By combining up-to-date satellite imagery, detailed land-use classification, and advanced spatial analytics with proven sampling methodologies, we are able to provide richer context for a variety of complex issues, including community health burdens, access to services, and social vulnerability.

This approach not only improves the precision of sample selection but also unlocks powerful, actionable insights—helping policymakers and program designers to target interventions more effectively while elevating the rigor and relevance of survey research.

Remote sensing uses indirect measurement to collect data from a distance, theoretically enabling continuous measurement of conditions on Earth (Hall, 2010; Taubenböck et al., 2009). Sensors, such as those aboard aircraft and satellites, allow analysts to monitor and capture ground conditions over time. Data obtained using remote sensing techniques are often made available in the form of images, which can be processed to support scientific interpretation.

The spatial, temporal, and spectral resolution of images determine the precision with which ground conditions can be determined.

• Spatial resolution determines the level of detail present in the image based on the size of a single pixel and the ground area it corresponds to; higher spatial resolution images improve analysts’ ability to identify ground features.
• Temporal resolution is the time elapsed between images and indicates how often a sensor revisits and observes a location; higher temporal resolution supports change detection by providing more points in time for comparison.
• Spectral resolution indicates the sensor’s ability to differentiate wavelengths of light on the electromagnetic spectrum; imagery can be produced with multiple bands, such as red, green, blue, and infrared, which can be combined to produce enhanced composite images of specific feature types on Earth’s surface, such as vegetation.

Public programs such as NASA and USGS’s Landsat have been capturing images of Earth for over 50 years, providing an archive of data for studying changes to the Earth’s surface. In recent years, data from small satellites, drones, and other unmanned aircraft have also become available, making the prospect of using remotely sensed images in social science research even more feasible. Historical archives of publicly available imagery, such as from the Landsat program, provide a baseline for studies of change over time at broader scales, while other data sources may offer imagery with finer spatial, temporal, or spectral resolution, which can be used to identify specific features on the ground at a point in time.

Integrating remotely sensed images with other kinds of social science data, such as surveys and administrative records, can improve data accuracy, efficiency, and completeness (Fox, 2003; Hall, 2010; Rindfuss and Stern, 1998). For example, using satellite imagery alongside an address-based sample frame can help analysts evaluate and mitigate potential biases by identifying areas where residences or businesses may have been omitted or reviewing addresses that fall outside of urban areas. Linking remotely sensed images with survey responses also enables the exploration of relationships between respondents, their characteristics, and the context in which they live, providing points of comparison that can aid analysis and interpretation (Taubenböck et al., 2009). For example, individual survey responses about personal health can be coupled with data from remotely sensed images that indicate the survey respondent’s distance from sources of pollution, such as factories (Demetillo et al. 2020). Remotely sensed imagery with the appropriate level of resolution can be used to identify features, such as specific types of facilities, and detect environmental changes, such as new construction, flooding, or drought.

Geospatial satellite imagery can be used to construct enumeration areas and provide field teams with data in the absence of other high-quality information to develop or validate sample frames. Images, or derived information like land use and land cover, adds valuable context to address frames, which can be used for further stratification and analysis. In the absence of official statistics or other updated data sources, imagery can also be used to produce population estimates that reflect current conditions, especially for areas experiencing rapid growth.

Understanding the spatial distribution of environmental burdens, as well as access to critical care services, can provide more insights into health outcomes and help identify areas with high social vulnerability or resilience. Policymakers and community advocates can also use this information to develop spatially targeted interventions or strategies to help mitigate burdens or improve access to services and care. Integrating remotely sensed data into projects with survey data enriches responses and enables context-rich spatial modeling.

Remotely sensed imagery supports a variety of social science applications at NORC, with prospects for further development. Remotely sensed data has the potential to improve data collection, enhance data quality, and enrich insights into conditions on the ground in near-real time. Integrating remotely sensed imagery into social science research supports a wide variety of tasks, from address frame construction and enhancement to health research. Spatial modeling with remotely sensed data can advance social science research by linking outcomes observed in administrative and survey data (e.g., households experiencing food insecurity) to environmental factors (e.g., zoning), providing missing context and opening new lines of research inquiry.