|
Abstract:
|
The Census Bureau has been working to improve measurement of same-sex couple households, especially for same-sex married couples. In 2015, several tests of the revised relationship question were conducted. The new question lists explicit categories for "opposite-sex husband/wife/spouse," "same-sex husband/wife/spouse," "opposite-sex unmarried partner," and "same-sex unmarried partner." A relatively small population such as same-sex married couple households are difficult to count since it only requires a very small percentage of opposite-sex couples to mismark relationship or sex to affect the estimates. In anticipation of respondent mismarks, an automated check was also added to the self-response internet and mobile device nonresponse follow-up instruments. Tests in the spring of 2015 included the new relationship categories and edit across three modes (mail, web, and personal visit). Using response data and paradata, the paper will report on item nonresponse rates, consistency of reports between sex and relationship, and degree of misreporting and distribution of categories between the test and control versions.
|
