1. The turnout rate used here is computed as: total votes/eligible voters. Eligible voters were determined by age, race, and sex determinations on a state by state basis. Some states allowed African-American men to vote long before passage of the Civil War Amendments and some states recognized the right of women to vote well before 1920. Finally, a few states allowed people younger than 21 years old to vote before 1970.

The data set used here was constructed to conduct long-term analyses of U.S. electoral behavior. Because Alaska and Hawaii were not added to the union until 1959, the data set does not include them. All references to U.S. totals should be read as reflecting totals for the continental U.S. only.

2. We felt that the forces affecting turnout would differ significantly across different types of cities both because of historical factors and more recent demographic changes. At the same time it proved challenging to differentiate cities over the period of this study. The size and relative ranking of the nation's largest cities varied over time; what are the largest center cities today were not the largest cities fifty years ago. Cities that have long been at the core of a metropolitan area have had different experiences over the past century than cities that have grown rapidly over the post WWII period.

To differentiate among these different large cities, we began by looking at populations aggregated over the twentieth century for all cities that had a population of 200,000 or more in 1990, 69 cities altogether. We aggregated both the city and the metropolitan area. Dividing these cities into two equal categories based on aggregated city population gave us a good basis for differentiating among cities. Six cities were eliminated from the tier 1 grouping because their aggregated metropolitan population did not match with the others in that group (San Jose, Birmingham, El Paso, Sacramento, Memphis, Rochester). At the same time several others were added to the top thirty because they had large metropolitan areas (Minneapolis, St. Paul, Seattle, Newark, Denver). Then the two groupings were examined to see if any “mismatches” appeared. Kansas City appeared to be the only “mismatch,” just barely missing classification in the tier 1 category. Because it appeared to have more in common with the cities classified as tier 1 it was reclassified. The remaining 37 cities were categorized as tier 2 center cities.

3. We collected actual city data for twenty of these cities, usually as far back as 1860. The exceptions were Phoenix, Oakland, San Diego, Miami, Atlanta, Buffalo, San Antonio, Dallas, Houston, and Fort Worth. In the latter cities we used county totals to represent the city.