U.S. Political Stats Methodology
Below you will find a detailed overview of the methodology behind data in U.S. Political Stats. To download only the source list, please click here.
- Biographical Characteristics
- Campaign Finances
- Economic Indicators
- Floor Votes
- Interest Group Scores
- Presidential Performance Indicators
- Supreme Court Cases
- Voting Scores
Congress Member Biographical Variables
U.S. Political Stats can be used effectively to ascertain the makeup of both chambers of the U.S. Congress based on a number of factors. Changes in the makeup of the U.S. House, for example, by military service can be achieved easily. By comparing a biographical characteristic with a floor vote, users can use this product to compare how members of the U.S. House voted on a veterans benefit bill based on their military service (or not). The variables listed here for members of Congress represent the total list of options for selection currently in the database. Some of these variables may not have been used for members of Congress from 1983–present.
- Highest Educational Attainment
- Military Service
- Party Affiliation
- Religious Affiliation
Highest Educational Attainment
Based on degree attainment and other educational information, the following designations were made. Individuals who studied for, but did not attain a degree will be classified based on whatever lower degree (if any) they attained. In some cases, the information available on a particular individual was insufficient to make a designation and are therefore classified as “Unknown”.
- High School Graduate
- Associate's Degree (AA, AS)
- Bachelor's Degree (BA, BS)
- Master's Degree (MA, MS, MEng, MEd, MSW, MBA)
- Professional Degree (MD, DDS, DVM, JD, LLB)
- Doctorate Degree (PhD, EdD, or other academic doctorate)
- African American
- Asian Pacific American
- Native American
- Served/Did Not Serve
- Business or banking
- Congressional aide
- Construction/building trades
- Labor leader
- Law enforcement
- Public service/ politics
- Real estate
- African Methodist Episcopalian
- Apostolic Christian
- Assembly of God
- Associated Reform Presbyterian
- Central Schwenkfelder
- Christian [Note: This is a self-designation and is not an aggregation of all Christian denominations]
- Christian Missionary Alliance
- Christian Reformed
- Christian Scientist
- Church of Christ
- Church of God
- Disciples of Christ
- Dutch Reformed
- Eastern Catholic
- Eastern Orthodox
- Eastern Orthodox Christian
- Evangelical Lutheran
- French Huguenot
- Greek Orthodox
- Independent Bible Church
- Independent Christian
- Not specified
- Reformed Church
- Reorganized Church of Jesus Christ of Latter-Day Saints
- Roman Catholic
- Serbian Orthodox
- Seventh-Day Adventist
- Society of Friends
- Southern Baptist
- United Church of Christ
- United Methodist
Characteristics by Congress Number
It is important to remember when looking at data by Congress (e.g. 110th, 111th, etc.) that the data represents every member of Congress that served during that two-year period. Therefore, when looking at the House or Senate one will likely see more than 435 or 100 members respectively for a given term. This occurs when a member leaves before a Congress ends and another member replaces them by special election or appointment.
CQ Press’s American Political Leaders, 1789–2005 and Guide to Congress, 6th ed., and CQ Weekly are used to confirm members’ sex, race, position held and religious affiliation as well as dates of birth and death. CQ Press uses multiple sources to confirm this information, including obituaries and excerpts from major newspapers, as sources do not always agree on precise dates. Biographical Directory of the American Congress 1774–1996 is another reliable source for confirming these dates and other biographical characteristics, as well as lists of congressional rosters and state delegations. This directory will also list any external biographies covering a given member if applicable.
Presidential/Vice Presidential Biographical Data
Presidential biographical characteristics and variables mirror those of the members of Congress. Because the database includes all presidents and vice presidents going back to Washington, the political party variables are more numerous than those for members of Congress that date back to 1983. For vice presidents, the following biographical variables are available: gender, party, highest educational attainment, and military service.
Presidential and vice presidential biographical data was originally published in CQ Press’s Guide to the Presidency and the Executive Branch, 5th ed., Vital Statistics on the Presidency, 4th ed., American Political Leaders 1789–2000. Other sources consulted for biographical data include: The Presidency of the United States: A Student Companion, 2nd ed. (New York: Oxford University Press, 2001), Biographical Directory of the United States, the websites of the Office of the Clerk: U.S. House of Representatives, and the U.S. Senate Historical Office.
The political party designations and congressional district numbers of presidents and vice presidents who served in Congress (to Richard Nixon and John Garner respectively) is derived from:
- Martis, Kenneth C. The Historical Atlas of United States Congressional Districts, 1789–1983. New York: Free Press, 1982.
- Martis, Kenneth C. The Historical Atlas of Political Parties in the United States Congress, 1789–1989. New York: Macmillan, 1989.
Supreme Court Justice Biographical Data
Biographical data on the U.S. Supreme Court justices focuses primarily on their legal training and career; however some of the variables overlap with members of Congress and presidents. Variables for Supreme Court justices include:
- Gender (see above for variables)
- Religious Affiliation
- Military Service (see above for variables)
- Childhood Surrounding
- Family Status
- Ethnicity/National Background
- Undergraduate Status
- Law School Status
- Read the Law
- Private Practice
- Teaching Experience
- Disciples of Christ
- Dutch Reform
- Roman Catholic
- Family farm
- Family plantation
- Not specified
- Small city
- Small town
- Urban; family ranch
- Not specified
- Hispanic/Puerto Rican
- Did not attend
Graduate School Status
- Did Not Attend
- Not Specified
Law School Status
- Did not attend
Read the Law
It was not always the norm for lawyers to obtain law degrees from accredited universities; law schools were few and far between in early America. Until the twentieth century, lawyers typically "read the law" by teaching it to themselves alone or with the help of one or more mentors. This method of entering the legal profession diminished in the early twentieth century, and the last Supreme Court justice to have "read the law" was appointed in 1941.
