ANES

1970 TIME SERIES STUDY

How to Cite
DATA
Return to Data Center
Errata
QUESTIONNAIRES
Post-election pdf
CODEBOOK
Introduction pdf   text
Variables pdf   text
Appendices pdf   text
FIELD MATERIAL
Interviewer Instructions pdf
Show Cards
  ABOUT THE DATASET:
Time Series Study
Completions: 1,694 post-election
Sample: fresh cross-section and separate supplement
Modes used: face-to-face
Weights: V700006-V700017

STUDY CONTENT HIGHLIGHTS:
In addition to content on electoral participation, voting behavior, and public opinion, the 1970 ANES contains questions in other areas such as political knowledge and partisanship. Some emphasis is placed on racial issues, including integration and racial composition of the respondent's environment. Other special-interest and topical content includes questions on the Vietnam war and student unrest, new issue scales on government health insurance, rights of the accused, government assistance to blacks, government action to control inflation, and pollution from private industry, as well as a new question on the relative conservatism of the major parties, and the first appearance of follow-ups to issue self-placements on importance of the issue to the respondent.

STUDY DESIGN HIGHLIGHTS:
The design of the 1970 ANES is unusual in several respects. First, the 1,694 total interviews include not only a traditional time series national cross-section, but also a supplement of 114 eligible black voters, in addition to a small number (73) of politically ineligible respondents from both the cross-section and black supplement samples. The ineligible respondents were too young to vote in 1970 but were included in an attempt to investigate the political attitudes of a group that would become eligible in the next Presidential election. The black supplement alone is not usable as a valid sample, but is intended for use together with the cross-section to create a combined sample, or with black respondents from the cross-section to produce a larger all-black sample. Because the black supplement portion of the combined sample slightly under-represents blacks in mixed neighborhoods, the combined sample is usable for cross-section analysis with exception for analysis on questions in which the racial environment might have an influence on the respondent's answer.

Another unusual aspect of the 1970 design lies in its use of the two questionnaire versions or 'forms,' which were randomly assigned to respondents as 'Type 1' or 'Type 2'. The objective of the questionnaire design was to maximize information obtained from black respondents while reducing the length of interview time for white respondents. Black respondents were asked all questions when assigned to either form, but white respondents were not, with the set of questions not administered to white respondents differing by form type. Questions related to racial issues were excluded for Type 1 white respondents, while questions on various other topics such as government responsiveness and the respondent's economic situation were excluded for Type 2 white respondents. Questions asked of everyone were designated Type 0. An additional feature of the 1970 design was a question ordering experiment, whereby the order of thermometer items differed according to questionnaire form.

Administration of the post-election interview began immediately after the November 3 elections, and continued until January 25, 1971. For the cross-section sample, the average interview length was 78 minutes, and 1,507 completions were yielded from 1,967 eligible.

WEIGHTS SUMMARY:
The 1970 ANES is unique to the time series in its use of variable-level weighting. Variables are weighted according to their (form) 'Type' and are also weighted differently within the different samples. For cross-section respondents, Type 0 variables are unweighted, and a weight for Type1 and Type 2 questions are present in both integer (V700006, V700007) and decimal (V700012, V700013) formats. For the combined sample, all three variable Types are weighted using integer (V700008, V700009, V700010) or decimal (V700014, V700015, V700016) weights. The all-black sample has a single integer (V700011) and a single decimal (V700017) weight for use with all variables. V700004 and V700005 are sample selection filters.