You are here: Home / Reports / October 2011 NSF Workshop

Executive Summary from October 2011 NSF Workshop on the Future of Observatories in the Social Sciences

This document summarizes the workshop held in October 2011.

See the final report prepared by Emilio Moran and Sandra Hofferth


The following reports on the results of a Workshop held for the National Science Foundation in the Hilton Arlington Hotel, Arlington, VA, October 12-13, 2011 attended by 9 leading scientists from social, behavioral and economic disciplines. The goal of the workshop was to develop a plan for the network of observatories, with particular attention to the design of a fully fleshed-out set of centers, and an initial call for proposals focused on a pilot or set of pilots to inform this overall design. The Workshop on the Future of Observatories in the Social Sciences that convened in October 2011 took the recommendations of the December 2010 workshop as its starting point.

Observatory Design

Substantial discussion focused on what a place-based design would be. This included discussion of whether sites should from the outset be based on a representative sample of the American population or instead be geographically representative and distributed to ensure that both large and small populations are represented. Another question was whether some should be sited in areas of low population to capture social processes currently not well represented in nationally-based representative samples. One observatory model, the National Ecological Observatory Network (NEON), focuses on land use and climate variability. “If you sample people, you sample coasts. If you sample land you get lots of areas that have no people.” Although the observatories should link social science and natural science data, our focus is people and the relationships among them.

The discussion confirmed the importance of placed-based data collection – helps us understand process in a local environment. “By getting a place, one can integrate information across studies that others around the table have done in other disciplines from other perspectives, for example, how drought changes households, and how nonprofit groups respond. Integration across levels is needed. Some disciplines are oriented to understand things this way across multiple levels, but others are not. “A place allows people to put these all the levels together. Then, one can compare across places.”

The consensus was also that the sample should be developed a priori and deliberately rather than accidentally. The Long-Term Ecological Research stations (LTERs) (of which there are now 25) are an accidental sample. One participant said that she would settle for a small sample design that’s well designed. Most participants supported the nationally representative design. Although not all thought it crucial, the consensus was that it needed to be national and probability-based. The focus would be on sampling people not land, but need to represent the entire nation. The purpose is to make a broad, deep statement about what is going on in enough places to be nationally representative. The study would combine the best of in-depth community studies and new data forms of passive data collection and administrative data with the strength of broad survey research. We would want to foster different degrees of granularity. The consensus was to bring the small and large-scale studies together to make data that are both broad and deep. “Let’s not do it by accident.” There was some concern that the sample would be drawn that would not “look representative” on the face of it. Need to address this concern head on.

The participants focused on the need for new ways of collecting data and new types of data within this observatory model. Because of the emphasis on innovation, there will also be a need for training of investigators and of students – technical assistance. Synthesis and integration will be important. This model will also produce access/data sharing issues. There will be a need for new work on confidentiality issues and development of agreements to merge different types of data.

The big remaining question is what are the critical dimensions in each field on which there should be variability? Need to decide on which criteria to stratify or sample on. Need to be able to specify the types of variation at different levels that make sense for different issues. Some dimensions mentioned are state, county, school system policy differences; county or city size, pollution, environmental risk, unemployment rate, occupational concentration such as manufacturing, inequality; household socioeconomic differences such as education, occupation, income; cultural differences such as religion, race and ethnicity, and immigration status. Select areas on potential risk? The county as the PSU seemed to make the most sense, but others were interested in PUMAs and other potential sampling units.

Substantive Themes of the Observatories

The second part of the discussion focused on the categories of questions that would be the focus of these observatories. What are the exciting SBE science questions that would be transformative for the science and society, and could benefit from a network of observatories? The discussion began with separate discussions of the 5 significant themes that developed from the December 2010 meeting: (1) Children, Education and Our Future; (2) Sustainable Cities: Land Use and Global Change; (3) Changing Social and Policy Networks; (4) Resilience and Vulnerability to Hazards; and (5) Rebuilding the Middle Class.

The discussion focused on how these themes could be better consolidated. For example, “children, education and our future” and “rebuilding the middle class” have a lot in common. The participants did not care for the wording of “rebuilding the middle class” because it felt too political. The participants regrouped the topics. The consensus was to make the first theme “Opportunity and Mobility.”

Resilience, vulnerability, sustainability, were then subsumed under a second theme entitled “Change and Adaptation.” Networks could be placed under either theme, for example, social capital under opportunity and mobility; political networks under Change and Adaptation.

The consensus was that there would be two themes and multiple objects of study in each theme: biomarkers, individuals, families, social institutions (government agencies, private firms, NGOs), and the built and the natural environment, to cite a few examples. There would also be novel data collection instruments and linkages across data.

See the final report prepared by Emilio Moran and Sandra Hofferth