This variable indicates whether or not the justice, during the course of their legal career, worked in a private legal practice.
This variable indicates whether or not the justice during the course of their legal career had a legal teaching position.
Although originally compiled from many varied sources, the Supreme Court justice biographical data is sourced to: Epstein, Lee, Jeffrey A. Segal, Harold J. Spaeth, and Thomas G. Walker. The Supreme Court Compendium: Data, Decisions, and Developments. 5th ed. Washington: CQ Press, 2012.
Candidates running in federal elections must disclose their campaign receipts and disbursements to the Federal Election Commission (FEC). During the election cycle, candidates are required to submit quarterly reports to the FEC providing details on a number of fundraising and spending particulars. U.S. Political Stats offers what CQ Press editors believe to be the most useful of those disclosure categories for each candidate represented here. It is inevitable that there will be inconsistencies and errors in the FEC's data, given the size and complexity of the FEC's reporting mandate and the tendency for human error in reporting. Presently, this database contains campaign finance totals for the 2008–2012 election cycles.
U.S. Political Stats contains FEC disclosure data for the following election cycles:
Primary and General Campaign Funds
For each election cycle, a candidate's primary and general election fundraising and spending are combined (assuming that they are for the same office).
U.S. Political Stats includes spending and fundraising totals as well as select component parts that editors thought would be most insightful. These variables include:
- Total Receipts
- Total Disbursement
- Contributions from Individuals
- Party Contributions
- Other Contributions
- Cash on Hand
Those interested in seeing all component parts of a candidate’s financial disclosure report to the FEC set for a candidate should visit the FEC's website and access their “Candidate Summary” file: http://www.fec.gov/data/index.jsp.
Multiple Races in a Single Election Cycle
There are times when a candidate might run for two offices in a given election cycle. For instance, a member of the House who is running for their House seat might also run in a presidential primary. In that case, the Representative would report campaign finance data for both office elections to the FEC. We are currently working on enabling U.S. Political Stats to accommodate such situations.Can funds reported in one election cycle actually have been raised/spent in a prior election cycle?
Yes. For example, it is possible that money raised/spent in the 2009–2010 election cycle could be reported by a candidate in the 2011–2012 cycle.
Population figures and other data characterizing the people of a given jurisdiction can provide a useful view of the electorate, particularly when combined with other data such as election results or roll call votes. Knowing, for example, that a House member represents an impoverished district might shed light on why he or she votes a certain way. The decennial U.S. Census provides data for a number of geographic boundaries, but users will find demographic data only at the state, county, and congressional district level in U.S. Political Stats.
Because the U.S. Census is essential to the reapportionment and redistricting process for the decade that follows, the Census Bureau has to wait until states approve new district boundaries before they can apply the Census data (now several years old) to those new boundaries.
Over the decades, the U.S. Census has changed the questions it asks, the way it asks them, and the way it reports data. Consequently, it is not always easy or accurate to compare demographics across time.
While some figures, such as total population or population by age group, remain fairly consistent over time, the terminology used in ascertaining race or ethnic background have changed over time. One important change was the transition away from the long form used through the 2000 Census and the subsequent adoption of the American Community Survey (ACS). The ACS, like the long form, gathers social and economic information about the U.S. populace. However, the ACS collects data on a rolling basis on a smaller sample set and releases its findings annually. The decennial Census of 2010 continued to collect data on all persons living in the United States, but it is now complemented by the ACS, not the long form, for the collection of social and economic data.
Selected Methodology and Sources: Congressional Districts
Population figures and other demographic data for congressional districts created after the 1980 U.S. Census were published in Congressional Districts in the 1980s, published by Congressional Quarterly, Inc. in 1983. Data is sourced from: Bureau of the Census, 1980 Summary Tape Files 1D and 3D, National Summaries. All data was either taken directly from the decennial 1980 Census or calculated from that Census data by Congressional Quarterly.Age of Population
Voting age population includes all persons 18 and over, that is, all potential voters, not just those who are registered.Occupation
Percentages of white-collar, blue-collar, service and farm workers were calculated by Congressional Quarterly from figures provided by the Census Bureau. The Census Bureau divides employed persons into 12 major categories that Congressional Quarterly has regrouped into four categories as follows:
- White collar: managerial and professional specialty occupations including executive, administrative and managerial occupations and professional specialty occupations; and technical, sales and administrative support occupations including technicians and related support occupations, sales occupations and administrative support occupations including clerical.
- Blue collar: precision production, craft and repair occupations; and operators, fabricators, and laborers including machine operators, assemblers and inspectors; transportation and material-moving occupations; and handlers, equipment cleaners, helpers and laborers.
- Service workers: service occupations, including private household occupations and protective service occupations, and service occupations except protective and household occupations.
- Farm workers: farming, forestry and fishing occupations.
Data on years of school completed refer to the adult population 25 years of age or over. Percentages were calculated by Congressional Quarterly.Housing and Residential Patterns
Housing units are broken down into owner-occupied and renter-occupied units. Housing units include both individual houses and individual apartments.
