Survey Design

Our cost-effective, tailored sampling methods are grounded in probability-based designs and best practices. For general population samples, designs include address-based and multistage sampling, as well as targeted methods such as small group sampling and network sampling. For surveys of specific populations such as Medicare recipients, healthcare providers or school-age children, we sample from population list frames using combinations of stratified, multi-stage, balanced and systematic sampling. To minimize sampling bias, we use fully reproducible random sampling at all selection steps, and ensure that the frames are representative of the population. Our designs achieve the sample sizes and precision levels required by the study objectives and include adaptive methods to address obstacles to equity. 

Westat has extensive experience designing and fielding cross-sectional and longitudinal studies, including some of most complex and longest-term civilian population studies conducted to date. These studies represent a variety of study designs, data collection modes, subject matter, populations, respondent incentive schemes and contact protocols, and data collection intervals. Scalable corporate based systems that support management of a range of study designs, from simple to very complex, ensure our teams can focus on quality implementation that promotes equity and project specific goals throughout the lifecycle. Our cross-sectional designs allow researchers to look at characteristics in a population. Our longitudinal study design helps clients explore trends, examine interventions, or track changes in attitudes or behaviors.

Survey methodologists, subject matter experts and operations specialists work together to carefully craft our survey instruments to ensure that questions are clear, unbiased, and aligned with research objectives. Our team is experienced in developing surveys for special populations, covering sensitive topics, and incorporating multicultural and multilingual designs. Our facilities include an onsite laboratory for conducting focus groups, cognitive interviews and usability testing. Additionally, our team uses modern approaches for virtual testing of survey instruments and questionnaires, building greater equity into our pretesting and question development processes.

Our multimode data collection strategies maximize survey response and data quality. We select optimal modes, combination of modes, and data collection protocols to fit the study population and research goals to minimize bias, while controlling costs. Our technological solutions allow us to effortlessly design, track, and analyze multimode research. Learn more about the Future of Multimode Survey Designs or  Westat’s data collection methods and capabilities.

Westat uses leading commercial and custom survey software that seamlessly integrates with powerful survey data management, dashboard reporting, and data delivery platforms to provide end-to-end solutions to our clients. Our survey software tools and platforms easily scale to accommodate small- to large-scale data collection efforts and support all data collection modes. Learn more about Westat’s key survey platforms, Blaise and Survey Builder.

We collect and analyze data on the survey process and data itself to monitor quality and identify areas for improvement throughout the lifecyle of a project. This allows us to identify and address issues as they arise, rather than waiting until data collection is complete. 

We systematically compare the characteristics of respondents to nonrespondents to identify possible biases in our results so that adjustments can be made to address them. When appropriate, we supplement this with comparisons to external information such as results of other comparable surveys or other data sources.

Interviewer monitoring plays a crucial role in minimizing errors and biases. This involves ensuring that interviewers are adhering to the study protocol, asking questions accurately and consistently, and collecting high-quality data from the respondents. We employ a variety of monitoring techniques to achieve this, ranging from basic observation and quality control procedures to cutting-edge machine learning (ML) and natural language processing (NLP) tools. Our custom computer-assisted recorded interviewing (CARI) system enables near real time assessment of both interviewer and question performance, while our custom EAGLE platform enables 100% of field monitoring at scale. Learn more about these and our other research support tools.

We use a variety of methods, including pretesting, response rate analysis, checks for consistency, data cleaning, statistical analysis, and post-survey validation to ensure that our survey results are reliable and accurate.