Within each household, one adult was designated by a random procedure to be the respondent for the survey, to gain equal access to all adults in the households.

Within each household, both in the new poll and the earlier polls, one adult was designated by a random procedure to be the respondent for the survey.

Within each household one adult was designated by a random procedure to respond to the survey.

Within each household one adult was designated by a random procedure as the respondent.

For both polls, one adult within each household was designated by a random procedure to be the respondent.