For each combination of demographic characteristics — each gender, age, race, education level, and/or weekly earnings level — the number of cases that answered “Yes” to the question of whether they had voted, was divided by all other cases except those listed as “Not in Universe.” According to the Census Bureau, “Not in Universe” generally refers to non-citizens and to American military personnel, and this is the method the Census Bureau uses to report turnout from the Current Population Survey.
These voting percentages numbers are approximate at best, and they occasionally reflect small sample sizes. Some combinations of the five characteristics are not reported because the subsample of respondents who fit those characteristics is too small; no voting percentages for particular subsamples are reported here when there were fewer than 100 respondents in that subsample.
Also, note that the CPS data was collected across four very different elections and political climates. Moreover, because the responses on voting are entirely self-reported, it is possible that the percentages will be higher than reality. People often respond to surveys with what they perceive as the societally desirable response. Proving the point, other official sources show the average voting age turnout from 2004-2010 was 46%, compared to a 55% turnout rate reported by CPS respondents.
More information about CPS methodology can be found at technical documentation for the CPS. Other sources of information about voter turnout can be found at the Statistical Abstract of the United States.
This interactive graphic shows the percentage of voting-age American citizens in various age, sex, race, education levels and earnings levels who say they voted in a recent election. These percentages come from a News21 analysis of data collected from 378,949 respondents by the Census Bureau's Current Population Survey in November of 2004, 2006, 2008 and 2010. MORE