Results from the November 1990 supplement to the Current Population Survey (CPS) show that 45 percent of persons 18 years old and over reported that they voted in the 1990 Congressional elections. This is 1 percentage point lower than in the Congressional elections of 1986.
The survey also shows that 62 percent of those 18 years and over were registered to vote at the time of the 1990 Congressional elections, two percentage points lower than in 1986.
If noncitizens are removed from the denominators and the rates recomputed, then 66 percent of citizens were reported as registered in 1990, and 48 percent were reported as having voted.
The findings in this report are based on answers to a series of questions asked of persons of voting age in the CPS conducted by the Bureau of the Census in November 1990. Data in the November CPS relate to the civilian noninstitutional population 18 years old and over.
Household surveys of registration and voting supplement official counts by providing estimates of registration and turnout for population groups not identified in official sources. The Current Population Survey is the largest of these surveys and provides the Nation's most reliable information on the social and economic characteristics of voters and nonvoters, as well as on the number and characteristics of persons registered to vote.
Voting data from household surveys, however, have the limitation of generally overestimating turnout. For example, the November 1990 CPS indicates that 82.0 million persons voted in the 1990 elections, 13.0 more than the 69.0 million who actually voted, according to the Election Research Center, using official election results from individual state boards of elections. Factors accounting for differences between official returns and CPS results are discussed in the section, "Evaluation of the Accuracy of the Data."
Because of inconsistent practices among States in purging registration records, CPS estimates of registration, although probably higher than true registration, are also probably more reliable than national estimates derived by summing official state registration records.