Exploring search strategy design to identify diverse literature for inclusion in systematic reviews

Systematic reviews used to inform policy-making bring together diverse research literature. Such research literature is identified from a carefully designed search strategy. This thesis explores the search strategy design process for diverse literature, and draws on the findings of five of my research publications. It develops theory from case studies. In doing so, it uses an integrated approach using framework analysis, a thematic analysis and summary to develop higher order themes. These are configured into a model. My candidate publications centre upon a common theme of designing of search strategies for use in social policy-relevant systematic reviews in public health and social care. The findings of these publications are combined with findings from analysing search strategies I have designed for ten systematic reviews and an examination of guidance and methods literature. Ten dimensions of search strategy design emerged from the analysis, which comprise four overarching categories: 1) domain and publishing norms; 2) research questions and scope; 3) terminology; and 4) information organisation. Seven dimensions within these categories can be used to characterise diverse literature: multiple subject domains, diffuse dissemination, multiple search resource types, multiple meanings of search concepts, wide search term diversity, broad indexing terms, and the low discriminating power of the search. The remaining three dimensions are integral dimensions to search strategy design though they do not necessarily characterise diverse literature: search sample, establishing search concepts and searching strategy. Seven categories describing the influences of search strategy design are configured from my analysis: people, processes, resources, evidence, technology, standards and creativity and the problem-solving processes within information searching. A model of search strategy design illustrates how the dimensions and categories relate to each other