People classified as urban by the Census Bureau were those persons living in urbanized areas and in places of 2,500 or more outside urbanized areas. People classified as rural were all persons not classified as urban. The Census Bureau defines an urbanized area as one in which at least 50,000 people are concentrated, usually consisting of a central city and the surrounding, closely settled, contiguous territory generally referred to as suburbs.Race and Ancestry
The Census Bureau established five main racial categories: white; black; American Indian, Eskimo and Aleut; Asian and Pacific Islander; and other. All people responding to the 1980 Census were asked to identify which race they were. Census then grouped as white all those persons who responded “white” or who wrote in answers such as Canadian or Polish or Lebanese. It classified as black all those who said they were black or Negro or who reported such answers as Jamaican, Nigerian and black Puerto Rican.
Persons in the American Indian, Eskimo and Aleut category were those who said they were a member of one of those specific groups as well as those who reported a specific Indian tribe or said they were French American Indians or the like.
Census classified as Asian and Pacific Islander all those who responded they were Chinese, Filipino, Japanese, Asian Indian, Korean, Vietnamese, Hawaiian, Samoan or Guamanian. In Census’s full-count tabulations — that is, in the data in which every respondent was counted — persons who listed their race as Cambodian, Laotian, Pakistani and Fiji Islander were placed in the “other” category. In its sample-count data — that is, that data in which the total for a given category was extrapolated from a sample count — persons who listed these four races were classified as Asian and Pacific Islanders. Congressional Quarterly has listed the percentages based on the full-count figures for white; black; and American Indian, Eskimo and Aleut; and on the sample-count data for Asian and Pacific Islanders. In some instances the sizable “other” category in the racial breakdown may be partially explained by the way in which the Census Bureau treated those persons of Spanish origin who listed their race as “other.” Census reported that more people of Spanish origin marked the “other” category in 1980 than did in 1970. Moreover, in 1970 the Census Bureau reclassified as “white” those persons of Spanish origin who marked the “other” column, but who also gave themselves a Spanish designation such as Mexican or Venezuelan. In 1980 such persons were left in the “other” category.
Persons of Spanish origin or descent were asked to identify themselves by responding to a specific question in the 1980 Census. Persons of Spanish origin can be of any race. Census provided figures for persons of Spanish origin from both its full-count and sample-count data. Congressional Quarterly has listed percentages based on the full-count figures.
Persons responding to the 1980 Census sample questionnaire were asked to identify their ancestry, which Census defined as “a person’s nationality group, lineage or the country in which the person or the person’s parents or ancestors were born before their arrival in the United States.” Census noted that the question reflected only a person’s identification with an ethnic group and “not necessarily the degree of attachment or association the person had” with that particular group. Congressional Quarterly has included all such ancestry groups that make up at least 0.5 percent of each district’s population.
Population figures and other demographic data for congressional districts created after the 1990 U.S. Census were published in Congressional Districts in the 1990s: A Portrait of America, published by Congressional Quarterly Inc. in 1993. All data was either taken directly from the decennial 1990 Census or calculated from 1990 Census data by Congressional Quarterly.Age of Population
Voting age population includes all persons 18 and over, that is, all potential voters, not just those who are registered.Occupation
Percentages of white-collar, blue-collar, service, and farm worker are in terms of employed persons age 16 and older and were calculated by Congressional Quarterly from census data. The Census Bureau divides employed persons into thirteen major industry groups that Congressional Quarterly lists in four categories:
- White-Collar Workers. Managerial and professional specialty occupations (including executive, administrative and managerial occupations and professional specialty occupations); and technical, sales and administrative support occupations (including technicians and related support occupations, sales occupations and administrative support occupations including clerical).
- Blue-Collar Workers. Precision production, craft and repair occupations; and operators, fabricators and laborers (including machine operators; assemblers and inspectors; transportation and material-moving occupations; and handlers, equipment cleaners, helpers and laborers).
- Service Workers. Service occupations (including private household occupations and protective service occupations, and service occupations except protective and household occupations).
- Farms Workers. Farming, forestry and fishing occupations.
Data on years of school completed refer to the adult population 25 years of age or older. The four education categories (e.g. “Less than grade 9”) represent discrete groups of individuals. Figures for high school graduates, for example, refer to persons with only a high school diploma or the equivalent. Percentages were calculated by Congressional Quarterly from census data.Housing and Residential Patterns
Housing units are divided between owner-occupied and renter-occupied units. Housing units include both individual houses and individual apartments.
People classified as urban by the Census Bureau were those persons living in urbanized areas and in places of 2,500 or more outside urbanized areas. The Census Bureau defines an urbanized area as one in which at least 50,000 people are concentrated, usually consisting of a central city and the surrounding, closely settled, contiguous territory often referred to as suburbs. People classified as rural were all persons not classified as urban.Race and Ancestry
The Census Bureau established these five main racial categories: white; black; American Indian, Eskimo and Aleut; Asian and Pacific Islander; and Other. All people responding to the 1990 census were asked this question based on self-identification. In reporting the data, the bureau noted the self-identification “does not denote any clear-cut scientific definition of biological stock. The data for race represent self-classification by people according to the race with which they most closely identify.”
Census included as “white” those persons who indicated their race as “White” or reported entries such as Canadian, German, Italian, Lebanese, Near Easterner, Arab or Polish. It classified as “black” all persons who indicated their race as “Black or Negro” or reported entries such as African American, Afro-American, Black Puerto Rican, Jamaican, Nigerian, West Indian, or Haitian. Persons in the American Indian, Eskimo and Aleut category were those who said they were a member of one of those specific groups as well as those who reported a specific Indian tribe or such entries as Canadian Indian, French American Indian or Spanish American Indian.
