A sentence can look well supported because it contains a link. The useful question is what happens after you open that link. Does the destination address the same subject? Is it the original document or another article repeating the statement? Does the source support the sentence as written, or only a narrower version of it?
For an editor working on CBD-related information, a source trail is a practical way to answer those questions. The result is a small record that connects a published statement to the document behind it. This guide focuses on that editorial task. It does not evaluate whether a product or treatment is appropriate for an individual.
Copy the exact sentence first
Start with the wording you need to check. Put the sentence in a working document before opening a series of browser tabs. If the wording changes during the review, keep the original beside the revision so the reason for the change remains visible.
Break a compound sentence into separate claims. A sentence may identify a substance, describe a study and suggest an implication for readers. Those are different statements, and one source may support only one of them. Give each statement its own row in the working record.
Write the question you are checking in plain language. “Does this document say what the article says it says?” is often a better starting point than a broad search for more information about the topic. The narrower question keeps the research from becoming an open-ended reading session.
Identify what kind of source you have
Open the linked page and record its publisher, title and date. Then classify the document. It may be an agency overview, a study, a review article, a press release, an opinion piece or a commercial page. Each can tell you something about its own contents, but the types should not be treated as interchangeable.
For example, the FDA's cannabis and CBD information page explains the agency's regulatory role and positions. When citing it, describe it as agency information. Calling that page “a study” would change the nature of the evidence even if the link itself were correct.
If the destination is a news story, look for its link to the original document. Follow that link and record both steps. The news story may be useful for context, but the source trail should make clear when the statement is based on another writer's summary.
Find the passage that does the work
Search within the document for the term or phrase connected to the claim. Read the surrounding paragraph, then the relevant section. A sentence removed from its context can appear broader than the author intended.
Record a short location note: section heading, table number, page number or another stable marker. A colleague should be able to reach the same passage without repeating the entire search. Keep quotations short and use them only where the exact wording matters to the check.
If you cannot locate a supporting passage, mark the claim as unresolved. Do not fill the gap with a second source that merely repeats the same unsupported statement. The working record is useful precisely because it can show where the trail stops.
Compare the scope of the sentence and source
Now compare the proposed sentence with the document. Is the article talking about the same population, substance, setting and outcome? Does it describe a possibility as a settled result? Does it turn a limited observation into a statement about every product or every reader?
NCCIH's guide to reading a scientific journal article provides a useful introduction to the structure of a paper. When a sentence rests on a study, the short abstract is a starting point. The methods, results and discussion help explain what the authors actually investigated and how they describe its limits.
Write a one-sentence account of the match. For example: “This document supports the description of the agency's stated position, but it does not establish a clinical outcome.” That note is more useful to an editor than a vague label such as “good source.” It explains what the citation can do in the article.
Work through a simple example
Consider a fictional draft that says, “A government report proves that every CBD product has the same effects.” The link leads to a general public-information page. The first problem is the word “proves.” The second is the move from a general source to every product. The third is the absence of a specific effect or study in the sentence.
The checker should identify those problems separately. The source record might say that the destination is an overview, that the draft does not identify a study and that the universal product statement is unsupported by the cited passage. The editor can then remove the sentence or replace it with a narrower, source-supported description relevant to the article.
The purpose of this example is the editorial method. It shows how a broad claim can be divided into questions that can actually be checked. The same process works for less dramatic wording, where the difference between a source and a paraphrase is easy to miss.
Record the result so someone else can review it
A compact source record can contain seven fields: proposed statement, source title, publisher, date, direct link, relevant location and editorial decision. Add the name of the checker and the date of the check. If a qualified reviewer is needed, identify that next step instead of marking the statement complete.
The editorial decision should be specific. “Retain with attribution,” “narrow to the stated population,” “replace the citation” and “remove pending support” describe actions. A general pass or fail label often leaves the writer guessing what to do next.
Keep the record connected to the draft version. If a later edit changes the meaning of the sentence, the earlier source check may no longer apply. A short note beside a changed claim can prevent an old approval from being carried forward automatically.
Finish with a second-reader check
Give the record to another editor and ask them to follow the trail. They should be able to find the source, locate the relevant passage and understand why the wording was retained or changed. If they cannot, improve the record while the research is still fresh.
The final check is to read the revised paragraph as a reader would. Accurate individual sentences can still create a misleading impression when placed together. Check the heading, caption and surrounding transitions as well as the sentence that started the review.
Begin with one claim from a current draft and complete the record before checking the next. A small, reproducible source trail is more useful than a large folder of links whose purpose only one person remembers. Over time, those records become a practical foundation for the publication's review and correction work.