The Problem with Research
Research is the part of writing a book that feels most like work and least like writing. It is also the part most likely to expand indefinitely. There is always one more source to track down, one more archive to check, one more expert who might have something useful to say. Left unchecked, the research phase becomes a form of productive procrastination — a way of feeling busy without committing to the harder work of making an argument.
I have written enough research-heavy books to have developed a system that keeps the research phase bounded and purposeful. It is not a system designed to find everything — that is impossible. It is a system designed to find enough: enough to make a credible argument, enough to answer the questions the outline raises, enough to write with confidence.
Starting with Questions, Not Sources
The most important shift I made in how I research was moving from source-first to question-first. Early in my career, I would begin by reading everything I could find on a subject and hoping that a structure would emerge from the accumulation. It rarely did. What emerged instead was a pile of notes with no clear relationship to each other and no obvious path toward a book.
The question-first approach starts with the outline. Before I open a single source, I have a premise statement and a chapter map — both described in How I Outline a Book [blocked]. The chapter map tells me what questions each chapter needs to answer. Those questions become the research agenda. I am not reading to learn everything about a subject; I am reading to answer specific questions in a specific order.
This sounds obvious, but it changes the experience of research substantially. Instead of reading a source and asking "what is useful here?", I am reading it and asking "does this answer any of my questions?" The answer is usually no, which means I can move on quickly. When the answer is yes, I know exactly where the material belongs in the structure.
Primary and Secondary Sources
For most of the books I write, the research involves both primary and secondary sources, and the relationship between them matters.
Secondary sources — books, articles, and reviews written by other scholars — give me the shape of the existing conversation. They tell me what has been argued, what is contested, and where the gaps are. They are also where I find leads to primary sources: the footnotes and bibliographies of good secondary sources are among the most valuable research tools available.
Primary sources — original documents, letters, interviews, archival materials — are where the book's distinctive contribution comes from. Any writer can summarise what other scholars have said. The work that makes a book worth reading is the work done with primary sources: the letter that contradicts the received account, the interview that reveals something the subject never said publicly, the document that changes the interpretation of everything that came before it.
For Searching for Bowlby, the primary sources included Bowlby's own published papers, unpublished correspondence held at the Wellcome Collection, and interviews with people who had worked with him or been influenced by his ideas. The secondary literature on attachment theory is substantial, but the book's argument depended on the primary materials — on what Bowlby actually wrote and said, rather than on what others had concluded about him.
The Scrivener Research Folder
All of this material — notes, PDFs, web clips, interview transcripts, photographs of archival documents — lives in the Scrivener research folder for each project.
Scrivener allows you to import almost any file type directly into the project. PDFs open in the built-in viewer. Web pages can be imported as formatted text or as web archives. Images display inline. The result is that every piece of research material is accessible from within the same application as the manuscript, without switching between windows or hunting through a file system.
I have written a full review of Scrivener that covers the research folder and other features in detail: Scrivener Review: Why Every Serious Author Needs This Writing Software [blocked]. The short version is that having the research and the draft in the same place changes how you write. When you need a specific detail, you click to the research folder and it is there. You do not lose the thread of the sentence you were writing while you go looking for it.
Within the research folder, I organise materials by chapter. Each chapter in the Binder has a corresponding folder in the research section. When I find a source that answers a question in chapter four, it goes into the chapter four research folder. By the time I start writing a chapter, everything I need is already gathered in one place.
Note-Taking: The Two-Pass Method
I take notes in two passes. The first pass is fast and permissive: I read a source and capture anything that seems potentially relevant, with a note of where it came from. I am not evaluating at this stage — I am collecting. The goal is to get through the source quickly and extract what might be useful without spending time deciding whether it is actually useful.
The second pass is slower and more selective. I go through the first-pass notes and ask, for each item: does this answer one of my research questions? If yes, I move it to the relevant chapter folder and add a note about how it fits into the argument. If no, I delete it or move it to a "miscellaneous" folder that I rarely look at again.
This two-pass approach keeps the research phase moving. The first pass is fast because it requires no judgment. The second pass is where the judgment happens, but by then I have the full picture of what a source contains, which makes the judgment easier.
On Writing Well by William Zinsser has a useful observation about research and writing that I return to often: the writer's job is not to report everything they found, but to select what serves the reader. The research folder is where everything lives. The manuscript is where the selection happens.
Knowing When to Stop
The hardest part of research is knowing when you have done enough. There is always more to find. The archive has more boxes. The bibliography has more footnotes. The subject has more dimensions than any single book can cover.
The answer I have arrived at is this: you have done enough research when you can answer all the questions in your outline, and when additional research is producing diminishing returns — when you are finding material that confirms what you already know rather than material that changes or complicates the argument.
This is not a precise threshold, and it requires judgment that only comes with experience. But the outline is the key instrument. When every question in the chapter map has an answer, and when the answer is supported by more than one source, the research phase is complete. The writing can begin.
The outline and the research are not sequential phases — they run in parallel, each informing the other, as I described in How I Outline a Book [blocked]. But there comes a point when the outline is stable and the research questions are answered, and at that point the only thing left to do is write.