Census classified as Asian or Pacific Islander all persons who identified themselves as Asian (including Chinese, Filipino, Japanese, Asian Indian, Korean, and Vietnamese or provided a write-in identification such as Cambodian, Laotian, or Thai), or Pacific Islander groups (including Hawaiian, Samoan, Guamanian, or provided a write-in response such as Tahitian, Polynesian, or Micronesian).
A category of “Other Race” was used for all persons not in the other groups. Persons reported here include write-in entries such as multiracial, multi-ethnic, mixed, interracial, Wesort or a Spanish/Hispanic group (such as Mexican, Cuban, or Puerto Rican).
Data on ancestry was obtained from a sample of persons in the 1990 census, based on self-identification. The Census Bureau said that ancestry “refers to a person’s ethnic origin or descent, ‘roots,’ or heritage or the place of birth of the person or the person’s parents or ancestry before their arrival in the United States.” The bureau noted the “ancestry question allowed respondents to report one or more ancestry groups. While a large number of respondents listed a single ancestry, the majority of answers included more than one ethnic entry.”
The census accepted “American” as a unique ethnicity if it was listed alone, with an ambiguous response, or with the name of a state. If a respondent listed any other ethnic identify, such as “Italian American,” generally the bureau did not include the person under “American.”
Population figures and other demographic data for congressional districts created after the 2000 U.S. Census were published in Congressional Districts in the 2000s: A Portrait of America, published by Congressional Quarterly Inc. in 1993. All data was either taken directly from the decennial 1990 Census or calculated from 1990 Census data by Congressional Quarterly.Population
The voting-age population shows all potential voters, including noncitizens and others not eligible to register, not just persons registered to vote.Occupations
Percentages of white-collar, blue-collar, service, and farm worker occupations are in terms of employed persons age 16 and older and were calculated from census data. The Census Bureau revised the classification of the occupational questions for 2000 to coincide with the standard occupational classification manual. The Census Bureau divides employed persons into 13 major industry groups that Congressional Districts in the 2000s combines into four categories in the table that are only generally comparable with previous combinations:
- White-Collar Workers. Managerial and professional specialty occupations (including executive, administrative, and managerial occupations, and professional specialty occupations) and technical, sales, and administrative support occupations (including technicians and related support occupations, sales occupations, and administrative support occupations including clerical).
- Blue-Collar Workers. Precision production, craft, and repair occupations; and operators, fabricators, and laborers (including machine operators; assemblers and inspectors; transportation and material-moving occupations; and handlers, equipment cleaners, helpers, and laborers).
- Service Workers. Service occupations (including private household occupations and protective service occupations, and service occupations except protective and household occupations).
- Farm Workers. Farming, forestry, and fishing occupations.
Data on years of school completed refer to the adult population 25 years of age or older. The data that appear in the columns of the table are for discrete groups of individuals. Percentages were calculated from census data.Housing and Residential Patterns
Housing units are divided between owner-occupied and renter-occupied units. Housing units include individual houses and individual apartments. People classified as urban by the Census Bureau were those persons living in urban clusters or urbanized areas, both census definitions. An urban cluster was a densely settled territory of at least 2,500 people but fewer than 50,000. An urbanized area was a location with at least one central place and adjoining territory with a population density of at least 1,000 people per square mile and a minimum residential population of at least 50,000 people. People classified as rural were all persons not classified as urban.
In explaining these groupings, the Census Bureau cautioned that the criteria for an urban area were extensively revised for the 2000 census. As a result, some territory that was classified as urbanized for the 1990 census was reclassified as rural, and some areas identified as urban areas in the 1990 count were reclassified as urban clusters in 2000.Race and Ancestry
Race. The racial categories that appear in Congressional Districts in the 2000s are taken from the Census Bureau’s listings and use the Bureau’s labeling: white, black or African American, American Indian or Alaska native, Asian, native Hawaiian or other Pacific Islander, some other race, and two or more races. Persons responding to the 2000 census were asked to choose the race or races with which they most closely identify. In reporting the data, the bureau noted that the categories are “socio-political constructs and should not be interpreted as being scientific or anthropological in nature.” The bureau also said the race categories included both racial and national-origin groups. People who identify their origin as Hispanic or Latino may be of any race.
The categories used in the 2000 census incorporated new guidelines that the U.S. Office of Management and Budget (OMB) recently created. The OMB said the revised guidelines reflected “the increasing diversity of our nation’s population, stemming from growth in interracial marriages and immigration.” One important change allowed respondents to report as many race categories as they wished. As a result of this and other definitional changes, the Census Bureau said the 2000 race data are not directly comparable with data from the 1990 and previous censuses.
Ancestry. Unlike the data in most tables in Congressional Districts in the 2000s, ancestry data are subject to different methodological analysis. This is because the ancestry question asked during the census was open-ended and allowed for multiple responses. A person did not have to indicate any ancestry or could have stated one or several. The Census Bureau tabulated the first two responses for purposes of reporting. Analysts can examine these data in different ways using different population universes such as total population, ancestry specified, and total ancestry responses, each of which will give somewhat different pictures of the ancestry of persons living in a congressional district area.
Clark Bensen of Polidata prepared the estimates for Congressional Districts in the 2000s. For purposes of this book, Polidata used the average number of responses to the ancestry question for any given district to estimate the number of persons who shared this characteristic. This number was then divided by the total population to attain an estimated percentage of the population that is comparable to other percentages. This method disregards any priority to the first response tabulated and does not count persons more than once. It is, however, just an estimate and is subject to error, and it may be affected by differential rates of response to the ancestry question.
