If President Carter feels like he's riding a public opinion roller coaster, he has good cause: The Macy's and Gimbel's of national opinion surveys can't seem to agree on anything.

The latest Gallup poll, released yesterday, claims that 6 persons in every 10 (59 per cent) express approval of the way Carter is handling his job. Only 24 per cent disapprove.

That, presumably, will come as good news in the White House.

But wait, the ink is hardly dry on a Louis Harris survey of just a week ago, which had the President's approval rating skidding to 48 per cent, with an equal percentage of Americans disapproving of the way Carter is leading the nation.

Who is a President to believe?

While the President stands higher in the Gallup poll than Harris', his October Gallup rating is 7 percentage points below his September approval level, and 6 points below his average monthly popularity rating since he took office.

The "approval" and "disapproval" lines on the Gallup charts seem headed toward a convergence, as they have already done on the Harris graphs.

This convergence is what the experts at Gallup's Princeton, N.J., offices and Harris' New York City shop point to when they try to explain the disparity between the two polls.

"The trend is the thing. The timing and the asking of slightly different questions may have some effect, but they should't affect the trend," said Humphrey Taylor, Harris' deputy in New York.

Says George Gallup Jr., "The trend is all important. As long as the measurement is consistent, it doesn't matter which questions you ask."

That would explain the consistency within each of the two polling organizations, but not the disparity between them.

Opinion survey experts emphasize tht polling is more of an art than a science, and that enormous variations in results can be attributed to question wording, timing and the context in which a question is asked.

Gallup and Harris conducted their surveys at about the same time - the end of September - and each interviewed approximately 1,500 adults from scientifically selected localities.

But the two organizations have a long tradition of differences.

In the early days of the Nixon administration and during Gerald Ford's presidency, Gallup tended toward higher approval ratins than Harris.

For example, in February, 1975, Gallup asked, "Do you approve or disapprove of the way Ford is handling his job as President." Fifty-five per cent approved and 28 per cent disapproved.

But when Harris posed his standard question, asking subjects to rate the President's performance as excellent, pretty good, only fair or poor, 46 responded with ratings Harris interpreted as positive, and 52 per cent gave a negative rating.

While deliberately trying to avoid saying whether he approves or disapproves of fellow pollster Harris, Gallup suggested it may be easier for an interview subject to move from "good" to "only fair" than it would be to move from approve to disapprove.

Also, other opinion analysts pointed out the Harris organization lumps "only fair" and "poor" together in a negative category, despite the fact that some respondents may not intend "only fair" to be distinctly negative.

All of which goes to prove that the people know best, whatever that is.