This chapter is devoted to data analysis and its critical role in analytics science. The reader is introduced to the science of inference from observations and experiments and learns about the main ideas in data analysis that have been influential in addressing societal problems. Real-world examples are used throughout to convey the main ideas and illustrate why data analyses performed without sufficient care can yield wrong insights. Successful examples of insight-driven problem solving approaches in data analysis are contrasted with those that can yield wrong insights, and the reader is taken on an engaging yet educational journey that depicts how and why successful insight-driven problem solving approaches using data can have significant public impact.

This chapter extends the analysis from Chapter 7 to multi-case settings and demonstrate how we can use the approach to undertake mixed-method analysis. We show how, when analyzing multiple cases, we can update our theory from the evidence and then use our updated theory to draw both population- and case-level inferences. While single-case process tracing is entirely theory-informed, mixed-data inference is thus also “data”-informed. We show how the approach can integrate information across any arbitrary mix of data structures, such as “thin” data on causes and outcomes in many cases and “thicker” process evidence on a subset of those cases.