Population figures and other demographic data for congressional districts created after the 2010 U.S. Census was collected and adapted from the U.S. decennial census dataset as well as data from the 2012 1-year American Community Survey estimates. Individual table/visualization source lines will indicate the exact source. Some percentages were calculated by CQ Press and did not originate with the Census data.Mid-Decade Redistricting
Census data in U.S. Political Stats is based on the first set of district boundaries drawn for each state following decennial redistricting. Data for new boundaries drawn mid-decade (due to legal challenges or for political reasons) are not yet available in U.S. Political Stats.
U.S. Political Stats contains a wide variety of statistics about the state of the U.S. economy, employment, poverty, and government spending among many other categories. While they are useful on their own they can also give context to other data found in U.S. Political Stats, such as floor votes. As the data is highly varied so are the sources and the methodology behind them. For more details, consult the source and notes sections of individual data sets.
Tracking elections across time and geography is easy with U.S. Political Stats. Users will find a wealth of election returns for president, U.S. Senate and House, and governor. These returns are available based on office and time at the state, county, and congressional district levels. Primaries and special elections are also easy to locate. (Although, remember to select a party in the drop-down menu on the elections page when accessing primaries). Use national maps to discover trends and to compare the changing landscapes of congressional districts. A specific candidate can be selected for viewing his or her full election record and the compare feature allows two candidates’ election returns to be viewed simultaneously over time. The following table gives the years and geographies included on this site for each office and election type. These data can be accessed primarily through the elections hub and any individual’s page.
The electoral landscape in the United States is diverse in its laws and there are many peculiarities across time, office, and geography. During the beta launch period and beyond, editors will work to iron out remaining inconsistencies and bugs in the visualizations and tables.
|Office||Election Type||Years||Smallest Geography|
|President||General (Electoral)||1984 to 2012||State Level|
|President||General (Popular)||1984 to 2012||County Level|
|President||Primary||1984 to 2012||State Level|
|Governor||General||1982 to 2012||County Level|
|Governor||Primary||1982 to 2012||State Level|
|House||General||1982 to 2012||Congressional District|
|House||Primary||1994 to 2012||Congressional District|
|Senate||General||1982 to 2012||County Level|
|Senate||Primary||1982 to 2012||State Level|
"Other" Votes and Vote Percentage Calculation
Election returns displayed in this product reflect the dominance of the Republican and Democratic parties in the twentieth and twenty-first centuries. In general elections, non-Democrat and non-Republican candidate vote totals are usually grouped together in the “other” category. The term “other” was chosen because the candidates included in this category represent a variety of affiliations or are running independent of a party. Therefore, the widely-used term “third party” does not always apply.
Any candidate who is not affiliated with the Republican or Democratic Party is included in the "Other" section of the vote totals. This includes Independent, third party, write-in candidates, as well as non-leading major party candidates. This database does not include the names of every non-Democrat and non-Republican who has ever run in a federal or gubernatorial election. Consequently, users will often simply find an aggregate vote total for all of those candidates. However, in cases where the database does include a name and, in particular, when one of those candidates wins a race, his or her name will appear in the "Highest Other" column. When an "other" candidate's name is shown, the vote totals are shown for that candidate only, not for all of the "other" candidates.
The American electorate can get a little complicated and in states such as California, where members of the same party can run against each other in the general election, U.S. Political Stats will use that "Highest Other" to show a second major party candidate's results. This is fairly rare so it will not appear very often.
Candidate vote percentages are always calculated based on the total number of votes cast.
A special election is a race or election held outside the normal election cycle to address special circumstances such as the death or resignation of an officeholder or a recall of a particular elected official. When browsing special election results by map, users will see only the elections that occurred in that year. Years without special elections for a given office will not be displayed.
Candidate Election Histories
Users will find general election histories for candidates in U.S. Political Stats that include their vote percentage in any gubernatorial, congressional, or presidential race from 1982 to the present. Use caution when comparing House elections over time as districts can change significantly from one decade to the next (even if they have the same number).
Open Primaries and Louisiana
For the purposes of U.S. Political Stats, an "open" primary refers to a primary open to candidates and voters of all parties. The top two vote-getters from the open primary proceed to a runoff election, except in some cases in Louisiana (see below). This usage differs from the oft-heard description of an open primary as allowing voters of all parties, but candidates from only one party.
Since 1975, Louisiana has used a unique two-tier election system for House, Senate, and governor. Its use was suspended for the 2008 and 2010 elections, but reinstated for the 2011 and 2012 elections. A "jungle primary" is held for all candidates of any party, and any candidate that receives 50% of the vote is elected to office. If no candidate receives 50% of the vote, the top two vote-getters enter a runoff election for office. Any "jungle primary" in which a candidate was elected is treated in this database as a general election, because it was a decisive election. Any "jungle primary" that was not decisive is treated as an open primary for the purposes of the Collection, and the runoff election is then treated as a general election.
Winning Candidate Party Abbreviations
The following list is alphabetical, but includes the most common parties at the top. This list includes only party designations (and the label of Independent) of candidates who won in a primary or general election.
|A Connecticut Party||ACP|
|American Independent Party||AIP|
|Peace and Freedom||PFP|
|Right to Life||RL|
|Socialist Workers Party||SW|
It is important to remember when looking at election returns that boundaries change over time. This is most apparent in congressional districts. With each decennial census, not only can congressional district boundaries change dramatically, but the number of districts apportioned to a state can change as well. For example, after the 1990 census, California went from 45 representatives in the House to 52. At the same time Michigan lost two seats in the House. Consequently, looking at districts over time can be difficult because they no longer exist or they may be redrawn to represent a completely different part of the state. To complicate things further, some districts have been redrawn in between decennial censuses due to court challenges or political actions. Editors have tried to include all such changes in the maps contained herein (so far as they are discernible in the level of detail provided in the maps).
