Separate Relevance From Credibility, and Keep a Source Log
In the last chapter you traced a claim back to the original source worth citing. But reaching the real source still leaves one question hanging: should this source make it into your answer at all? Two completely different things decide that, and beginners almost always squash them into one. This chapter pulls them apart, then hands you a simple habit, a source log, to keep track of everything you find without it turning into a pile of forgotten tabs.
Tell Relevance and Credibility Apart
Picture your search returning two results for the scoped question you built earlier: “Does a magnesium supplement help healthy adults fall asleep faster?”
The first is a careful, well-run scientific study, about magnesium levels in farm soil. Rock solid, and completely useless to you. The second is a forum post from a stranger called “SleepyJoe92,” and it is exactly about magnesium supplements and sleep, but you have no idea who they are or whether to believe a word they say.
Those two sources fail you in opposite ways, and naming the difference is the whole skill here.
Relevance is how well a source actually fits your scoped question: the who, the what, and the outcome you defined when you framed it. A relevant source answers the question you are actually asking, not a neighbouring one.
Credibility is how much you can trust what a source says: who made it, how they know it, and whether they have any reason to mislead you.
The trap is that these two are completely independent. Any source can be high on one and low on the other, which gives four combinations.
Relevant and credible: the keeper. This is what you are hunting for.
Relevant but not credible: the seductive trap. On-topic, so it feels right, but you cannot trust it yet (the forum post).
Credible but not relevant: the distraction. Trustworthy, but it answers the wrong question (the soil study).
Neither: discard it and move on.
If a source is clearly trustworthy, isn’t that good enough?
No, and this is exactly where careful people slip. A flawless study about the wrong question is still about the wrong question. Trust does not create relevance, and relevance does not create trust. You have to check both, separately, every time.
Relevance you can judge right now, against the scope you set before searching: does this source address that exact question, yes, partly, or no? Credibility runs deeper, who is behind it, how solid is their evidence, what do they want from you.
Keep a Source Log
As soon as you are weighing more than two or three sources, memory quietly fails you. You forget which page made which claim, which one was just a headline, and which you already decided to drop. The fix is small and boring and it works: a source log.
A source log is a simple running record of every source you seriously consider, holding the few facts you need to compare them now and cite them later. It pulls together everything you have learned in this part. For each source you keep, record a short row:
Source and link, so you can find it again.
What it claims, the one finding that matters to your question.
Type, its two labels from earlier: provenance (primary or secondary) and intent (informational, opinion, or promotional).
Original source, the primary source you traced it back to last chapter, not the headline that led you there.
Relevant?, whether it actually addresses your scoped question: yes, partly, or no.
Trust notes, a quick first read on credibility, plus anything you want to verify later.
Date checked, since sources and facts go stale.
Three habits make the log worth keeping. Log a source the moment you decide it is worth keeping, not in a panic at the end. Always record the original you traced to, not the messenger that pointed you there. And mark relevance honestly, even for a source you were hoping to use.
Isn’t keeping a log just extra busywork on top of the research?
It feels that way for about five minutes, until you have fifteen tabs open and cannot remember which one actually had the data. The log is what turns a messy pile of sources into something you can compare at a glance, and it is the exact artifact your project asks you to produce.
Watch a Source Log Come Together
Let’s fill a few rows for a real question. Suppose your scoped question is the civic one from earlier: “In our city, did adding protected bike lanes reduce cyclist injuries over the last five years?” Three sources come back, and watch how relevance and credibility split them apart.
Source 1, the city road-safety statistics portal. It publishes the city’s own injury figures and shows cyclist injuries falling on the streets that got protected lanes.
Claim: injuries dropped on the lane-equipped roads over the period.
Type: primary, informational.
Original source: this is the original, the first-hand dataset.
Relevant? Yes, it matches the scope exactly.
Trust notes: official data, high first impression.
Decision: keep, and cite this as the evidence.
Source 2, a cycling-advocacy blog post headlined “Our lanes are a huge success, share this!”
