Week 12 · Chapter

Chapter 12 — Evaluating Sources and Evidence

Critical Thinking Free sample edition

The module's primary reading · ~35–45 minutes · Objective 7 (with Objective 4 carried forward) · Read it first this week; the lecture, the slides, the tutorial, the quiz, and the Argument Workshop all build on it.

In plain language

This week is about where a claim comes from and whether its source has earned your trust. A source's credibility (how far it deserves trust on one claim) rests on four questions you can ask of any page, post, or person. You will learn to follow a claim back to its primary source (the original record). You will check what other people say about a site before you read it, and spot a real photo or quotation used out of its setting. You will also learn to treat a chatbot's citations and a familiar voice on the phone as things to verify, not things to trust on sight.

Why you already care about this

One evening you scroll past 3 posts in 10 minutes. A photo shows a street under brown floodwater, shared with the name of a city 2 hours from you (illustrative). A sales page for a sleep supplement says "a study proves it works" in 4 places and links nothing (illustrative). And a classmate pastes a chatbot reply into the group chat that cites a journal article with authors, a title, and a year, all in perfect format (illustrative).

Each post asks for your trust. Each one looks like evidence. Before Week 12 is over, you will know how to test all three, and you will find that each one fails once you check its source.

Last week you asked what a number in a claim says. This week you ask where a claim comes from, and whether the source behind it deserves your belief.

By the end of this chapter you will be able to:

judge a source's credibility on one claim with four questions, and tell relevant expertise from authority outside the field;
tell a primary source from a secondary one, and trace a claim back through its retellings to where it started;
leave an unfamiliar page to check what independent sources say about it before you read it closely;
check a photo or a quotation against its original context;
spot a fabricated citation or a cloned voice, and verify through a channel you already trust;
run the whole source check on a new claim and say the verdict in words.

Carry this line with you: trust is earned by the source, not by the format.

Where we are

Already covered: Week 1 taught you to find an argument, and Week 2 to put it in standard form. Week 7 taught the appeal to authority and when it becomes a fallacy. Weeks 9 to 11 taught you to judge a sample, a causal claim, and a number.
This week: you test the source behind a claim: its expertise, its record, its independence, and the evidence it shows, plus where the claim started and whether its context survived.
How they connect: every test from Weeks 9 to 11 assumes the data are real and reported fairly. This week checks that assumption. Week 7's appeal to authority returns as the first of the four credibility questions.
Where we're going next: Week 13 asks what would count as a test of a scientific claim, the step after asking where the claim came from.

Week 7 handed you the idea you lean on most. An appeal to authority is good inductive evidence when the authority has relevant expertise, the experts in that field broadly agree, and the authority has no serious conflict of interest. It becomes a fallacy when any of these fails. This week turns that test into a habit you can run on any source, from a news story to a chatbot.

1. Ask four questions: what makes a source credible

Picture two people telling you a storm is coming. One is a neighbor who saw a dark cloud. The other is the national weather service, which tracks storms for a living. You trust both a little, and one a lot more. That difference has a name.

Credibility is how far a source deserves trust on a particular claim, judged by its relevant expertise in that field, its track record of accuracy, its independence from the outcome, and whether it shows the evidence behind the claim.

Each part of that definition is a question you can ask.

Relevant expertise is training and experience in the field a claim belongs to; a source can have real expertise in one field and none in the field of the claim at hand.
A track record is a source's history of accuracy on earlier claims, including whether it corrects its mistakes openly when it gets something wrong.
Independence from the outcome is the condition of a source that gains or loses nothing from whether its audience believes a claim, which is missing when the source has a conflict of interest, such as a seller praising its own product.
The fourth question asks whether the source shows its evidence or only asserts the claim.

That makes 4 questions, and you ask them about one claim at a time. A doctor is a strong source on a medicine and a weak one on a used car.

Worked example — the actor (illustrative, called back from Week 7). "A famous actor says this herbal supplement cures colds, so it works."

Relevant expertise? Acting is not expertise in medicine. The source fails the first question.
What follows. The whole case rests on the actor's word, and the actor has no standing in the field of the claim.
The verdict. The passage commits the appeal to unqualified authority (ad verecundiam).

In words: acting is not expertise in medicine, so the source fails the first question, and the passage commits the appeal to unqualified authority (ad verecundiam).

