The Research and Development Survey (RANDS), which began in 2015, is a series of cross-sectional surveys from probability-sampled commercial survey panels. Two of the surveys also included a nonprobability sample for research purposes. RANDS has been used for methodological research at the National Center for Health Statistics (NCHS), including the use of close-ended probe questions and split-panel experiments for evaluating question-response patterns and the development of statistical methodology for the calibration of survey estimates that utilize the strength of national survey data. Survey results have been used to evaluate estimation approaches for health outcomes from recruited survey panels, including propensity score adjustment and calibration. In addition, the targeted embedded probe questions in RANDS have expanded findings from NCHS’ Collaborating Center for Questionnaire Design and Evaluation Research cognitive interviews to a wider sample.
Seven rounds of surveys have been completed, with responses collected during fall 2015, spring 2016, spring 2019, summer 2020, winter 2022, summer 2022, and fall 2022 (referred to as RANDS 1–RANDS 7, respectively). RANDS 1–3 collected responses using web administration, while RANDS 4–7 used web and phone administration. Each survey examined a sample of U.S. adults aged 18 and over, with respondent sample sizes ranging from 2,304 (RANDS 1) to 6,896 (RANDS 5). The questionnaires ask about health behaviors and conditions and were designed to be completed within 15 to 20 minutes. Rounds 1, 2, 3, 4, and 7 of RANDS include existing questions from the National Health Interview Survey (NHIS). RANDS 5 includes existing questions from the National Intimate Partner and Sexual Violence Survey (NISVS) and RANDS 6 includes questions from the National Survey of Family Growth (NSFG). In addition, RANDS 2–7 included embedded probe questions for cognitive evaluations. The specific topics in each round of RANDS have varied. Topics such as access to and use of health care, chronic conditions, food security, general health, health insurance, opioid use, physical activity, psychological distress, smoking, and disability have been included in the completed rounds of RANDS.
RANDS is administered by contractors using their proprietary recruited probability panels. Gallup used the Gallup Panel to conduct RANDS 1 and 2, while NORC at the University of Chicago used the AmeriSpeak Panel to conduct RANDS 3–7. Each round was conducted as a probability survey using recruited panels, with sampling strata assigned by demographic factors such as age group, sex, race and ethnicity, and education level. The panels and sample weighting methods differed between the two contractors, although both contractors adjusted the data using poststratification weighting to maintain proportionality of demographic groups in the population.
In addition to the seven rounds of RANDS noted above, in response to the COVID-19 pandemic, NCHS adapted RANDS to collect timely information on COVID-19. This special version of RANDS was named RANDS during COVID-19 to distinguish it from previous RANDS surveys. Two rounds with a longitudinal design were conducted in the summer of 2020 (RANDS during COVID-19 Round 1 and RANDS during COVID-19 Round 2), and a third round from an independent sample was conducted in the spring of 2021 (RANDS during COVID-19 Round 3). NORC conducted the surveys for all three rounds via web and phone administration using their AmeriSpeak Panel. For RANDS during COVID-19 Rounds 1 and 2, additional web responses were collected using the Dynata opt-in commercial survey panel. Similar to other rounds of RANDS, the RANDS during COVID-19 surveys were designed to be completed within 15 to 20 minutes. However, the RANDS during COVID-19 surveys featured larger sample sizes compared with most of the regular RANDS collections. RANDS during COVID-19 Round 1 included 6,800 AmeriSpeak Panel respondents and 6,220 opt-in panel respondents, while Round 2 included 5,981 AmeriSpeak Panel respondents and 5,502 opt-in panel respondents, and Round 3 included 5,458 AmeriSpeak Panel respondents. The RANDS during COVID-19 surveys were used to publicly release a set of experimental estimates on selected topics, including loss of work due to illness with COVID-19, telemedicine access and use before and during the pandemic, and reduced access to specific types of health care.