Just two days before the 1980 election, American voters remain narrowlyy divided in their choice for president and unusually volatile, according to new Washington Post polls taken after last Tuesday's debate.

The uncertainty of Tuesday's outcome is conveyed by differing findings in two new nationwide polls conducted by The Post. A national poll of 1,100 registered voters conducted Wednesday and Thursday showed President Carter leading Ronald Reagan 43 to 39, with John B. Anderson at 7 percent.

These figures are in line with a new Gallup Poll taken for Newsweek magazine and released yesterday, giving Carter a 44-to-41 lead among registered voters but Reagan a 44-to-43 edge among what Gallup calls "likely voters." Both The Post and Gallup polls have a margin of error of plus or minus 3 points.

The Post's second nationwide poll was based on reinterviews with 1,800 voters originally surveyed in September, and was designed to measure change in opinion among specific blocs of voters since then. That poll also suggested an extremely close election, but showed more voters -- including Democrats and former Anderson supporters -- moving toward Reagan than toward Carter.

Among those reinterviewed, 30 percent have changed their minds since early September about how they will vote -- either by switching candidates or becoming undecided.

The new polls indicate that 10 million Americans who say they plan to vote still have not made a choice among the candidates. Thirteen percent of both Democrats and independents and 7 percent of Republicans still say they are undecided, so the outcome of this election could well be determined by the size of the turnout. A big turnout is particularly crucial for Carter in some key states such as Texas.

Given the closeness of the race and the huge number of still-undecided voters, today's events in Iran and President Carter's reaction to them could easily be the decisive factor Tuesday.

Additional, new Washington Post polls of eight key states, all based on reinterviews with people originally polled in September, give Carter a large edge in New York, Reagan the advantage in Michigan, and show the others all virtually even: Texas, Pennsylvania, Illinois, Ohio, Florida and New Jersey. Shifts in voter opinion in these states since the first interviews have tended to favor Reagan.

The Post findings are based on the opinions of registered voters in general. Because a higher percentage of registered Republicans than of registered Democrats usually turns out to vote. The Post polls probably understate Reagan's vote slightly.

However, The Post's polls show that the remaining pool of undecided voters contains a disproportionate number of people who would naturally be expected to vote for the Democrat in a "normal" election, a factor that may favor Carter. For example, the president is still having trouble convincing traditional Democrats in many states to vote for him. In Texas, 19 percent of those who call themselves liberal Democrats are still undecided.

The Post's polls indicate that Carter still holds a considerable advantage in the South, that the Northeast and Midwest continue to be closely contested, and that the West is now closer than previous polls have suggested.

These new surveys show that although support for Anderson has been cut nearly in half during October and could easily decline further, the Anderson factor is still potentially crucial in many states, including Florida, Illinois, New Jersey and Pennsylvania.

The latest national survey shows that less than 3 in 10 Anderson supporters feel "very strongly" about their stated intention to vote for him.

The new polls also show the volatility of the electorate this year. In the eight key states a high percentage of voters -- as many as 30 percent -- changed their minds about whom to vote for since they were originally polled a month ago. In a nationwide resurvey of 1,800 voters -- 1,200 of whom were reached, 30 percent also said they had changed their voting intentions since early September.

The Post's polls produced a conflicting conclusion about the impact of last Tuesday's debate between Carter and Reagan. On one hand, of the people in The Post surveys who said they made up their minds how to vote since last Tuesday -- nearly 10 percent of those polled -- a majority now favors Reagan. However, of those who say that the debate was an important factor in their decision, the majority favors Carter.

The new polls show a striking difference between men and women voters this year. For example, in the national survey of 1,100 people who said they were registered voters, men favored Reagan 42 to 41, but women favored Carter 45 to 35. In some states, this difference is even more striking. There are also more women than men who are still undecided -- 14 vs. 10 percent in the national poll, and even greater differences in some key states.

The Post got two different results in its two national polls from two different polling procedures. One of these surveys was a straight telephone poll of 1,100 voters conducted on Wednesday and Thursday of last week, producing a 42-to-39 edge for Carter.

The second poll involved contacting participants in a telephone survey of 1,800 voters The Post questioned last September. On Wednesday and Thursday of last week, interviewers were able to reach 1,200 of those 1,800 people, and could compare their September views with what they felt after the debate. By weighting the new findings mathematically, the overall September findings could be compared with last week's results, which gave Reagan a 42-to-39 edge.

This second procedure, which pollsters call a "panel back," is considered the best way to trace changes in opinions among particular voting blocs. For example, reinterviews of Jewish voters in the eight key states showed that Carter's standing with this group rose from 26 to 37 percent during October, while Reagan moved from 25 to 29 percent.

