Funders? mandates and changing social norms encourage social scientists to share and archive their research data so they can be reused for replication or to generate new knowledge. However, archiving data does not ensure social scientists will be able to find and reuse it. In this study, researchers at the Inter-university Consortium for Political and Social Research (ICPSR) will determine if transforming data search systems to include recommended results can help social scientists discover datasets. Recommender systems have enhanced and reshaped search in other contexts, such as book and video recommendation, but they have not been widely applied to research data. Furthermore, research on data recommendation should engage with theories of human-information behavior (HIB) to inform effective system design. Funders? mandates and changing social norms encourage social scientists to share and archive their research data so they can be reused for replication or to generate new knowledge. However, archiving data does not ensure social scientists will be able to find and reuse it. In this study, researchers at the Inter-university Consortium for Political and Social Research (ICPSR) will determine if transforming data search systems to include recommended results can help social scientists discover datasets. Recommender systems have enhanced and reshaped search in other contexts, such as book and video recommendation, but they have not been widely applied to research data. Furthermore, research on data recommendation should engage with theories of human-information behavior (HIB) to inform effective system design. We propose a study to understand how social scientists search for data and to test whether changes to search processes and providing data recommendations facilitate data discovery and secondary use.
This project will take place in two phases. First, to understand how social scientists search for data, we will develop and refine an HIB model of data search behavior. We will leverage a range of data, including semi-structured interviews and behavioral trace data, to validate our model and establish recommendation criteria. Second, we will use our model to prototype a data search system that includes recommendation. We will employ an iterative design process that uses low- and high-fidelity prototyping and user engagement, and the result will be an improved HIB model that describes research data-seeking behavior and a prototype search system that incorporates recommendation.

This project will take place in two phases. First, to understand how social scientists search for data, we will develop and refine an HIB model of data search behavior. We will leverage a range of data, including semi-structured interviews and behavioral trace data, to validate our model and establish recommendation criteria. Second, we will use our model to prototype a data search system that includes recommendation. We will employ an iterative design process that uses low- and high-fidelity prototyping and user engagement, and the result will be an improved HIB model that describes research data-seeking behavior and a prototype search system that incorporates recommendation.