CNN and the Kaiser Family Foundation, a nonprofit group that researches health care and other issues, partnered to conduct the Survey of Americans on Race in order to produce a nationally representative poll on race and ethnicity to analyze how different groups feel about important issues facing their daily lives. The results of the survey provide a look at how Americans of different races and ethnic groups feel about race relations, discrimination, unfair treatment by police and social and economic inequality. Both CNN and KFF contributed funding for the survey and participated in the development of the questionnaire. Each organization is responsible for its own content related to the results of the survey.
The survey was conducted August 25 through October 3, 2015 among a randomly selected probability-based sample of 1,951 adults. Interviews were conducted by SSRS of Media, Pa. in English or Spanish, depending on the preference of the respondent. In order to ensure there were enough interviews to reflect the views of African Americans and Hispanics, additional interviews were conducted with randomly selected people from those groups (commonly referred to as an “oversample”), for a total of 500 interviews with African Americans and 501 with Hispanic Americans.
Sampling for the poll consisted of three elements: 1) General population interviews conducted among those reached by random digit dialing either landline telephones or cellphones; 2) Respondents reached via random digit dialing on either landline or cellphones but interviewed only if they were black or Hispanic; 3) Prescreened respondents reached by calling back phone numbers originally part of a random digit dial sample where black or Hispanic respondents had been reached before.
A multi-stage weighting process was applied to ensure an accurate representation of the national adult population. Weighting parameters corrected for the overrepresentation of blacks and Hispanics in the sample, the probability of selection of individual respondents to the sample, and selected demographic and phone usage traits.
For results among the full sample of 1,951 adults, the margin of sampling error is plus or minus 3 percentage points. Among the full sample of white respondents, it is plus or minus 4 points, and for blacks or Hispanics, it is 6 points. For questions asked of a half-sample overall, it is 4 points. For those asked of a half-sample of whites, it is 6 points, and for a half-sample of blacks or Hispanics, it is typically 8 points for each group.