However, some pollsters feel that the second round of interviews in a panel back survey does not produce as accurate an overall picture as does a fresh survey of a new national sample. These pollsters say there is a slight built-in bias because some of the people being reinterviewed will be embarrassed to admit changes of heart to the interviewer. rIn addition, pollsters generally believe it is easier to reach Republicans than Democrats in telephone surveys of this -- or any other -- type. The Post decided to take a second national survey last week to provide a fresh benchmark for comparisons.

A panel back procedure was also used in the new polls of eight key states. In these cases there were 600 respondents in each of the original polls, and interviewers were able to get back to 350 to 450 of them in each state.

For the national polls, the mathematical margin of error is plus or minus 3 percentage points, meaning the real sentiment in the country when the polls were taken could have been as much as 3 points higher or lower than the figures given in this article. For the state polls based on smaller samples, the margin of error is plus or minus 4 percentage points.

Because of these margins of error, it is impossible to say whether Carter or Reagan holds a lead in this extraordinarily close election. Moreover, the new Post polls suggest that voters' opionions are still in flux. For example, about half The Post's new National sample was interviewed Wednesday night, and the other half on Thursday. The Thursday sample was substantially more favorable to Carter.

Judging from this movement, it seems likely that the 1980 election, like the ones in the 1948, 1968 and 1976, for example, will be decided only in the final hours before the polls close.

The Post's new state polls neatly illustrate Carter's problems in trying to win reelection Tuesday. A rundown of some of them:

Florida. Though the Reagan camp believes it has Florida wrapped up, The Post poll suggests that the race there is still close, largely because women and black voters favor Carter so heavily. Carter loses Florida 48 to 38 among white voters in the new poll, but wins it 52 to 36 among women and by nearly 9 to 1 among blacks.

Overall the poll gives Reagan a 1-point lead, with 7 percent still undecided. Twice as many of the undecided are women as men. Anderson still has 6 percent in Florida, according to the new poll, including support from a bloc of liberal, independent voters who would be expected to go for Carter if they left Anderson.

Without a heavy turnout Tuesday, though, Carter will not get the poorer and minority voters who are crucial for him.

Illinois. Since they were polled a month ago, 18 percent of the voters in The Post's sample who had then supported Carter have moved away from him, while just 7 percent of Reagan's supporters a month ago have changed their minds.

Of thos who drifted away from Carter, half went to Reagan, a fourth to Anderson and a fourth to undecided. Of both Democrats and independents in Illinois who classify themselves as liberals, 13 percent are still supporting Anderson, a significant loss of potential Carter votes. Women are heavily for Carter in Illinois, and among remaining undecided voters women outnumber men by more than 2 to 1.

Overall The Post poll gives Reagan a statistically insignificant 2-point edge in Illinois.

Texas. There has been extraordinary volatility in Texas, according to The Post's resurvey of voters there. One out of five of the voters who said they favored Carter a month ago have moved to new categories. Of this group, two-thirds now consider themselves undecided, one-third has switched to Reagan.

Reagan supporters have been more loyal, but one in seven of them has also changed opinions. Half of that group has switched to Carter, half to undecided.

Men and women have sharply different attitudes in Texas, where Reagan leads among men by 10 percent, and Carter among women by the same margin. In Texas, as elsewhere, the preponderance of undecided voters appears to be people who would normally vote for the Democrat. The question of turnout, especially among black and brown voters, is crucial.

Pennsylvania. The most striking figure here involves registered Democrats who decline to call themselves liberals -- presumably ethnic and bluecollar voters who have made up the backbone of the Pennsylvania Democratic Party. Fully one-fourth of them, according to The Post poll, are still undecided.

Among similar Democrats in Pennsylvania who have made up their minds, Carter is favored over Reagan by 2 to 1. Whether the undecideds in this category decide to vote and split the same way could determine the outcome in Pennsylvania, where the poll now gives Reagan a 3-point edge. Overall, 15 percent of the voters remain undecided in the state, the largest undecided group in the special eight-state poll.

Michigan. This is the state which the poll shows moving more decisively toward Reagan than any other surveyed. In fact, Carter also gained marginally in Michigan over the last month, but Reagan gained more than twice as much. In early October Reagan had a 2-point lead, according to The Post poll; now he has a 7-point lead.

However, the undecideds in Michigan are still high -- 13 percent, according to the poll. Moreover, a high percentage of liberal Democrats -- 14 percent -- still say they are undecided. Both candidates still see this as a battleground stage. Reagan spent yesterday in Michigan.

In The Post's national sample, 8 percent of the respondents said they had made up their minds how to vote since last Tuesday, when Carter and Reagan debated on television. Of them, 48 percent said they had picked Reagan, 37 percent Carter, and 15 percent had chosen Anderson. This high Anderson number would suggest that one effect of the debate may have been to convince some voters that neither major candidate deserved their vote.

In the eight state polls, 10 percent of the respondents said they had made a choice since last Tuesday. Of them, 47 percent had chosen Reagan, 43 percent Carter.

However, of those questioned who said the debate had been important in helping them reach a decision, more described themselves as Carter supporters than Reagan supporters.