Blog · November 15, 2026

How I Turn a Research Question into a Useful Reading Note

A reading note is not a miniature version of a book. It is a record of what a source can help me answer, what it actually says, and what I still need to find out.

I begin a reading note with a question, not a summary. That distinction sounds modest, but it changes the work: instead of trying to preserve everything I have read, I make a small, deliberate record of the source’s claims, its evidence, and the pages where I found them. Later, when I am comparing sources or drafting a piece, I want the note to return me to the argument—not merely to the agreeable sensation of having read it.

My basic structure has three columns: claim, evidence, and question. I add a page number to every entry, and I write down the research question the source might help answer. The method is plain enough to fit on a single sheet of paper. Its usefulness comes from the distinctions it forces me to make.

Those distinctions matter most after the first read. A familiar subject can make separate arguments seem more alike than they are. Two authors may use the same word for different things; one may offer a measured result while another builds a case from archival records or interviews. A note that records only “what the text is about” smooths over these differences. A note that records claims and evidence leaves them visible.

Start with a question small enough to travel

Research questions often arrive in generous clothing. “How does place shape opportunity?” might be a promising opening, but it is not yet a useful guide to a reading session. It contains several questions: What kind of place? Which opportunity? For whom, and over what period? A source could touch the subject without helping answer any of those questions in a way I can later use.

Before I read closely, I make the question more specific. I might ask, for example, How do studies of neighborhood access define proximity to public transit? That is not necessarily the final question for a project. It is a working question—a temporary handle on the material. It gives me something to look for while leaving room for the source to complicate my assumptions.

A good working question does not have to contain the whole project. It has to be narrow enough to help me decide what belongs in the note. If a chapter concerns transit funding but says nothing about how proximity is defined, that may still be useful, but I will know why it belongs: perhaps it offers context, a counterexample, or a question for another source. The note should make that reason legible.

I write the question at the top of the page, often with a short phrase beneath it: “definition,” “method,” “historical context,” or “possible counterexample.” These labels are not a filing system for the ages. They tell me what I am trying to learn today. A question can change as I read. When it does, I revise it rather than pretending that the original wording still fits.

The three columns

The table is simple by design. I want a structure that can survive a dense article, a chapter with a long argument, or an untidy afternoon when my attention keeps wandering. Three columns are enough to slow me down without turning note-taking into a second research project.

Claim Evidence Question
What does the author argue that I may need to recall? What material or method supports that claim? What remains uncertain, limited, or worth comparing?
Record the claim in my own words, without making it broader. Name the evidence precisely and note the relevant page. Write a real question, not a vague reaction.

Claim is not a topic label. “Public transit” tells me what a passage concerns; it does not tell me what the author says. A claim is closer to: “The study treats access as a measure of travel time rather than distance.” I phrase it in my own words, but I keep the scope and qualifications of the source intact. If the author argues that a pattern appears in one city, I do not quietly turn it into a claim about cities in general.

Evidence answers the question, “What is the claim resting on?” The answer might be a dataset, an archival collection, a close reading of policy language, a set of interviews, an experiment, or a comparison between cases. “The article uses research” is not an answer. Neither is “the author proves this.” I want to know what the author examined and, where possible, how.

Question is the space I reserve for what the source does not settle. It might be about a definition, a boundary, a comparison, or a consequence. “Interesting” is a response, not a question. “Would the same definition of access produce a different result in a low-density area?” gives me a next step. It may send me to another source, or it may reveal that I need to revise my working question.

Put the page number beside the thought

I record page numbers as I go, rather than planning to add them later. Later has an alarming talent for arriving after the library session, the browser tab, and the precise context of a quotation have all disappeared. Page numbers are not decoration. They are the route back to a claim, a passage, or a qualification that matters when I begin writing.

For a useful point, I note the page where it appears and, when the argument unfolds over several pages, the range. If the source is in a format without stable page numbers, I record the most precise locator available: a section heading, chapter, paragraph number, or timestamp for a talk. The principle is the same. A note should let me find the passage again without asking my memory to do the indexing.

I distinguish a paraphrase from a direct quotation. A paraphrase goes in my own language, followed by a page number. A quotation is marked as a quotation, copied exactly, and accompanied by the page. That visual difference helps prevent a familiar drafting mistake: returning to an old note and no longer knowing whether the polished sentence is mine or the source’s.

When exact wording is especially important—perhaps because a term is being defined or the author’s phrasing carries the argument—I copy only the necessary passage. I do not collect quotations because they sound good in isolation. A sentence can be elegant and still be irrelevant to the question I am trying to answer.

A worked example, with the limits kept visible

Suppose my working question is: How do studies of neighborhood access define proximity to public transit? I am reading a hypothetical article that compares two measures: straight-line distance to a stop and estimated walking time. The example is invented to show the structure, not to report a finding from a real study.

My note might contain an entry like this:

  • Claim: The article treats walking time as a more behaviorally relevant measure of proximity than straight-line distance.
  • Evidence: The authors compare both measures in their analysis; note the specific method and results on pp. 6–8 of the hypothetical article.
  • Question: Does the walking-time measure account for street crossings, hills, or barriers between a home and a stop?

