The basic principles and practices of SMMR are introduced based on the simplest scenario that one can encounter in applied social science research: a single sufficient condition. The main types of cases and various fundamental principles of SMMR are detailed, the same as the first formulas introduced that distinguish better from worse cases for within-case analysis if fuzzy sets are used. Learning goals: - Understand how membership of cases in the QCA solution and the outcome is used for sorting cases into one of the four basic case types - Learn about the first set of SMMR principles that guide the selection of cases for within-case analyses on the mechanism linking the condition to the outcome - Get acquainted with the seven possible SMMR designs: three single-case and four comparative SMMR designs - Become familiar with the basic logic of the smmr() function - Learn about formulas distinguishing between better and worse case choices for within-case analyses - Understand the different scenarios in which a typical and an iir case can hold membership in the condition, the outcome, and the mechanism and the implications of these scenarios for causal inference

Chapter 7 shows how RIO can facilitate algorithmic case selection. We outline how algorithms can be used to select cases for in-depth analysis and provide two empirical analyses to illustrate how RIO facilitates a deeper understanding of how cases relate to one another within the model space, and how they align with the theoretical motivations for different case selection strategies.

Chapter 9 demonstrates how RIO facilitates a field-theoretic approach to regression models. The chapter draws parallels between the data representations made possible by turning regression models inside out and the geometric data analysis (GDA) that is central to field theoretic approaches to social research.

Chapter 2 introduces the logic, basic mathematics, and some of the benefits of turning regression inside out in the context of Ordinary Least Squares (OLS) regression. We do this through an in-depth reimagining of a classic analysis of the effects of welfare state spending on poverty. The chapter introduces novel techniques for regression decomposition, data visualization, and geometric data analysis.

Chapter 6 demonstrates one way that RIO can be used for exploratory data analysis: identifying statistically significant interaction terms. We show how exploring the relationships among cases offers important insights into the relationships between variables.

Chapter 1 introduces the general logic and motivations behind turning regression models “inside out.” Here, we explain how Regression Inside Out (RIO) facilitates a case-oriented approach to regression, detail the benefits of such an approach, and provide a roadmap for the rest of the book.

Chapter 3 demonstrates how the mathematics of turning Ordinary Least Squares (OLS) regression inside out can be generalized to Generalized Linear Models (GLM) including logistic, Poisson, negative binomial, random intercept, and fixed effects models.

Chapter 10 concludes our book, outlining the benefits of a case-oriented approach to regression. We review key substantive findings from the analyses presented in previous chapters and highlight directions for future research.