Full dataset from survey on relative partisanship used in Atsusaka, Yuki, & Kim, Seo-young Silvia (2025). Addressing Measurement Errors in Ranking Questions for the Social Sciences. Political Analysis, 33(4), 339-360. https://doi.org/10.1017/pan.2024.33
This data contains Americans' rankings of four sources of their identity, including political party, religion, gender, and race, for 1,082 respondents. The columns consist of marginal rankings to the main identity ranking question and the corresponding anchor question, as well as whether they have answered the anchor questions "correctly." The anchor ranking question is used to estimate the proportion of random responses and to correct for measurement error bias.
Format
## `identity` A data frame with 1,082 rows and 16 columns:
- app_identity
Full ranking profile for the main identity ranking question.
- party
Marginal ranking for party (main identity ranking question).
- religion
Marginal ranking for religion (main identity ranking question).
- gender
Marginal ranking for gender (main identity ranking question).
- race
Marginal ranking for race (main identity ranking question).
- anc_identity
Full ranking profile for the anchor ranking question.
- household
Marginal ranking for household (anchor question).
- neighborhood
Marginal ranking for neighborhood (anchor question).
- city
Marginal ranking for city (anchor question).
- state
Marginal ranking for state (anchor question).
- anc_correct_identity
Whether the respondent answered the anchor questions correctly. This is a binary variable that 1 if the respondent correctly answers the anchor ranking question and 0 if otherwise.
- app_identity_recorded
Recorded responses for the main identity ranking question.
- anc_identity_recorded
Recorded responses for the anchor ranking question.
- app_identity_row_rnd
The order in which the items were randomly presented for the respondent in the main ranking question.
- anc_identity_row_rnd
The order in which the items were randomly presented for the respondent in the anchor ranking question.
- s_weight
Survey weight.