.