Claim: the lanes dramatically improved safety.
Type: secondary, opinion (an interested party arguing a position).
Original source: it leans on the same city report, compressed into a slogan.
Relevant? Yes, it is on topic.
Trust notes: a campaigner who wants this result; useful as a lead, but verify every number against the primary data before trusting it.
Decision: keep as a lead, not as evidence.
Source 3, a respected national study on bicycle-helmet effectiveness. Carefully done, widely cited, genuinely trustworthy.
Claim: helmets reduce head-injury severity in crashes.
Type: primary, informational.
Original source: itself.
Relevant? No. It is about helmets, not protected lanes. Different question.
Trust notes: highly credible, simply not about your scope.
Decision: drop it, despite how solid it is.
Look at what the log just did. Source 3 is the most credible of the three and still gets cut, because credibility could not save it from being irrelevant. Source 2 is perfectly relevant and still cannot be cited, because relevance could not make it trustworthy. Only Source 1 passes both tests. That is the four-way split from the first section, working on real sources.
Your Turn!
Time to keep a short log of your own.
The situation. You are researching a consumer question you scoped earlier:
“Do phone cases sold as ‘compostable’ actually break down in a home compost bin within about a year, according to independent testing?”
Two sources come back:
A report from an independent consumer-testing organisation that put several “compostable” cases through home-composting conditions for twelve months and published what happened.The case manufacturer’s own product page: “100% compostable, eco-friendly!”, with a “buy now” button and a line saying “lab certified.”
Your task. Build a two-row source log. For each source, record: source and link, what it claims, its type (both axes), the original source it traces to, whether it is relevant to your scoped question, and a quick trust note. Then make a call on each: keep and cite, keep but verify, or drop, with one line saying why.
Deliver: a two-row log plus your keep / verify / drop decision for each. Keep it, this is exactly the Source Map you build on in your project.
Then check your answer against the model solution found at the end of this chapter.
Let’s Recap!
Relevance is whether a source fits your scoped question; credibility is whether you can trust what it says, and the two are independent.
The most common trap is treating an on-topic source as automatically trustworthy.
A relevant but untrustworthy source and a trustworthy but irrelevant one both fail you, for opposite reasons.
A source log records, for each source, what it claims, its type, the original source you traced to, whether it is relevant, and a quick trust note.
Always log the original you traced to rather than the headline, and judge relevance against the scope you set before you searched.
You have now reached the end of the course. Look back at what you can do: you can take a vague curiosity and turn it into a sharp, researchable question, set its scope and success criteria, and plan a search that actively tries to prove you wrong. When the results come back, you can name what each source is on both axes, trace any claim to the original worth citing, and tell whether a source is relevant apart from whether it is credible, keeping the whole picture in a source log. Now head to one more short quiz to check your understanding.
Model Solution
There is no single perfect wording, but a strong log makes its decisions visible. Here is one good version.
Source 1, independent testing report.
Claim: most “compostable” cases tested did not fully break down in home composting within a year.
Type: primary, informational (it ran and reports its own test).
Original source: itself, the test is the first-hand record.
Relevant? Yes, it matches the scope exactly: compostable cases, home conditions, about a year, independent testing.
Trust notes: independent, not selling anything, shows how it tested. High first impression.
Decision: keep and cite. Relevant and credible.
Source 2, manufacturer’s product page.
Claim: the case is “100% compostable.”
Type: secondary and promotional (it leans on a certification it did not run, and its job is to sell).
Original source: not here, you would have to trace the “lab certified” claim back to the actual certification or standard before trusting it.
Relevant? Yes, on topic.
Trust notes: the seller has a clear interest, drawbacks are missing, and a slogan is not a test result.
Decision: keep but verify at most, and never cite the sales page itself as the evidence. Relevant, but not yet credible.
If your log named both axes, recorded the original (or noted that you would need to trace it), judged relevance against your scope, and left a trust note, you have got the habit. That habit is what you carry straight into your project’s Source Map.
