The poll in this study is part of an on-going series of surveys developed by researchers at the Harvard Opinion Research Program (HORP) at Harvard T.H. Chan School of Public Health in partnership with the Robert Wood Johnson Foundation and NPR. The research team consists of the following members at each institution.
Harvard T.H. Chan School of Public Health: Robert J. Blendon, Emeritus Professor of Health Policy and Political Analysis and Executive Director of HORP; John M. Benson, Senior Research Scientist and Managing Director of HORP; Mary G. Findling, Assistant Director of HORP; Chelsea Whitton Pearsall, Research Coordinator.
Robert Wood Johnson Foundation: Carolyn Miller, Senior Program Officer, Research-Evaluation-Learning; Jordan Reese, Director of Media Relations; Martina Todaro, Research Associate, Research-Evaluation-Learning.
NPR: Andrea Kissack, Senior Supervising Editor, Science Desk; Joe Neel, Deputy Senior
Supervising Editor, Science Desk; Vickie Walton-James, Senior Supervising Editor, National
Desk; Marcia Davis, Supervising Editor, Race and Identity, National Desk.
Interviews were conducted online and via telephone (cellphone and landline), August 2 – September 7, 2021, among a nationally representative, probability-based sample of 3,616 adults age 18 or older in the U.S. Data collection was conducted in English, Spanish, Mandarin, Cantonese, Korean, and Vietnamese by SSRS (Glen Mills, PA), an independent research company. The survey examined experiences of households in the U.S. as a whole, in the four largest U.S. cities (New York City, Los Angeles, Chicago, and Houston), on households by race/ethnicity (including Black, Latinx, Asian, and Native Americans) nationally, on households with children, and on households in rural America. The margin of sampling error, including the design effect, was ±3.4 percentage points at the 95 confidence level for national data.
The core of the sample was address-based, with respondents sampled from the United States Postal Service’s Computerized Delivery Sequence (CDS) file. Sampled households were sent an invitation letter including a link to complete the survey online and a toll-free number that respondents could call to complete the survey with a telephone interviewer. All respondents were sent a reminder postcard, which also included a QR code they could scan to be linked to the survey via a smart device. Households that could be matched to telephone numbers and that had not yet completed the survey were called to attempt to complete an interview. In order to represent the hardest-to-reach populations, the address-based sample (ABS) was supplemented by telephone interviews with respondents who had previously completed interviews on the weekly random-digit dialing (RDD) SSRS Omnibus poll and online using the SSRS Opinion Panel, a probability-based panel.
A total of 3,177 respondents completed the questionnaire online, 83 by calling in to complete, and 356 were completed as outbound interviews.
Possible sources of non-sampling error include non-response bias, as well as question wording and ordering effects. Non-response produces some known biases in survey-derived estimates because participation tends to vary for different subgroups of the population. To compensate for these known biases and for variations in probability of selection within and across households, the samples were weighted to match the distribution of the population based on data from the U.S. Census Bureau’s 2020 Current Population Survey (CPS). Weighting parameters included: gender, age, education level, race/ethnicity, region, and party identification.
Respondents who were the only person living in a household were asked about their own experiences. Respondents who had anyone else also living in their household were asked about the experiences of anyone living in the household. Together these responses represent the experience of the household.