Although much less common, counties or their equivalent (e.g. parishes, election districts, or independent cities) can also change. Sometimes new counties are formed from old ones, independent cities are subsumed into existing counties, or boundaries change on a regular basis, as in Alaska.
Election Districts, Independent Cities, and Parishes
The word county, when it comes to election returns, is used somewhat broadly in this product. While most states are made up of counties, there are exceptions. Virginia has both counties and independent cities that report election returns. Alaska has election districts (that are redistricted every ten years) and Louisiana has parishes. Because these entities act similarly to counties and all report election data we have included them with counties for ease of use.
U.S. Political Stats draws on a number of authoritative CQ Press sources for elections data. Data have been adapted from these CQ Press sources:
- America Votes biennial series
- America at the Polls series
The America Votes series, covering presidential, gubernatorial, House, and Senate general elections and primaries every election cycle, was first created by Richard M. Scammon and Alice V. McGillivray of the Elections Research Center, Washington, D.C., in 1956. Since 1996 the series has been edited and compiled by Rhodes Cook. The series includes the following:
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 15. Washington: CQ Press, 1983.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 16. Washington: CQ Press, 1985.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 17. Washington: CQ Press, 1987.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 18. Washington: CQ Press, 1989.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 19. Washington: CQ Press, 1991.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 20. Washington: CQ Press, 1993.
- Scammon, Richard M., Alice V. McGillivray, and the Elections Research Center. America Votes 21. Washington: CQ Press, 1995.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 22. Washington: CQ Press, 1998.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 23. Washington: CQ Press, 1999.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 24. Washington: CQ Press, 2001.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 25. Washington: CQ Press, 2003.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 26: 2003–2004. Washington: CQ Press, 2006.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 27: 2005–2006. Washington: CQ Press, 2008.
- Scammon, Richard M., Alice V. McGillivray, and Rhodes Cook. America Votes 28: 2007–2008. Washington: CQ Press, 2010.
- Cook, Rhodes . America Votes 29: 2009–2010. Washington: CQ Press, 2012.
- Cook, Rhodes. America Votes 30: 2011–2012. Washington: CQ Press, 2014.
The data in this series are compiled from final, official results obtained from election authorities in each state. On occasion, states may belatedly report vote total changes that occur after publication of the volumes. The editors and CQ Press have made every attempt to incorporate changes and corrections in the data. Where possible, footnotes and notes regarding specific elections or vote totals have been retained in this Collection or integrated into the data. Presidential general elections and primaries can also be found in:
- McGillivray, Alice V., Richard M. Scammon, and Rhodes Cook. America at the Polls 1960 to 2000: John F. Kennedy to George W. Bush. Washington: CQ Press, 2001.
The data in this series are from the official state canvass reports.
Interest Group Scores
Annually, many interest groups rate members of the House and Senate on how well each member's voting record conforms to each group's principles or legislative aims. Organizations provide these scores at least in part to provide constituents and other groups with an easy-to-read barometer of a member's ideological leaning or policy preference. Because some groups have provided consistent ratings across time they can provide a useful longitudinal glimpse into a member's voting history.
Within this product users will find the following organizations represented:
Americans for Democratic Action (ADA)
An organization dedicated to individual liberty and economic and social justice at home and abroad. (1983–present)
American Conservative Union (ACU)
An organization aimed at effectively communicating and advancing the goals and principles of conservatism through one multi-issue, umbrella organization. (1983–present)
American Civil Liberties Union (ACLU)
An organization dedicated to preserving, defending, and expanding application of the constitutional guarantees and freedoms set forth in the Bill of Rights. (1983–present)
American Security Council (ASC)
Under the slogan "Peace Through Strength" the organization promotes educational programs in the areas of U.S. foreign policy, national security, and the global economy. (1983–1994)
Chamber of Commerce of the United States (CCUS)
An organization that advocates for business and free enterprise before Congress, the White House, regulatory agencies, the courts, the court of public opinion, and governments around the world. (1983–present)
Consumer Federation of America (CFA)
An association of hundreds of consumer groups dedicated to advancing pro-consumer policy on a variety of issues before Congress, regulatory agencies, and the courts. They disseminate information on consumer issues to the public and the media, as well as to policymakers and other public interest advocates. They provide support to national, state, and local organizations committed to the goals of consumer advocacy and education. (1983–2000)
Committee on Political Education of the AFL–CIO (COPE)
As part of the American Federation of Labor and Congress of Industrial Organizations, this committee is made up of union members that work to inform and mobilize union families on a variety of political and policy issues. They provide candidate funding and support of legislation and candidates at all levels of the political process. (1983–present)
League of Conservation Voters (LCV)
An organization dedicated to holding members of Congress accountable for their environmental votes and to help elect a pro-environment majority to Congress. (1983–present)
National Education Association (NEA)
A professional employee organization that advocates on issues relating to public education. Beginning in 2005, NEA switched from a numerical rating to an alphabetical rating. For databasing reasons, these alphabetical ratings have not been included. (1983–2004)
National Taxpayers Union (NTU)
A citizen group that works for lower taxes, smaller government, taxpayer rights, and accountability from elected officials at all levels. (1983–present)
Individual Interest Group Methodologies
For more information on how scores have been devised for each interest group, please consult Michael Sharp Directory of Congressional Voting Scores and Interest Group Ratings, vol. 1, pages vii-xii.
