An important part of my job is figuring out how research studies can target a specific kind of person, often defined by geography, demographics, or some other factor. To do that, we need to consider the best way to create a comprehensive list or sample frame for any population of interest.

I’ve been interested in this issue since early in my career when we needed to identify hard-to-survey households across multiple languages and levels of literacy for a neighborhood health survey. I learned that designing sample frames provides the opportunity for creativity and outside-the-box thinking since you need to integrate multiple pieces of information from various sources.

Sample frames are such a basic ingredient in social science research that they are often taken for granted. A frame determines the universe of households, individuals, or businesses that may be included in a study, implicitly shaping research outcomes. Without a well-constructed sample frame that includes nearly all our target population, even the most robust sampling and data collection designs may be called into question or potentially lead to misleading research conclusions.

Inclusion in a sampling frame determines who may be part of a given study, and thus whose experiences shape public understanding and policy. As such, we’d ideally have a perfect list of all households, people, or organizations in the actual population for a given area, providing “full coverage” of that population.

In reality, it rarely is possible to create such a perfect frame for a number of reasons, including household mobility, database aging, and limitations in linking datasets. Any efforts to maximize coverage and “fill in” people who may be missing are fundamentally efforts to enhance survey representation.

Another example of specialized survey frames is the National Survey of Early Care and Education (NSECE), the definitive study of the availability and use of early care and education in the United States. NSECE uses GIS to create a custom sample frame of households that are within a convenient distance of geocoded child care providers, allowing us to collect data that help us understand the needs of families and the child care workforce. NSECE shows how we can be creative to capture specialized populations of interest, in this case, families who may need child care and providers who are near them.

NORC has continued to extend sample frame methodological research by investing in and integrating newer technologies and approaches. Our research has advanced topics including using aerial imagery to develop address lists in rural areas through remote listing and how best to efficiently target hard-to-survey populations with Big Data Classifiers (English, et al., 2025; Dutwin, et al., 2023). Such innovations allow us to reach populations beyond traditional methods, from rural households to transient urban communities. By combining technological advancement with methodological rigor, we ensure that our frames remain current, comprehensive, and flexible.