Broad questions often contain several separate claims that need individual attention. Beginners benefit from separating these claims early and keeping a clear record of where each one came from. This approach prevents later confusion about which piece of information supports which part of the original question and makes it easier to check accuracy without retracing steps unnecessarily.
The method also supports later verification when new sources appear or existing ones change. By treating each claim as a distinct item from the outset, researchers maintain control over the scope and avoid mixing unrelated details that could dilute the focus of the work.
Listing the separate claims inside a broad question
Begin by copying the question word for word into the working document. Read the text slowly and underline or highlight every statement that asserts a fact or relationship. Convert each highlighted section into a standalone sentence that could be examined on its own. This step reveals whether the original wording already mixes several ideas or leaves some claims implicit. A common error is to accept the question as a single indivisible unit; breaking it down shows that most broad enquiries contain between three and eight testable assertions. The list should stay faithful to the original phrasing at this stage so that later comparisons with sources remain exact. If a statement contains two connected ideas joined by “and” or “because”, split them into separate entries immediately. This prevents later disputes about which part of the combined sentence a source actually supports. The resulting list forms the skeleton of the research task and should be reviewed once before any sources are consulted.
Judging whether a sentence qualifies as a claim requires checking that it makes a verifiable assertion rather than posing a further question or expressing an opinion. Boundary cases often arise with conditional statements such as “if X occurs then Y follows”. Record the conditional clause as one claim and the consequence as another. This separation allows the researcher to test the condition and the outcome independently. The process works best when the original question is written in ordinary language; highly technical phrasing may need a brief plain-language gloss beside each claim, but the gloss must never replace the source wording.
Choosing a simple recording format
A table with three columns—claim, date noted, source code—provides sufficient structure for most beginners. Rows can be added without reformatting, and the layout remains readable when printed or viewed on a small screen. Numbered lists serve the same purpose when a table is impractical, provided each entry occupies its own line and the source code appears at the end of the line. The format should be chosen once and kept unchanged for the duration of the project; switching midway introduces transcription errors. Columns for page numbers or paragraph references can be added later if sources are long documents, but they are unnecessary at the first recording stage.
Common mistakes include using a word processor’s automatic numbering that restarts or using colour coding that disappears when the file is opened on another device. A plain text or basic spreadsheet format avoids these problems. The chosen format must allow easy sorting by source code so that all claims attached to one location can be reviewed together. This capability becomes useful when a single source is later found to be unreliable and every linked claim requires re-examination.
Assigning an identifier to each source
Each distinct location receives a short code such as S1 or S2. The code is entered once in a separate source list that contains the full URL, document title and access date. This list grows only when genuinely new locations are consulted. Reusing an existing code for repeated visits to the same page keeps the claim table short and prevents duplication. The identifier should be unique within the project; reusing codes from earlier projects risks confusion when files are archived together.
Judgement is required when a source appears in several versions, such as a webpage that is updated or a PDF that carries a revision date. Treat each version as a new source and assign a fresh code. The applicable condition is that the content visible to the researcher has changed in a way that could affect any claim. Minor typographical corrections usually do not require a new code, but alterations to dates, figures or conclusions do. Recording the access date beside the code allows later checks against archived copies if the page disappears.
Linking each claim to its originating source
Place the source code immediately after the claim text, separated by a consistent marker such as a square bracket. If a claim is supported by two locations, list both codes in the order they were consulted. This record shows the sequence of discovery and indicates whether the claim rests on a single source or appears across multiple independent ones. The link must be created at the moment the connection is noticed; postponing the step invites omission or incorrect attribution.
Errors commonly occur when researchers copy text from a source without noting the code at the same time, relying on memory to add it later. Memory is unreliable once several claims have been processed. The method works within projects of any size provided the source list remains complete; it becomes less practical when sources are oral interviews or private correspondence that cannot be assigned stable codes.
Checking that claims remain within the original scope
After the first batch of sources has been added, compare each recorded claim against the wording of the starting question. Claims that introduce new topics or extend the time period beyond the question’s limits should be marked “out of scope” rather than deleted. The mark preserves the trail in case the research question later widens. Scope checks are best performed after every five new sources rather than continuously, because frequent interruptions slow progress.
The judgement criterion is whether the claim, if verified, would help answer the original question as first stated. Claims that merely provide background context but do not address any part of the question fall outside this boundary. Beginners often over-include such material because it seems relevant; the scope check forces explicit decisions and keeps the task manageable.
Reviewing the log before expanding the search
Read the entire claim list and its attached codes in one session before moving to additional sources. Confirm that every code still corresponds to an accessible record and that no claim has been left without a code. This review also reveals duplicate claims that can be merged and gaps where a source code points to material that does not actually contain the stated information. Corrections made at this stage prevent error propagation.
The review is most effective when performed at the end of each working day or after a set number of sources. Its purpose is verification rather than analysis; deeper interpretation belongs to a later phase once the initial log is stable. The condition for stopping the review is that every claim has a valid code and every code points to retrievable material.
Continuing the log as new sources appear
New claims are added only when they directly support or contradict an existing entry. Insert the new claim directly beneath the related one and assign the next source code. This placement keeps related material together without requiring complex cross-references. The log grows incrementally rather than in large batches, which reduces the risk of overlooking connections.
A useful practice is to count the total number of claims and sources at the close of each session. The counts provide a simple progress indicator and highlight any sudden increase that might signal scope creep. The method remains practical as long as the project stays within a single document or spreadsheet; when the log exceeds several hundred entries, splitting it by sub-topic becomes necessary to maintain readability.
Once the initial log is complete, the researcher possesses a clear map of every assertion contained in the original question together with the precise locations that must be examined to test each one. This map can be handed to another person or returned to after an interruption without loss of context.
