Using Your Own Data
Most of the tools in this book accept your own interviews, codes, and documents. This page explains how to prepare them, which tool fits which task, and how this book works with its companion.
Before You Open Anything
Qualitative data are personal. Chapter 5 covers what a researcher owes participants. Three steps apply before any transcript goes into any tool.
- Remove names. Replace each one with a pseudonym or a label such as “Participant 4.”
- Remove details that identify a person. Employers, street names, and rare events can point to one person. The Identifying Details Checker in Chapter 5 helps find them.
- Check your consent text. Use the material only in the ways participants agreed to.
Your Material Stays With You
The tools run inside your web browser. A document you open is read on your own device. It is not uploaded, and no copy is sent to the publisher or to anyone else. The Coding Workbench keeps a copy of your project in the browser between visits. Clearing the browser’s data erases that copy, so save the project to a file when you stop.
Which Tool for Which Task
| Your task | Tool | Chapter | What you bring |
|---|---|---|---|
| Check an interview guide | Interview Guide Checker | 2 | Your questions, one on each line |
| Check a quotation before publishing it | Identifying Details Checker | 5 | The quotation |
| Code interviews or documents | Coding Workbench | 6 | Text files or pasted text, and your codes |
| Judge whether to keep interviewing | Saturation Tracker | 4 | The coded segments from the workbench |
| Sort codes into themes | Theme Builder | 7 | The coded segments from the workbench |
| Measure agreement between two coders | Coder Agreement | 8 | A file with one column for each coder |
| Search and count in a set of documents | Text Explorer | 9 | A file with one row for each document |
File Forms
Documents for coding. Plain text files, one document in each, or text pasted into the tool. Save a word-processor file as plain text first.
Coded segments. The Coding Workbench downloads a CSV file with four columns, named document, code, quote, and definition. The Theme Builder and the Saturation Tracker open this file. A file made in a spreadsheet works too, as long as it has a document column and a code column.
Two coders. A CSV file with one row for each coded item. One column holds the first coder’s code and another holds the second coder’s. A third column can hold the text of the item.
Documents for searching. A CSV file with one row for each document and a column that holds the text. A plain text file also works, with a blank line between documents.
A Path Through Your Own Study
- Draft the interview guide and check it with the Interview Guide Checker. Then try it on one person.
- Conduct and transcribe the first few interviews. Remove names and identifying details.
- Code them in the Coding Workbench. Revise the codebook as Chapter 6 describes.
- Open the coded segments in the Saturation Tracker after every few interviews. Compare the result with the stopping rule you set in advance.
- Ask a second person to code a sample. Compare the two sets of codes with the Coder Agreement tool.
- Sort the codes into themes with the Theme Builder. Read every segment under each theme.
- Before quoting anyone, run the quotation through the Identifying Details Checker.
What the Tools Cannot Do
The tools keep records, count, and display. The analysis remains the researcher’s.
- They do not choose codes. A code is a judgment about what a passage means.
- They do not decide what a theme claims. The Theme Builder asks for a sentence and cannot tell whether the segments support it.
- They do not know the setting. A detail that looks harmless in a tool can identify a person in a small town.
Large projects with many coders or hundreds of documents are better served by dedicated software for qualitative analysis. The tools here are meant for learning the method and for small studies.
Using This Book With Before the Numbers
This book covers research that produces words. Its companion covers research that produces numbers. Many studies need both, in one of two orders.
Words first, then numbers. When a problem is poorly understood, start with interviews or focus groups. They show which reasons exist and what words people use. A survey then measures how common each reason is. The answer choices in the survey come from what people said.
Numbers first, then words. When a survey produces a result that needs explaining, follow it with interviews. A survey can show that patients without a car miss far more visits. Interviews can show what a missed ride looks like on the day, and what would have helped.
Table 2 shows where each step of the research process is covered in the two books.
| Step | Beyond the Numbers | Before the Numbers |
|---|---|---|
| Formulate the problem | Introduction, Chapter 1 | Introduction, Chapter 1 |
| Find what exists | Chapter 9 | Chapter 2 |
| Choose the design | Chapters 1 and 3 | Chapters 3 and 4 |
| Design the method and the forms | Chapters 2, 3, and 5 | Chapters 5 and 6 |
| Plan the sample | Chapter 4 | Chapter 7 |
| Collect the data | Chapters 2, 3, and 5 | Chapter 8 |
| Analyze | Chapters 6, 7, and 8 | Chapters 9 and 10 |
| Report | Chapter 9 | Chapter 10 |
The two books share one example. The clinic with missed appointments opens both, and Chapter 9 of Before the Numbers analyzes a survey of its patients. Written comments from a survey are a common meeting point. Open the comment column in the Text Explorer to search it, and code a sample of the comments in the Coding Workbench.