Worked example — the weather service (illustrative, discussed and never keyed). A national weather service forecasts a storm's path, and an event planner cites the forecast as a reason to move an outdoor event indoors.

Relevant expertise? Forecasting storms is the weather service's own field. Yes.
Track record? Weather forecasts are checked against the weather every day, in public. Yes.
Independence? The service gains nothing from whether the event moves. Yes.
Evidence shown? Forecasts come with maps, tracks, and stated chances. Yes.

In words: the weather service passes all 4 questions on a question in its own field, so citing it is good inductive evidence, not a fallacy. You discuss this 1 contrast passage and key none.

This week keys no new passage from the course's fallacy set. The actor passage is 1 of the 19 passages keyed through Week 10, and the set holds 20, so 20 − 19 = 1 passage is still to come in a later week.

Misconception check — expertise travels. Students often think a person who is an expert at something is an expert in general. That is tempting, because real experts sound confident and use careful language. But expertise belongs to a field. A famous chemist talking about a sports trade, or a top athlete talking about tax law, is speaking outside the field. Ask whether the claim sits inside the source's field before you ask anything else.

Memory hook: ask four questions per claim; a famous name is not one of them.

2. Trace the claim: primary and secondary sources

Your roommate tells you that losing sleep hurts how you do on exams. You ask where they heard it. "A newsletter." Where did the newsletter get it? That second question is the start of a trace.

A primary source is the original record of an event, study, or statement, such as the study a news story reports on or the full text a quotation was taken from.

A secondary source is a source that reports, summarizes, or interprets a primary source, such as a news story about a study or a post that quotes a speech.

Tracing a claim is following a claim back through the sources that repeat it until reaching the primary source where it started, then checking that the original says what the retellings say.

Each retelling is a chance for a claim to change. A study about one group can become a headline about everyone. A finding that something "can" happen can become a claim that it "will".

Worked example — the sleep claim traced (illustrative). The running question's strongest objection says that losing sleep can hurt exam performance. A campus newsletter repeats the claim and cites a news story, and the news story cites a study.

List the chain. Newsletter, then news story, then study. The chain has 3 links.
Sort the links. The newsletter and the news story report on the study, so they are 2 secondary sources. The study is the original record, so it is 1 primary source.
Read the original. You open the study itself and check three things: that it says what the retellings say, that it is about the same kind of people, and that it claims the effect with the same strength.

In words: the newsletter is 2 steps away from the evidence, so you read the study itself before you rely on the claim, and you check that its people and its strength match the retelling.

The same move works on a number. The running question's cost premise, P4, says that keeping the library open overnight for finals week costs about 1,600 dollars.

Worked example — the cost premise traced (illustrative). Rebuild the figure from the staffing numbers it rests on.

Find the inputs. 2 staff work 8 hours a night for 5 nights at 20 dollars an hour.
Rebuild the figure. 2 × 8 × 5 × 20 = 1,600 dollars.
Compare. The rebuilt figure matches P4's 1,600 dollars.

In words: traced back to its inputs, the cost premise matches the staffing numbers it rests on, so P4's figure holds up. That does not settle whether 1,600 dollars is a "modest" cost; that is still a judgment for the argument to defend.

Misconception check — more sources means more evidence. Students often think a claim repeated on five sites is better supported than a claim on one. That is tempting, because repetition feels like agreement. But if all five sites copied the same news story, you have one source repeated five times. Tracing shows how many independent roots a claim has, and often there is only one.

An honesty note: you will not always reach a primary source. A study can sit behind a paywall, or a quotation can come from a speech no one recorded. When the trace stops, say where it stopped, and give the claim only the trust that the last source you reached has earned.

Memory hook: follow the chain until it stops; trust the link where it started.

3. Leave the page: what independent sources say

A sales page for a sleep supplement looks wonderful. It has clean fonts, photos of smiling people, a seal that says "clinically studied", and the line "a study proves it works". Nothing on the page tells you who wrote it, or whether the study exists.

The careful move is to stop reading the page and open new tabs. Before you trust an unfamiliar site, you leave it to check what other sources say about who runs it and how reliable it is.

An independent source is a source with no tie to the one being checked and no stake in the claim, such as a reference site or a news report about who runs a website, consulted by leaving the page before reading it closely.

