Map & Data Resources

Digging down into voter turnout

Map & Data Resources | ByMike Alberti,Lori Bikson | Elections

June 20, 2012 — Based on an analysis of recently-released Census Bureau data, this visualization allows you to combine multiple demographic features — race and ethnicity, sex, age, and educational attainment — to get a detailed portrait of voter turnout and registration in national elections since 2006. The viz is set up to allow you to compare one demographic composite (using the selectors at the top of the viz) with another demographic composite (using the selectors at the bottom of the viz).

In ease case, first select an election year and then construct your demographic composite by selecting race and ethnicity, sex, age, and educational attainment. For example, you might wish to view the voter turnout in the 2008 Presidential election among Hispanic men between the ages of 18 and 29 who are high school graduates (26.4 percent), or among White, non-Hispanic women between the ages of 60 and 69 with a bachelor’s degree (86.4 percent).

The filters with blue titles control the blue bars. The filters with orange titles control the orange bars. 


Why do some of the filter options disappear when I make selections?

This visualization is limited to those subgroups for which the survey sample size was large enough to result in a standard error of less than 10 percent of the voter turnout estimate. (Note: this means that, in the overwhelming number of cases, the margin of error is less than +/- 10 percentage points.) If the standard error was larger, the demographic subcategory has been excluded. Sometimes a selection along one dimension (age, for example) will mean that some options along a second dimension (educational attainment, for example) will "disappear" from the viz. To reset the filters, simply select the “All” values in each filter.

The sample size of the Census Bureau’s survey on voter turnout is larger for Presidential elections, meaning that more subcategories have a sampling error of less than 10 percent of the estimate in 2004 and 2008. The sample size was also generally larger for White, non-Hispanics than for Black, non-Hispanics or for Hispanics of any race.

 All data are from the biennial Voting and Registration Supplement of the Current Population Survey, collected by the U.S. Census Bureau. The methodology used to calculate the sampling error of voter turnout estimates is also from the Census Bureau, and is available here.



Request to republishSend a letter to the editor