Question wording varied from year to year, but this variable is always coded as family income. In the early years of the time-series (1958–1966) the question focused on the family’s expected income for this year. Afterwards the question focused on the family income in the previous year. Coding: 0 if 1–16th percentile, 0.25 if 17th–33rd percentile, 0.50 if 34th–67th percentile, 0.75 if 68th–95th percentile, 1 if 96th–100th percentile.
Coded using variables v580501, v640269, v660235, v720420, v742549, v763507, v800686, v700388, v941404, P023149.
Question wording has varied from year to year, but for most of the time series it is possible to construct stable categories. Coding: 0 if 8 grades or less, 0.25 if 9–12 grades with no diploma or equivalency, 0.50 if 12 grades, diploma, or equivalency, 0.75 if some college, 1 if college degree or higher.
For 1994, the 1992 panel data was used. For the 2002 year, the 2000 response was used. Individuals who reported having community college or junior college degrees were coded as 0.75.
Coded using variables v580478, v640196, v660197, v720300, v700269, v742423, v763398, v800445, v941209, P023131.
Questions that asked about employment status of respondent wording changed somewhat from year to year. From 1958 to 1966 this data was only collected about the head of the household rather than the respondent. Coding: 1 if unemployed, 0 otherwise.
Coded using variables v580479, v640202, v660199, v700275, v720306, v742443, v763409, v800515, v941216, P025183.
Coding: The square of the above response.
Question wording: “Now I’d like to read some of the kinds of things people tell us when we interview them. Please tell me whether you agree or disagree with these statements.” “I don’t think public officials care much what people like me think.” “People like me don’t have any say about what the government does.” Coding: for each item, coded 0 if agree, 1 if disagree, then summed and rescaled to zero–one interval.
Question wording: “Sometimes politics and government seem so complicated that a person like me can’t really understand what’s going on.” Coding: 0 if agree, 1 if disagree.
Question wording: “If a person doesn’t care how an election comes out then that person shouldn’t vote in it.” Coding: 0 if agree, 1 if disagree. We note here that this variable does not appear in the final Rosenstone–Hansen mode, but we wished to include it in this analysis. In future versions of this paper this variable may not be included.
Strength of Party Identification
Question wording: “Generally speaking, do you usually think of yourself as a Republican, a Democrat, and Independent, or what?” (If Republican or Democrat) “Would you call yourself a strong (Republican/Democrat) or not very strong?” (If independent, other, or no preference) “Do you think of yourself as closer to the Republican or Democratic party?” Coding: 0 if independent or apolitical, 0.33 if independent leaning toward a party, 0.67 if a weak partisan, 1 if a strong partisan.
Affect for Party
Question wording: “Is there anything in particular you like about the Republican party?” “Is there anything in particular you dislike about the Republican party?” “Is there anything in particular you like about the Democratic party?” “Is there anything in particular you dislike about the Democratic party?” Coding: the absolute value of the difference between two sums, coded to the zero–one interval: the sum of Democratic party “likes” and Republican party “dislikes” minus the sum of Democratic party “dislikes” and Republican party “likes.” For the 2002 respondents, their responses from 2000 were used. For 1974 respondents, their responses from the 1972 surveys were used.
Question wording (Presidential year): “Generally, speaking, would you say that you personally care a good deal which party wins the presidential election this fall, or don’t you care very much which party wins?” Question wording (Mid-term): “Now I’d like to talk with you a bit about the elections which took place this fall. As you know, representatives to the Congress in Washington were chosen in this election from congressional districts all around the country. How much would you say that you personally cared about the way the elections to congress came out: very much, pretty much, not very much, or not at all?” Coding: 1 if care a good deal, pretty much, or very much. 0 otherwise (including non-response).
Wording does change somewhat from year to year. This variable was coded using variables v580312, v640020, v660063, v700164, v720029, v742026, v763030, v800061, v940209, P023007.
Affect for Candidate
Question wording: “Is there anything in particular you like about [the appropriate Republican candidate]?” “Is there anything in particular you dislike about [the appropriate Republican candidate]?” “Is there anything in particular you like about [the appropriate Democratic candidate]?” “Is there anything in particular you dislike about [the appropriate Democratic candidate]?” Coding: the absolute value of the difference between two sums, coded to the zero–one interval: the sum of Democratic candidate “likes” and Republican candidate “dislikes” minus the sum of Democratic candidate “dislikes” and Republican candidate “likes.”
Question wording (1952–1968): “Would you say you go to church regularly, often seldom, or never?” Coding: 0 if never, 0.33 if seldom, 0.67 if often, 1 if regularly. Question wording (1970–2002): “Would you say you go to (church/synagogue) every week, almost every week, once or twice a month, a few times a year, or never?” Coding: 0 if never, 0.33 if a few times a year, 0.67 if once or twice a month, 1 if every week or almost every week. In 1994 an experimental version of this question appeared, so 1992 responses were used instead.
Years in Community
Question wording: “How long have you lived here in your present (city/town)?” Coding: actual number of years. When respondent chose “all of my life” their age was imputed here. When this variable was used on the right hand side, it is transformed using a natural logarithm to induce normality.
Question wording: “The political parties try to talk to as many people as they can to get them to vote for their candidates. Did anyone from one of the political parties call you up or come around and talk to you about the campaign? Which party was that?” Coding: 0 if not contacted, 1 if contacted.
Observed by interviewer. Coding: 1 if lives in Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Texas, or Virginia, 0 otherwise.
Observed by interviewer. Coding: 1 if lives in Missouri, Kentucky, Maryland, Oklahoma, or West Virginia, 0 otherwise.
The question wording on race and ethnicity have probably changed more throughout the ANES time-series than any other variable here. Throughout most of the time-series blacks, “negro”, or African-American is presented as one option. Coding: 1 if black, 0 otherwise. In the 2002 survey respondents were allowed to mark multiple racial and ethnic categories. All respondents who marked more than three categories were coded as missing and otherwise were coded as 1 if any of their choices included black or African-American.
This variable is missing for 1958. In the early years of the time series (1964) the best we were able to do was include the “other” category (coding: 1 if other, 0 otherwise) as this seemed to be the category that shifted most when Hispanic options were added in 1966. From 1966 until 1976 respondents were given the option to identify themselves as Mexican–American or Puerto Rican (coding: 1 if Mexican or Puerto Rican and 0 otherwise). For 1980 and 1994 ethnicity was coded separately and all those of Hispanic origin are coded as 1, and respondents were coded as 0 otherwise. In 2002 multiple choices were allowed, and we followed the analogous scheme as described in the “black” variable above.
Race and ethnicity variables were coded using variables, v580469, v640183, v660237, v720425, v742554, v763513, v800721, v800722, v700391, v941435, v941419, P023150.