Worked example — the product page (illustrative). A supplement's sales page says "a study proves it works" and links nothing.

Leave the page. You search for the site's name in a new tab, and independent sources describe the site as the seller's own store.
Independence? The seller profits if you believe the claim. The page fails the independence question.
Evidence shown? "A study" is asserted, not linked, named, or described. The page fails the evidence question.
Count. The page fails 2 of the 4 questions.

In words: the sales page fails 2 of the 4 questions, independence and evidence shown, so "a study proves it works" is an assertion you have no way to check, not evidence.

There is a real standard behind this. The Federal Trade Commission's guidance for advertisers says ads must be truthful and not deceptive, advertisers must have evidence to back up their claims before the ad runs, and health and safety claims generally need competent and reliable scientific evidence. An ad that says "a study proves it" is making a claim the advertiser is supposed to be able to back up.

Misconception check — a professional-looking site is credible. Students often trust a page that looks polished: good design, no typos, a seal, a ".org" or a lab-coat photo. That is tempting, because sloppy pages often are unreliable. But design costs money, not honesty, and anyone can buy a clean template. The look of a page answers none of the four questions. Leave the page and find out who is behind it.

A limitation: leaving the page tells you about the source, not about the claim. An independent source can tell you a site sells the product. It cannot tell you on its own whether the product works. You still need evidence you can check.

Memory hook: before you read it, find out whose it is.

4. Check the original: images and quotations out of context

A real photo can tell a false story. The flooded street from the opening is a real photo of a real flood. But suppose the flood happened years ago, in a different city, and a post now shares it with your neighbor city's name and the words "happening now". Every pixel is true. The caption is not.

Out of context is said of a real photo, video, or quotation shown apart from its original date, place, event, or surrounding words, so that it suggests something the original did not.

The check is the same for a photo and for a quotation: find the original, and read what surrounds it.

Worked example — the re-captioned photo (illustrative). A post shares the flood photo with a new city's name.

Look for earlier copies. You search for the image and find copies posted long before the post you saw.
Read the earliest caption. The earliest copy names a different city and a different storm.
The verdict. The photo is real, but the new caption moved it to a place and a time where it did not happen.

In words: the photo is authentic and the claim built on it is false, because the post changed its place and its time.

Quotations can be cut the same way, and a cut can flip the meaning.

Worked example — the shortened quotation (illustrative). A library staff member wrote, "Opening overnight would be a mistake unless the library adds staff for safety." A post about the running question quotes only "Opening overnight would be a mistake."

Find the full sentence. The original includes the condition "unless the library adds staff for safety".
Compare the meanings. The full sentence supports overnight hours on a condition. The cut sentence opposes them flat out.
The verdict. The quotation is accurate word for word, and it misleads in meaning.

In words: cutting the condition turns conditional support into flat opposition, so the quotation is accurate word for word and misleading in meaning.

This links back to Week 7. Quoting only the cut sentence and then attacking it answers a weaker version of the staff member's view, which is how a straw man often starts.

Misconception check — quotation marks mean accurate. Students often trust a quotation because it sits inside quotation marks. That is tempting, because quotation marks promise exact words. But exact words can be cut from a sentence that said something else. The marks tell you the words were said. They do not tell you what the whole sentence meant. Read the sentence around the quote.

Memory hook: a true picture can carry a false caption.

5. Verify what AI hands you: fabricated citations and cloned voices

Back to the chatbot reply in the group chat. It cites an article with two authors, a title, a journal, and a year. The format is perfect. You search for the article and cannot find it anywhere.

That happens because a chatbot produces fluent, confident text by predicting likely words, and a likely-looking citation is not the same as a real one. Chatbots can produce text that includes invented statistics, studies, quotations, and citations.

AI-generated misinformation is false or misleading content produced with an AI tool, such as a chatbot's invented statistic, study, quotation, or citation, or a cloned voice that imitates a real person.

A fabricated citation is a reference to a book, article, or study that does not exist, even though its authors, title, and year are set out in correct format.

Worked example — the chatbot citation (illustrative). A chatbot answer cites a journal article with authors, a title, and a year, all in correct format.

