In the last chapter we explored the three key principles of effective communication: (1) communicate substance, not statistics; (2) when performing inference, convey uncertainty; and (3) graph data and results. We also offered some general advice for creating visually effective displays of data, urging you to aim for clarity, to iterate, and to write detailed captions.

In other disciplines, adherence to these principles has generated benefits for the producers and consumers of empirical research, and we have no doubt that research relating to law and legal institutions will see similarly salutary effects. Most crucially, as we explained in Chapter 10, moving towards more appropriate and accessible data presentations will enhance the impact of empirical legal scholarship—regardless of whether the intended audience consists of other scholars, students, policy makers, judges, or practicing attorneys.

At the same time, however, we realize that legal researchers require more than general guidelines; they need on-the-ground guidance so that they can convey the results of their (carefully designed and executed!) studies to members of legal and policy communities. So here we aim to get far more specific, offering analysts advice on how to translate data and results into powerful visual presentations.

In setting out the various strategies to follow, we divide the material into two sections. In the first, we focus on communicating data; in the second, on the presentation of results. We split the material in this way because the presentation of data and results are somewhat different tasks. For example, when performing inference, authors have an obligation to convey the level of uncertainty about their results, as we stressed in Chapter 10. But when researchers are merely displaying or describing the data they have collected—and not using their sample to draw inferences about the population that may have generated the data—supplying measures of uncertainty, such as confidence intervals, may be overkill.

This is a distinction between the two sections to come. The commonality, though, may be more important. In both we adhere to the general principles laid out in Chapter 10—though none more so than the very basic idea that researchers should almost always graph their data and results.

- end of excerpt -