Harvard T.H. Chan School of Public Health/National Public Radio/Robert Wood Johnson Foundation: Discrimination in the United States Survey, 2017 (ICPSR 38387)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Study Purpose View help for Study Purpose

The study intended to document the self-reported prevalence of discrimination against African Americans, Latinos, Asian Americans, Native Americans, women, and LGBTQ (lesbian, gay, bisexual, transgender, and queer) adults across multiple institutional and interpersonal domains, including health services (health care) and social services (education, employment, housing, political participation, police, and the criminal justice system), as well as interpersonal domains that affect health outcomes, including slurs, microaggressions, harassment (sexual and nonsexual), and violence.

Study Design View help for Study Design

The study utilized an overlapping landline/cell phone sample frames RDD design. The sample plan consisted of three basic components: 1) stratified main sample RDD, where every adult qualifies and the general adult population respondents are reached by random-digit dialing (RDD) of cell phones or landlines; 2) stratified RDD oversample, where respondents are reached by RDD cell phone or landline, and interviewed only if they were members of one of the racial or ethnic groups at the focus of the study, or if they were members of the LGBTQ community; and 3) prescreened omnibus sample, meaning callbacks to telephone numbers where respondents who were previously interviewed on the SSRS weekly omnibus poll indicated they were members of one of these racial or ethnic groups or fell under various definitions of LGBTQ.

The questionnaire was translated into Spanish (n=255) and Chinese (n=33).

Sample View help for Sample

The sample includes 3,453 national adults, including oversamples of 802 African Americans, 803 Latinos, 500 Asian Americans, 342 Native Americans (defined as the Census category "American Indian or Alaska Native," or AI/AN), and 489 LGBTQ adults.

Universe View help for Universe

National adult, including oversamples of 802 African Americans, 803 Latinos, 500 Asian Americans, 342 Native Americans (defined as the Census category "American Indian or Alaska Native," or AI/AN), and 489 LGBTQ adults.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Description of Variables View help for Description of Variables

Usage Notes: There are multiple variables for gender in the dataset. The variable to use in analysis is SR_IR_REPORTED_GENDER. This variable is a composite of two measures. One is self-reported gender. The other is interviewer-reported gender, which is ONLY used to identify gender when the respondent does not self-report one. The LGBTQ questions are very complex, and it is best not to use them individually. When looking at LGBQ, use the variable LGBQVAR. When looking at transgender, use the variable CISTRANSVAR. However, if you are concerned specifically with LGBTQ responses, please be aware that the wording of the screening questions was changed a couple of weeks into the field period to try to improve the measures. Both versions are in the full final questionnaire (not in the summary version), with the dates they were asked. The two versions were coded into the summary variables listed above, so no important responses were lost. For racial and ethnic groups, use RACEVAR. For weighting, use the variable WEIGHT, not BASEWEIGHT.

Response Rates View help for Response Rates

The overall response rate for this study was 10%.

Original Release Date View help for Original Release Date

2022-03-10

Weight View help for Weight

The study dataset contains weight factors that should be employed in any data analysis. Typically weights are used in an attempt to assure that the survey sample more accurately represents the population. For weighting in this study, use the variable WEIGHT, not BASEWEIGHT.

The survey data were weighted to provide nationally representative and projectable estimates of the adult population 18 years of age and older both overall and for each of the racial/ethnic groups and LGBTQ adults. Each race-defined group was weighted using the following steps: 1) probability of selection (total), 2) correction for oversampling of racial/ethnic groups and strata, 3) recontact propensity correction, 4) LGBTQ adjustment. Then, with the base-weight applied: 5) the sample underwent the process of Iterative Proportional Fitting (IPF) for poststratification weighting ("raking"), 6) weight truncation ("trimming"), and 7) adjustment for race oversampling.