Search the journal. You search the named journal's own site for the title. No such article appears.
Search a library database. You search the title and the authors' names. No such article appears there either.
The verdict. The citation is fabricated, however real it looks.

In words: a chatbot's citation is unverified until you find the original, and this one has no original, so it is fabricated.

Voices can be faked too. Voice cloning is using an AI tool to imitate a particular person's voice from a short recording of it, so that a call or message can sound as if that person made it.

Worked example — the cloned voice (real guidance, taught in words). The Federal Trade Commission's 2023 consumer alert warned that scammers can clone a family member's voice from a short audio clip.

The setup. A call sounds like a relative in trouble, asking for money right away.
The advice. The alert advised calling the person back on a number you know is theirs before acting.
Why it works. A number you already trust is a channel the caller does not control.

In words: when a voice asks for money, the voice itself is not evidence of who is calling, so you verify through a channel you already trust.

The same habit covers the chatbot. Its answer is a lead, never a source. Check each figure, study, and quotation against the original before you repeat it.

Misconception check — a well-formatted citation is real. Students often trust a citation because it has everything a citation should have: authors, a year, a journal, even page numbers. That is tempting, because real citations look exactly like that. But a chatbot learned the format from a vast number of real citations, and it can produce the format without the article. The format proves nothing. Only finding the article does.

A caveat: none of this means a chatbot is always wrong. It means you cannot tell from the reply alone when it is right, so every claim it offers stays unverified until you check it.

Memory hook: a chatbot gives leads, not sources; a familiar voice is not proof.

6. Run the whole source check

You now have five moves. Put them together and you have one check you can run on any claim.

A source check is the set of moves for weighing where a claim comes from: ask the four credibility questions, trace the claim to its primary source, see what independent sources say, check the original context, and verify through a channel already trusted.

Worked example — a fresh claim (illustrative). A flyer on a community bulletin board says, "A local fitness coach says this protein bar cuts the risk of knee injury in half, according to research."

Ask the four questions. A fitness coach may know training, but knee-injury risk is a medical question. Track record unknown. If the coach sells the bars, independence fails. "According to research" names no research, so evidence is not shown.
Trace the claim. The flyer gives no study to trace. The trace stops at the flyer.
Leave the page. Search the coach's name and the bar's name to see who sells it.
Check the context. Look for the coach's full statement, if there is one, to see whether "cuts the risk in half" was the coach's claim at all.
Verify. Anything you find on the bar's own site, or in a chatbot's summary, still has to be found in an original study.

In words: the flyer fails the evidence question and the trace stops at the flyer itself, so the claim has earned very little trust until someone shows the research.

Notice what the check does not do. It tells you how much trust a source has earned on one claim. It does not prove the claim false. A source that fails one question can still be right, and a source that passes all four can still be wrong. The check is a tool for weighing evidence, not a machine that settles a case.

The same holds on the running question. Tracing the sleep claim and the cost premise tells you how well those premises are supported. It does not decide whether the library should stay open around the clock. That question is still yours to argue, from either side.

Misconception check — a failed check proves the claim false. Students often treat a source that fails the check as a source that lied. That is tempting, because the check feels like a verdict. But a weak source can repeat a true claim, and the actor might even be right about the supplement by luck. The check tells you the source gave you no good reason. To learn whether the claim is true, find a source that does.

Memory hook: ask, trace, leave, check, verify.

