The speculation that projections may have caused significant numbers of registered voters to either give up their franchise or fail to vote their conscience seriously underestimates the electorate. I have more faith than that in the American voter. And it's important to keep in mind that those speculations are based on anecdotes that just don't match the hard data on voting behavior that we have examined.
A smaller percentage of the voting age population turned out in 1980, continuing a 20-year trend. Nevertheless, seven of the 10 states with the highest percentage turnout closed their polls after the time of the NBC projections.
The these figures deal with turnout of the voting age population. But since only registered voters can vote, it's even more important that we look at them. In 1980, registered voter turnout rose over 1976 in 27 of the 45 states for which we have reliable data. In the West, the turnout was up in 9 of the 13 states in the Pacific and Mountain time zones.
Even in California, which was the focus of much of the outcry against broadcast projections, 75.6 percent of its registered voters cast ballots for president in the last election. This was in keeping with turnout levels for the entire western region. In Los Angeles County, which was the source of one specific claim that NBC's early projection caused a lower turnout, the Los Angeles County registrar, Leonared Parish, concluded after a detailed study of the results that "the early network projection and the president's concession had little, if any impact on turnout."
One specific and much-publicized claim of such an impact came out of the defeat of Oregon Rep. Al Ullman. We have examined the turnout in Rep. Ullman's district for the last several elections. Perhaps as a result of the increase in the voting age population, 1980 the total number of votes cast in his congressional race actually rose by almost 50,000 votes over 1976 when there was no early presidential projection. Although Rep. Ullman lost the election by 3,765 votes, or less than 1 percent of the total votes cast, he personally drew 35,000 fewer votes in 1980 than in 1976. These facts suggest that the vote in any congressional district is very much a product of the specific local factors at work in that district in that year. They certainly do not suggest that our projection had any effect on the outcome.
We believe the American people want and expect to be told of the results of their presidential elections whenever that information is available. We believe dissemination of information through our constitutionally protected news media is critical to effective self-government in a democratic society. Access to accurate information on vote results also represents a safeguard against fraud or manipulation of supressed returns. We have developed techniques to analyze and explain Election Day events with remarkable accuracy and integrity.
In the past 20 years, NBC News has projected more than 1,300 contests, only six of them incorrectly. During that period, NBC News has never -- on a presidential general election night -- incorrectly "called" any state for president.
We are convinced that broadcast election projections do not have any meaningful impact on voter turnout or voter choice. It would be wrong to adopt solutions for what may be a phantom problem. It would be wrong to consider legislation that could do violence to the interests of the public and to the guarantees of the First Amendment.