Preface

Most courses on data begin with a dataset. This book ends with one.

A dataset is the product of many decisions. A researcher decides what to study and how to define it. The researcher decides whom to ask, what to ask, and how to record the answers. By the time the file is ready for analysis, most of the choices that affect the result have been made. A statistical test cannot repair a question that was worded badly or a sample that left people out.

This book is about those earlier decisions. It follows the order in which a researcher makes them.

How the Book Is Organized

An introduction comes first. It explains what research is, the steps it follows, and where this book fits. Readers who are new to research should start there.

Eight chapters follow in four parts. The parts follow the order of the work.

  1. Foundations. Chapter 1 turns a question into a measure. Chapter 2 lists the ways a survey number can be wrong.
  2. Design and Measurement. Chapter 3 covers experiments and other designs. Chapter 4 covers scales, reliability, and validity.
  3. Instruments and Samples. Chapter 5 covers question writing. Chapter 6 covers sampling and coverage.
  4. Data, Analysis, and Reporting. Chapter 7 covers data collection and cleaning. Chapter 8 covers the analysis plan and the report.

Each chapter opens with a public case and uses that case throughout. The cases cover homelessness, immigration, hiring discrimination, food insecurity, community health, the census, online panels, and pay. These are topics where a number affects how money is spent or how a group is treated.

What Each Chapter Contains

Each chapter has the same parts. An opening case comes first, followed by the main sections. Key terms are in bold where they first appear, and a glossary at the end of the chapter defines them. Two kinds of boxes appear in each chapter.

  • An Ethics Moment describes a choice that affects the people being studied.
  • An AI Reality Check describes what a language model can and cannot do for that step of the work.

Each chapter also has a section called Work With the Data. It appears at the point where the idea is taught and is part of the chapter. In it you use a tool on the chapter’s data, with numbered steps and a note on what to look for. In the online edition the tool sits on the page. Readers of the print edition can open it in any web browser.

Each chapter ends with exercises at three levels. Check Your Understanding covers concepts. Apply It asks for a calculation or a design. Think Deeper asks for judgment, and those questions do not have one right answer.

What You Need

You need no software to read this book. The tools run in a browser. The data files are plain CSV files, and any spreadsheet program can open them. Readers who use R can find the code for every figure in the online edition.

Some statistics appear in Chapters 4, 6, and 8. Each is explained where it is used. A reader who wants more can turn to Margin of Error, the statistics book in this series.

For Instructors

The eight chapters can fill a short course or form one part of a longer course. An instructor guide is available on request. It has a session plan for each chapter, an activity for each tool, student worksheets, and answers to the exercises. Instructors can write to patilv@gmail.com.

The Companion Volume

This book covers research that produces numbers. Its companion, Beyond the Numbers, covers research that produces words, such as interviews, focus groups, and written comments. The two books can be read alone or together. Many research questions need both.