Key terms

credibility — How far a source deserves trust on a particular claim, judged by its relevant expertise in that field, its track record of accuracy, its independence from the outcome, and whether it shows the evidence behind the claim. (first used: 1. Ask four questions: what makes a source credible)
relevant expertise — Training and experience in the field a claim belongs to; a source can have real expertise in one field and none in the field of the claim at hand. (first used: 1. Ask four questions: what makes a source credible)
track record — A source's history of accuracy on earlier claims, including whether it corrects its mistakes openly when it gets something wrong. (first used: 1. Ask four questions: what makes a source credible)
independence from the outcome — The condition of a source that gains or loses nothing from whether its audience believes a claim, which is missing when the source has a conflict of interest, such as a seller praising its own product. (first used: 1. Ask four questions: what makes a source credible)
primary source — The original record of an event, study, or statement, such as the study a news story reports on or the full text a quotation was taken from. (first used: 2. Trace the claim: primary and secondary sources)
secondary source — A source that reports, summarizes, or interprets a primary source, such as a news story about a study or a post that quotes a speech. (first used: 2. Trace the claim: primary and secondary sources)
tracing a claim — Following a claim back through the sources that repeat it until reaching the primary source where it started, then checking that the original says what the retellings say. (first used: 2. Trace the claim: primary and secondary sources)
independent source — A source with no tie to the one being checked and no stake in the claim, such as a reference site or a news report about who runs a website, consulted by leaving the page before reading it closely. (first used: 3. Leave the page: what independent sources say)
out of context — Said of a real photo, video, or quotation shown apart from its original date, place, event, or surrounding words, so that it suggests something the original did not. (first used: 4. Check the original: images and quotations out of context)
AI-generated misinformation — False or misleading content produced with an AI tool, such as a chatbot's invented statistic, study, quotation, or citation, or a cloned voice that imitates a real person. (first used: 5. Verify what AI hands you: fabricated citations and cloned voices)
fabricated citation — A reference to a book, article, or study that does not exist, even though its authors, title, and year are set out in correct format. (first used: 5. Verify what AI hands you: fabricated citations and cloned voices)
voice cloning — Using an AI tool to imitate a particular person's voice from a short recording of it, so that a call or message can sound as if that person made it. (first used: 5. Verify what AI hands you: fabricated citations and cloned voices)
source check — The set of moves for weighing where a claim comes from: ask the four credibility questions, trace the claim to its primary source, see what independent sources say, check the original context, and verify through a channel already trusted. (first used: 6. Run the whole source check)

Summary

Credibility is judged one claim at a time, with four questions: field, record, stake, and evidence shown.
The actor fails the expertise question; the weather service passes all four in its own field.
A primary source is the original record; a secondary source retells it.
Trace a claim to where it started, and check that the original says what the retellings say.
Leave an unfamiliar page and see what independent sources say about who runs it.
A real photo or quotation can mislead out of context; find the original and read around it.
A chatbot's citation is unverified until you find the article; a familiar voice is verified by calling back on a known number.
The source check weighs trust in a source; it does not prove a claim true or false.

Check your understanding

Work these before looking at the answers — they are ungraded, and the answers (with reasoning) are at the bottom of the page.

A video by a popular car mechanic says a new allergy spray is "the best on the market" (illustrative). The mechanic is paid by the spray's maker and shows no test results. Which of the four questions does the video fail, and how many?
A community-garden flyer says a plot costs about 360 dollars a season (illustrative). The garden's own fee sheet lists 12 weeks in a season at 30 dollars a week. Trace the figure back to its inputs. Does it match?
A town council member said in a meeting (illustrative), "The new skate park is a great idea, but only if we can fund the lighting first." A post quotes the council member as saying, "The new skate park is a great idea." What happened, and what should a careful reader do?
You get a voice message that sounds like your older cousin, saying their wallet was stolen on a trip and asking you to send a gift card code right away (illustrative). What should you do first, and why?

Answers (worked reasoning — read after attempting)

It fails 3 of the 4 questions. Expertise: fixing cars is not expertise in allergy medicine, so it fails. Independence: the mechanic is paid by the maker, so it fails. Evidence: no test results are shown, so it fails. Track record is not something the video tells you either way. With 3 failures, the video gives you no good reason to believe the spray is the best.
12 weeks × 30 dollars a week = 360 dollars, so the flyer's figure matches its inputs. Tracing confirms the arithmetic. It does not tell you whether 360 dollars is a fair price; that is a separate judgment.
The quotation was cut, and the cut removed the condition "only if we can fund the lighting first". The full sentence gives conditional support, and the cut version sounds like full support. The careful reader finds the original, here the meeting record, and reads the whole sentence before repeating the quote.
Do not send anything yet. Call your cousin back on a number you already know is theirs, or reach them through another channel you trust. A voice can be cloned from a short clip, so the voice alone is not evidence of who sent the message, and a number you already know is a channel the sender does not control.

Next up in this module: the lecture, the slide deck, the Readings & Resources page, the lecture tutorial, the practice set, and the quiz. Next week, in Science and Explanation, the source check meets the test that asks what evidence could prove a claim wrong.

Built with The Course Maker · samples.thecoursemaker.com