Washington — The presidential race is on - and it could be extremely close. No, no, not that presidential race. The other one, between CBS, ABC, and NBC. Every big TV network, armed with whirring computers, statistical experts, and reams of data, will be racing to win the presidential sweepstakes and be the first to tell us next Tuesday evening who has won the White House for the next four years.
They will also be striving to win as many of the smaller races as possible for governorships, the US Senate, and the House of Representatives.
''By 8:30 p.m. on Tuesday, or even 8:10, we should all know who won the presidency,'' says a TV network official.
What the networks don't like to talk about is that all this electronic razzle-dazzle on election day worries some experts, and it makes some politicians fume. Many of them liked the old way of deciding who won elections - counting one vote at a time.
Democrats and Republicans alike ask questions about election projections, such as:
What if the computers are wrong?
Do the early projections of winners cause lower voter turnout?
Should the TV networks use greater self-restraint, even if that means getting defeated by their competition?
Some politicians, such as Sen. John Glenn (D) of Ohio, worry that the electoral process could be eroded by the intense competition between the networks to be the first to name the winners on election night. The senator says that if things go too far, we could see ''our sacred electoral system (take) second place to selling underarm deodorant, dog food, and beer.''
What does Senator Glenn mean by too far? Every politician seems ready to offer an example:
* Rep. Timothy E. Wirth (D) of Colorado recalls TV coverage of the Iowa caucuses in February. Two of the three TV networks, using poll samples, told the nation who had won the Democratic presidential caucuses ''even before one vote had been cast,'' Mr. Wirth says. This can be done by polling voters as they walk into the caucuses.
* Charles T. Manatt, the Democratic Party chairman, recalls an experience in 1980 when he was going door to door in North Hollywood, Calif., to get out the vote. As soon as the networks declared that Ronald Reagan had won, voter interest withered, even though California's polls were still open, he declares.
At the root of many complaints against the networks are so-called ''exit polls.'' In key locations across the country, each network this year will be interviewing as many as 40,000 people as they emerge from the voting place. Based on those voters, it is statistically possible to predict with great reliability the outcome of many elections.
With the help of exit-poll data, TV newscasters should be able to tell us who won races for the White House and Congress long before all the votes are counted.
They can also tell us much more. They can tell us which candidates did best among men and women, rich and poor, blacks and whites, young and old, Southerners and Northerners, Protestants, Roman Catholics, and Jews.
There are some signs, however, that the networks are beginning to respond to concerns about their polls and projections, especially at NBC. The new NBC News president, Lawrence K. Grossman, has decided that no projections for the presidential race in any state will be aired until local officials release some of the actual voting returns. Mr. Grossman concedes that ''this procedure may slow down our state-by-state reporting slightly'' - but probably not more than a matter of minutes.
At ABC and CBS, exit polls will be used in some cases to declare Ronald Reagan or Walter Mondale the winner of certain states - but only if the margin appears to be very great. In closer contests, ABC and CBS will await actual vote data.
Furthermore, none of the networks will declare Mr. Reagan or Mr. Mondale the winner of a state until most or all of the polls are closed in that state.
Although some politicians express unease over exit polls, there is very little hard data upon which to base a debate about their use. But political science Prof. Michael Delli Carpini of Rutgers University has done some research on the subject.
Professor Carpini told a House subcommittee that various studies have found that early projections of winners do seem to decrease turnout somewhat. The range appears to be somewhere between 2 and 12 percent. But such research still appears to be preliminary.
It also appears that well-educated voters are influenced more than others, perhaps because the better-educated ones are more likely to listen to newscasts.
A number of network executives have trudged up to Capitol Hill to talk about exit polls and projections. Among them is George Watson, vice-president of ABC News.
Mr. Watson observed that the main concern about projections was that it decreased voter turnout.
''We have seen no conclusive evidence to support the theory that the outcome of any particular election has been affected,'' Watson said. Even so, he continued, ''we share the ... larger concern that voter apathy is unhealthy for our democracy.''
The tough question for TV networks is: Can they really withhold information about who won an election, once they have that information?
''Practically speaking, suppressing information does not work in our free and open society,'' Watson said. ''Certainly, every political organization and both major parties will know what is happening. Given the inevitable flow of information, accurate reporting by news organizations is preferable to information spread by rumor.''
One problem often mentioned is that closing time in the East generally comes before closing time in the West. If the presidential election is a landslide, the winner could become very apparent long before polls close out West.
Even if Western voters cannot change the outcome of a presidential race, they are still electing all of their US House members, a third of their US senators, and large numbers of state and local officeholders. Early projections may reduce turnout for those local races - and may even change the outcomes.
One idea that's been getting favorable attention would be to close the polls at the same time in every state. Polls could close at 11 p.m. Eastern time, 10 p.m. Central time, 9 p.m. Mountain time, and 8 p.m. Pacific time, for example. That would avert situations where voters learn that Candidate X has ''won'' even before California finished voting.
A number of congressmen who have studied exit polling say it would be unwise - or even impossible - to put legal restraints on the networks because of the First Amendment, which guarantees free speech and press.
Both houses of Congress, however, have passed a resolution asking that ''broadcasters and other members of the news media should voluntarily refrain from characterizing or projecting results of an election before all polls for the office have closed.''
Congressman Wirth says that some newscasters do promise not to project winners before the polls close, but then duck under that ban by ''characterizing'' the race.
''We (might be told) that 'if this trend continues, it could be a very big night for Mondale,' for example. That's just like a projection,'' Wirth suggests.