Interest Group Score Averages
Whenever an average is given that value represents all representatives or senators from a particular group with an interest group score. This could potentially include or exclude scores for people who did not serve the full year or the Speaker of the House (who votes at his or her discretion) depending on the interest group's methodology. If you want to see if a specific person received a score in a particular year, use the person finder on the interest group hub page.
Interest group scores for 1983–2004 are sourced to: Sharp, J. Michael. Directory of Congressional Voting Scores and Interest Group Ratings. 4th ed. Washington, DC: CQ Press, 2006. Data for subsequent years has been provided to CQ Press by the author.
Presidential Performance Indicators
U.S. Political Stats has a wealth of data on presidential actions that can be accessed and viewed in a number of ways. These actions, from speeches and appearances to executive orders, provide a useful glimpse into a specific presidency and an insightful comparative measure between presidents. Because of the diversity of data there are many different methodologies and sources used to compile the data. Users should consult the notes and source information for each individual data set for detailed information. This is particularly the case for presidents that assumed office outside of the normal election route or did not finish their term.
Users looking for high-level term totals that can be easily compared across presidents should use the "Office" button on the hub page. Simply select the category and indicator of interest and a bar chart will appear that shows total numbers by presidential term. Users looking for more in-depth annual data can use the "Person" button and search by president. Once an indicator and category are selected, users can find line charts that show totals by year. For presidents who served multiple terms, users can access data by year for an individual term or for all years of service in the White House.
For presidents who served more than one term, users will see a "I" or "II" (or higher, in Franklin Roosevelt’s case) beside a president’s name to clarify from which term the data is coming from. To find data by an individual president’s term, use the "Person" search on the Presidential Performance page and then select the category and indicator you are interested in.
Statistics on presidential actions in U.S. Political Stats were originally published in: Ragsdale, Lyn. Vital Statistics on the Presidency. 4th ed. Washington: CQ Press, 2014. Specific information on original data collection methodology and sourcing can be found in each data set’s notes and source sections.
While U.S. Political Stats does contain all roll call votes in the House and Senate from 1983 to the present, it does not represent all legislative activity on the floors of those bodies. Voice votes or standing/division votes do not record the votes of members by name and are therefore not included. Quorum calls are also not included here.
It is important to remember that a bill might have quite a few roll call votes as it winds through the legislative process. Therefore it is important to look at the date and, perhaps most importantly, the vote description to see what legislators were voting on.
Roll Call Voting Options
A lawmaker’s response to a roll call vote in the House and Senate can be a “yea”, “nay” or “present”. However, there are other ways for members who cannot make it to a vote to express how they would have voted or counteract their lost vote.
The official roll tally from the House and Senate includes only the designations of aye (or yea), no (or nay), present, and did not vote. Of those that did not vote, Congressional Quarterly (now CQ Roll Call) provides additional information, when available, indicating their position on the vote even though they did not cast one.
Lawmakers who miss a vote can subsequently make their voting preference known by having it published in the Congressional Record. A member doing this would still be listed as “did not vote” in the official roll call, but here users will see the “announced for” or “announced against” designation beside their name. In terms of the roll call totals in this product, these legislators would be included in the “did not vote” category since they did not officially vote. Often, though, legislators will simply not make a vote and will not follow up in any way. This would register simply as a "did not vote" in the data.
Vote pairing in the House is one way for an absent legislator to impact the outcome of a roll call, although it is an uncommon practice. When a member cannot make a vote, they may enter into an agreement with a member of the opposing side to form a pair. The present member of the pair withdraws his/her vote, votes present, and announces the pairing. The action is recorded in the Congressional Record. Had the absent member been there and both had voted, the votes would have canceled each other out. While U.S. Political Stats will indicate such a pairing, the official roll call would only indicate a “present” vote for one and a “did not vote” for the other.
Roll Call Variables:
Yea (voted for)
Nay (voted against)
Did not vote
Indications of opinion for those that did not vote:
Mapping Senate Roll Call Votes
Due to the complication of having two vote outcomes per state, U.S. Political Stats uses diagonal striping when a state's senators were not in agreement. In cases where two senators cast the same vote or if a senate seat was vacant and only one vote was registered, the state will appear as one solid color. In all cases, consult the map key for color coding or mouse over the state for more specific information.
When reviewing maps of roll calls, users will occasionally find districts (for House members) or states (for senators) without a data point. One possible cause of a missing data point is a seat vacancy. When a representative or senator leaves office in the middle of a term it can take some time to appoint or elect a replacement. Until a replacement is sworn in there would be no roll calls attached to that particular seat.
As each roll call vote is databased, CQ Press staff assigns to it one or more relevant topic categories from our congressional taxonomy. For example, an agricultural appropriations bill may be assigned to both an appropriations and an agricultural topic. For this reason when one looks at the total number of roll call votes by policy or topic area the total number of roll calls presented in that data set will exceed the total number of roll calls taken that year.
Committee of the Whole Votes
U.S. Political Stats does not include the four delegates from the District of Columbia,U.S. Virgin Islands, Guam, and American Samoa, nor does it include Puerto Rico's resident commissioner. Generally speaking, these representatives do not have voting powers. During periods of Democratic control of the U.S. House in the 1990s and 2000s, however, these representatives were given limited voting rights when the House constituted itself as a Committee of the Whole to vote on legislation. Consequently, users will find in some floor vote descriptions that the vote total number of votes exceeds the usual 435.
