SSRS Opinion Panel
Rigorous Standards Combine Accuracy, Flexibility, and Affordability
Our highly experienced SSRS Opinion Panel team stands ready to support both data collection and analytic client needs. We work closely with clients from the early project planning stages to developing and delivering effective customized solutions. Our services include questionnaire design, web questionnaire programming and hosting, user experience (UX) testing, cognitive interviews, data collection, statistical weighting, and data analysis and reporting. Specific SSRS Opinion Panel services include:
- Methodological consulting
- SSRS Opinion Panel custom surveys
- Sub-population specific surveys such as surveys of Hispanics, people with specific health insurance statuses, caregivers, parents, registered voters, and people with different chronic health conditions
- Establishment and maintenance of custom proprietary panels
SSRS Opinion Panel members are recruited randomly based on nationally representative Address Based Sample (ABS) design (including Hawaii and Alaska). ABS sample is drawn from the Delivery Sequence File (DSF) maintained by the US Postal Service. Population coverage of DSF is in 98%-99% range.”
The SSRS Opinion Panel is a multi-mode panel. All internet respondents participate via web and all non-internet/web reluctant respondents participate via phone. SSRS’s own research on non-internet respondents suggests that more than 8 in 10 non-internet respondents are unwilling to take surveys via web. SSRS has therefore made a decision not to provide internet access to non-internet respondents as providing internet access does not effectively address the coverage bias of non-internet respondents. By definition, non-internet cases need to be contacted via a non-internet mode.
For studies that need completion via a singular web mode yet seek to ensure national representation, the SSRS Advanced Methods team has developed a proprietary propensity weighting solution that takes estimates derived for the U.S. adult internet population and adjusts them to mirror the full U.S. adult population (see the section on weighting).
SSRS web surveys comply with industry best practices. They are optimized for smartphone/mobile device administration and are adapted to all operating systems and browsers.
SSRS Opinion Panel data is weighted to be representative of the entire residential adult population in the United States. The first step in the weighting is the application of a base weight that accounts for the ABS sample design and the within-household sampling of one adult. We then apply a proprietary model-based non-internet propensity adjustment so that results are representative of all adults and not just those who have access to the internet. (This adjustment is made only to studies that have no telephone component for non-internet/web reluctant respondents.)
Following the base weight, survey data are calibrated to correct for differential nonresponse along demographics such as age, race, sex, region, and education. Benchmark distributions are obtained from the most recently available data from Current Population Survey (CPS) or American Community Survey (ACS). Additionally, online panelists are known to be more civically engaged than the general population. To correct for this potential bias, we often include volunteerism and interaction with neighbors when calibrating online samples. These benchmark distributions are derived from the Civic Engagement and Volunteer Supplement to the Current.
Most recent research suggests that probability samples interviewed by telephone or the internet provides accurate data. Surveys of a probability sample combined with an opt-in sample are less accurate than surveys of full probability samples. Least accurate are internet surveys comprised entirely of opt-in panel sample*. The SSRS Opinion Panel is a probability-based panel because it recruits randomly selected panelists from a probability-based sample source that reflects the full U.S. adult population (ABS). Results obtained from this panel can statistically represent the U.S. adult population, with a known margin of error. Unlike typical opt-in panels, no one can “volunteer” to be part of the SSRS Opinion Panel. They must be selected randomly and invited to participate.
*The Accuracy of Measurement with probability and non-probability survey samples by Bo MacInnis, Jon A. Krosnick, Annabell S. Ho, Mu-Jung Cho, POQ, Vol 82, No. 4, Winter 2018, pp. 707-744
During the recruitment process, full demographic information on panelists is collected. This data is stored securely and used to determine eligibility for specific studies (if needed). We also rely on these data to improve the survey experience for panelists by avoiding the need to re-ask demographic questions for each survey. Samples are drawn among panel members to best meet the study criteria. Selected panelists are sent an email invitation to participate in the survey, including unique survey login credentials. The respondent logs in and completes the self-administered online survey. Our contact protocol also includes a text survey reminder for panelists that have consented to receiving text messages.
The SSRS Opinion Panel also includes panelists who are primarily Spanish speakers. We translate and conduct surveys in Spanish as required by a specific project.
SSRS Opinion Panel standard deliverables include access to an online dashboard that provides near real-time project updates during data collection, a weighted SPSS data file including standard demographic variables, and an AAPOR compliant methodology statement with response rate calculations provided upon completion of all data collection and weighting.
SSRS is a full-service non-partisan public opinion firm. In addition to standard deliverables, our deliverables and other offered services are fully customizable to client needs.
A battery of demographic and other questions is asked during the recruitment of panelists. These questions are stored securely and kept for identifying target respondents, used for weighting, and can be appended to study data to expand what is known about respondents. Demographics are typically not re-asked for every survey, unless there is reason to believe that the data could have changed in a meaningful way. Key demographics are provided in the table.