Summary :
The analysis in this book is largely though not entirely statistical, and we have thought it might be helpful to introduce at the outset some technical concepts and terms in statistics that may be unfamiliar to many readers.1 As we emphasize in the general introduction, we want this book to be accessible to judges, lawyers, law professors, law students, and others who are not adept at statistical analysis, as well as to those members of the legal profession who are, and of course to social scientists. We also explain conventions we use to make our analysis at once clear and compact. A good deal of our analysis involves testing hypotheses about the relations between a dependent variable (what we are trying to explain—for example, the ideological valence of judicial votes) and an independent variable or variables (what our theory suggests may explain the dependent variable—for example, judges’ ideological leanings, workloads, and interactions with colleagues). Sometimes we summarize the strength of the relation with a statistic called the correlation coefficient (r) or, more commonly, r-squared (that is, r multiplied by itself—usually, and by us in this book, written R2). An r of .5 or, equivalently, an R2 of .25 between judges’ ideology and their votes would indicate that judges’ ideology explained 25 percent of the variation in votes. Multiple regression analysis, which we use extensively, enables the separate effect of each independent variable on the dependent variable to be estimated. Thus we can estimate, for example, whether judges’ ideology affects judicial votes separately from the effect of other variables on their votes, such as workload.

Copies :