Roll call tallies are published in CQ Weekly and databased from that source several months following the vote. This includes relevant metadata including: roll call date, bill number, party affiliation of each member, vote totals, and vote description. Vote descriptions that describe both the legislation at hand and the purpose of the vote are written by CQ Roll Call staff. Historical roll call data going back to 1983 were databased from the annual print book series Congressional Roll Call, published by Congressional Quarterly, Inc. (and later CQ Press).
Supreme Court Cases
Years versus Court Terms
Users will find ruling outcome data assigned to a single year. Currently, any data by year include opinions handed down from that calendar year, not the Supreme Court’s term (which is also denoted by a single year). Data by term is forthcoming.
Ruling outcomes (e.g. 5–4; 9–0) are a short-hand way of describing how the justices ruled on a particular case. Generally speaking, the numbers represent those that sided with the majority opinion and those that dissented from it. However, they are often not that simple. Many cases yield numerous opinions that can make the outcome less clear cut. CQ Press relies on legal journalist Kenneth Jost's interpretation of case opinions in order to determine the outcomes used here.
Although there are nine justices on the Court, occasionally a justice will excuse himself or herself in the consideration and ruling of a case. Justices do this for a variety of reasons. There could be a financial conflict of interest or prior professional conflicts. In the case of Justice Kagan, for example, there were cases that she was involved with previously as U.S. Solicitor General that she did not subsequently participate in as a justice.
Assignment of Case Topics
One or more topics is assigned to each case in the database based on the primary merits of the case. These range from societal issues to individual articles of the Constitution. Some cases do have more than one topic coding assigned to it, although that is avoided when possible. As a result aggregate counts by topic may exceed the total number of cases.
The Supreme Court statistics are derived in part from the database created by Professor Harold Spaeth (Michigan State University). Since 2001 CQ Press editors, in consultation with legal journalist Kenneth Jost, have independently updated the data during the annual Court term.
For decades, Congressional Quarterly (now CQ Roll Call) has tracked voting habits of members of the House and Senate. By analyzing the roll call votes of each member CQ is able to generate a value that reflects his/her frequency of voting (voting participation), alignment with party positions (party unity), and alignment with the stated legislative positions of the president (presidential support). While not perfect indicators of ideology or attendance these scores provide a powerful statistical indicator for comparing members and looking at trends over time. Here you can find values for particular members, find aggregate averages, and compare member scores to party averages or other members.
Conservative Coalition Support/Opposition
The “conservative coalition” refers to a voting alliance between Republicans and Southern Democrats against Northern Democrats. This voting alliance was particularly relevant and influential in Congress during the second half of the twentieth century. Congressional Quarterly (CQ) ceased publication of the scores after 1998 when the alliance became politically “moribund.”
The calculation began with CQ’s selection of conservative coalition votes. A vote was classified as such when a majority of voting Southern Democrats and Republicans opposed the position taken by a majority of voting Northern Democrats. Votes where there was an even divide amongst any of the three groups were not included.
The conservative coalition support score represents the percentage of total conservative coalition votes in which a member voted (yea or nay) in agreement with the position of the conservative coalition. The opposition score represents the percentage of conservative coalition votes in which a member voted against the position of the coalition.
In CQ’s calculations southern states include: Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, and Virginia.
This value represents the percentage of recorded votes in a given year on which a member was eligible and present, and voted “yea” or “nay”. While this is not a perfect indicator of attendance, it is the best approximation of a member’s legislative participation. A 100% would mean that a member voted “yea” or “nay” in every vote for which they were eligible and present.
Part Unity represents the percentage of recorded party unity votes in a given year on which a member voted “yea” or “nay” in agreement with a majority of his or her party. (Party unity votes are those on which a majority of voting Democrats opposed a majority of voting Republicans.) Party Opposition is the percentage of recorded party unity votes in a given year on which a member vote “yea” or “nay” in disagreement with a majority of his or her party. For both unity and opposition percentages are based on votes cast; therefore, failure to vote does not lower a member’s score.
The Presidential Support score is the percentage of recorded votes cast in a given year on which the president has taken a position and on which the member voted “yea” or “nay” in agreement with the president’s position. The Presidential Opposition score is the percentage of recorded votes cast in a given year on which the president took a position and on which the member voted “yea” or “nay” in disagreement with the president’s position. For both of these calculations a failure to vote does not lower an individual’s score.
Notes about Calculations
Vote scores are most useful when a member is in office for the entirety of a legislative session (year). However, for a host of reasons members leave the House and Senate before their term is up or are prevented from voting, for example, due to illness. Their vote score and the person replacing them needs to be considered in the context of their time served. Notes on members who joined or left Congress during a term are included alongside the vote score for a given year.
In cases where members of the House have joined the Senate due to a vacancy mid-term CQ often gives them a score for both the House and Senate. In these cases, the member’s Senate score is listed as their primary score and their House score is included in the note.
Whenever an average is given that value represents all representatives or senators from a particular group with a voting score. This could potentially include scores for people who did not serve the full year or the Speaker of the House who votes at his or her discretion. For a member who moved from the House to the Senate in a given year, their Senate score would be the one included in an average.
If you would like further information on CQ’s vote calculations, please consult the CQ Almanac for the year in question. CQ Almanac also provides alternative measures that remove some procedural votes from the calculations.
Individual member voting scores were originally published in the 1983–2012 editions of Congressional Quarterly (and later CQ) Almanac. Most recent scores can be cited to: Austin, Jan, and Jennifer Rubio, eds. CQ Almanac 2012. Washington: CQ Roll Call, 2013.