Even in this small example, the columns do different work. The claim preserves the argument without pretending that the article has settled every question about access. The evidence records the comparison the authors made, rather than merely declaring the article “quantitative.” The question identifies a possible limit. If the source does not say whether the route calculation accounts for barriers, I do not fill the silence with an assumption. I mark the uncertainty and look for the answer elsewhere.

The page range in this demonstration is fictional, too. In an actual note, I would replace it with the exact locator in the source I had read. That small discipline protects the difference between an example and evidence. A worked example can show how a method functions; it cannot stand in for a publication or be cited as one.

What I leave out

I rarely need a chapter-by-chapter account of a source. If a sequence of sections matters, I will record it. But an exhaustive recap often gives me a reassuring amount of text without telling me what I can do with it. When the time comes to write, I do not need a second version of the book. I need a reliable map to the parts that bear on my question.

I also leave out reactions that have not yet become useful. “This is confusing” may be true, but it does not tell me what kind of confusion I have encountered. I try to sharpen it: Is the central term undefined? Does the evidence seem to support a narrower claim than the author makes? Have I lost track of which population a number describes? The more exact question may expose a gap in my understanding—or a genuine ambiguity in the source.

Nor do I use the note to settle an argument prematurely. A source that seems to support my view may rely on a different definition from the one I am using. A source that appears to contradict another may simply be studying a different place, period, or population. The note should preserve those conditions. If I erase them to make a neat pair of opposing conclusions, I will have made the literature easier to summarize and harder to understand.

Compare methods before comparing conclusions

Once I have notes from several sources, I can compare what they say—but I try to compare their methods and definitions first. Imagine two studies that both address access to transit. One measures distance to the nearest stop; another estimates travel time along a walking route. Their conclusions may differ, yet the difference might begin in what each one calls “near.” Without that detail, I might describe them as disagreeing when they are measuring different things.

This is where the three columns become more than a tidy page. The claims give me the arguments; the evidence column shows how those arguments were built; the questions reveal where one source might meet or challenge another. I can sort entries by a shared term, then ask whether the term means the same thing in each source. I can also see where a claim rests on a method unlike the one used in a neighboring study.

Comparison is not a contest in which the newest source automatically wins. A historical account, a statistical analysis, and a close reading may answer different kinds of questions. The point is not to force them into a single scale, but to understand what each one can establish. A careful note keeps the source’s contribution attached to the form of evidence that made it possible.

The note is for the future reader

When I reread a note, I am reading it as a version of myself who has forgotten the article’s structure and may not remember why the source seemed useful. That future reader needs more than a fragment such as “access—walking time.” They need the question, the claim, the evidence, and a location in the text. A few extra words now can save a long, fruitless search later.

I include enough bibliographic information to identify the source: author, title, publication details, and date, where available. The citation does not need to occupy the three-column table, but it belongs with the note. Otherwise, a precise page reference can become an orphaned coordinate—accurate, perhaps, but impossible to return to.

Digital notes and handwritten notes can both work. I have no interest in turning one format into a moral position. What matters is whether I can retrieve the note, connect it to the source, and distinguish my words from the author’s. A folder full of files named “article notes” is not a method. A paper notebook can become just as opaque if it has no index or identifying details.

A short routine that keeps the structure useful

My reading-note routine has four moves. First, I write the working question and identify the source. Second, I read for the argument rather than transcribing as I go. Third, I record only passages that help answer the question, with claims, evidence, page numbers, and a question of my own. Finally, I spend a few minutes after reading checking that I can tell what each entry means without reopening the source.

That final check is often the most revealing. If an entry says only “important finding,” it has not done its job. If the claim is larger than the evidence I recorded, I narrow it. If I cannot tell whether a sentence is quotation or paraphrase, I mark it clearly. If I have written a question that the source already answers, I revise the note rather than letting the mistake travel into a draft.

The routine is not a guarantee that I will read perfectly or remember everything. It is a way to make forgetting less expensive. I do not expect a note to contain the whole intellectual life of a source. I want it to hold onto the few things I may need again, while preserving enough context to use them responsibly.

From notes to a draft

When I begin writing, I do not simply turn each note into a paragraph. First I arrange notes around problems, terms, or disagreements. A cluster might reveal that several sources use the same word differently. Another might show that a question has been studied through one method but rarely through another. The notes help me see those patterns, but the argument in the draft has to be made afresh.

Page numbers make that transition safer. When I am ready to cite a claim, I can return to the source and confirm its wording, context, and scope. My note is a guide, not a substitute for checking the evidence. This matters especially when a claim has become convenient—when it fits a paragraph so neatly that I am tempted to remember it more confidently than the source warrants.

The question column helps here, too. It keeps unfinished work in view. A question may become a paragraph’s hinge, a reason to find another source, or a sign that the draft is trying to claim more than the reading can support. Sometimes the most useful outcome of a note is not a fact I can quote but a boundary: this source does not answer that question.

I think of a reading note as a small promise to my future self: I will not ask you to trust a vague memory when the page can be found; I will not confuse a source’s evidence with my interpretation; I will leave a mark where the argument still needs work. The promise is manageable because the note is small. One question, three columns, and exact page numbers are usually enough to make the next encounter with a source more intelligent than the last.

Theme