Lecture: Research: Finding & Evaluating Sources
Lecture script · about 15–20 minutes spoken · Week 9
Opening
Think about a nursing student checking a claim about a new treatment. Or a small-business owner reading a market report before applying for a loan. Or a voter looking at a link a neighbor shared. Three different situations, and the same question sits under all of them: does this source deserve my trust before I act on it? That question is what this lecture is about.
Here's where we are. In Week 1 you learned that every text has a writer, an audience, and a purpose. In Week 7 you named the running draft's grounds, the research on schools that banned phones. Last week, in Week 8, you added the three-sentence habit: what a text says, how it persuades, and whether you agree, each in its own sentence. This week the text in front of you is a source you might use. So a new question comes first: should you use it at all? Next week, in Week 10, you'll bring the sources you choose into your own sentences.
Most of my examples come from the running draft, the student proposal about phones in lecture classes that this course keeps revising. It has no named author. This week it gets its sources: four real ones, plus one website I made up so we can practice on it.
1. Turn a topic into a research question
Imagine walking up to a librarian and saying one word: "phones." You'd get a shelf of books on phone design, phone bills, and phone history. Now imagine asking a question instead. You'd get a short list.
That's the difference between a topic and a research question. A topic names a subject and asks nothing. A research question is a focused question that evidence can answer, that asks more than a trivial yes or no, and that is arguable enough to lead to a thesis. A good one tells you what to search for and what to skip.
Let's run the running draft. Its topic is the phone policy in lecture classes. Too broad would be "How do phones affect education?" That covers every grade and every habit, so no essay can answer it. Too small would be "Do students bring phones to lecture?" You can answer that by looking around the room. Here's the research question the draft uses: "When schools restrict phones during class, what happens to students' learning, and which students gain the most?" It's focused on one change. Studies and reports can answer it. It asks what happens and to whom. And it leads straight to the draft's working thesis.
Pause and try this. "Do students bring phones to lecture?" Why isn't that a research question?
Think about how you'd answer it.
Here's the answer. One look around the room settles it. A question a glance can answer leads to no thesis, so it can't drive research.
2. Sort sources by what they are
Think about a concert. The recording of the show is the thing itself. A newspaper review of the show is writing about it.
That's the split between primary and secondary. A primary source is the thing itself that a writer studies: an original study, a speech, a law, or a set of official guidance. A secondary source is writing about the thing itself, such as a news report on a study or a book that discusses many studies.
Here's the twist. The same item can be either, depending on how you use it. The running draft has a news article by Jeff Amy, published by PBS News, about the majority of U.S. states that now have laws banning or regulating cellphones in schools. If you want the laws it reports, it's secondary. But suppose your research question were how news outlets covered those laws. Then the article is the thing you're studying, so it's primary. Same article, same words. Only your question changed.
Pause and try this. Eisenhower's farewell address, from 1961, sits in front of you. You're writing about what the address argued. Primary or secondary?
Hold your answer for a second.
Here's the answer. Primary. You're studying the address itself, so it's the thing, not writing about the thing.
3. Sort sources by who they are written for
Picture three readers interested in heart care. A cardiologist reads a medical journal. A parent reads a health magazine in a waiting room. A hospital purchasing manager reads a magazine about supplies. Same subject, three audiences, three very different publications.
A scholarly source is written by specialists for specialists and passes peer review before it is published. Peer review is the check in which other experts in the field read the work and judge its methods and claims. A popular source is edited for general readers. A trade source is written for people who work in one field, such as a magazine for school leaders.
None of those labels means good or bad. Each one tells you what kind of evidence you're holding. In the running draft, the article by Beland and Murphy, in the journal Labour Economics, from 2016, is scholarly. The news article is popular. Nicholas Carr's book, The Shallows, in its tenth-anniversary edition from 2020, is popular too. And the fourth source, the Department for Education's guidance page on GOV.UK, is neither. It's an official document. So the count is one scholarly, two popular, and one government document: four sources.
Pause and try this. Which of the draft's four sources passed peer review before it was published?
Run through the four in your head.
Here's the answer. The Beland and Murphy article. It's the scholarly one, so other experts checked its methods before it appeared.
4. Know where you are searching
Imagine two stores. One sells only items its buyers have inspected. The other is a giant flea market where anyone can rent a table. You can find good things in both, but you'd shop them differently.
A library database is a searchable collection, reached through a library, that screens what it indexes, so its results are already a filtered set. The open web is everything a general search engine can reach. It's unfiltered, because anyone can publish there and pages are ranked by many signals rather than by how credible they are.
That ranking is worth a moment. The top result is the most optimized page, not the most credible one. A page can climb because someone paid for an ad, or used popular search words, or got a lot of clicks. Rank tells you about attention, not truth.
For the running draft, you reach the Beland and Murphy article through a library database, and you find Carr's book through the library catalog. The news article and the guidance page came from the open web. Two through the library, two from the open web, four in all.
Pause and try this. A page sits at the very top of your search results. What does its rank tell you about its credibility?
Give it a moment.
Here's the answer. Nothing by itself. Rank reflects optimization and clicks, so you still have to judge the page on its own merits.
5. Run the five-question checklist
When you hire a mechanic, you probably ask a few questions. How long have they done this? Do they work on your kind of car? Do people recommend them? You're running a checklist without naming it.
The credibility checklist is a five-question test for judging a source: currency, relevance, authority, accuracy, and purpose. The version this course teaches is the CRAAP test, named for the first letters of those five questions, from Blakeslee's 2004 article, "The CRAAP Test". It's one common checklist among several, and like every model in this course, it's a simplification you can hold in one hand.
Here's the news article, run through the five. Currency: it reports recent state laws, so its date matters. Relevance: it covers schools, not college lecture halls, so it fits only part of the question. Authority: it carries a named reporter and a named news outlet. Accuracy: you can cross-check the laws it reports against other news reports and the states' own pages. Purpose: it exists to inform.
Now the scholarly article. It appeared in 2016, so you ask whether newer studies agree. It studied schools in four English cities, not college lectures. That's the gap the draft's critics will name, and the draft should name it first.
Pause and try this. A website ends in .org. Which checklist question does that settle?
Think before I answer.
Here's the answer. None of them. Anyone can register a .org, .com, or .net address. Only .edu and .gov restrict who may register, so an ending proves nothing about reliability.
6. Read laterally
Suppose a stranger at a bus stop offers you a great deal on a used car. You wouldn't judge the deal by asking the stranger how honest they are. You'd ask someone else about them.
That's lateral reading. Vertical reading is judging a website by staying on it and reading its own pages, such as its About page, logo, and design. Lateral reading is leaving a website and opening new tabs to see what independent sources say about it, so you learn the source's context before you trust its content. Wineburg and McGrew, in work from 2017 and 2019, found that professional fact checkers read laterally, while students and historians read vertically and were more often misled.
Here's the website I made up. It's polished. It has a .org address and a calm blue logo, and its headline reads "The Science of Screen-Free Focus." It says any phone in a classroom ruins every student's concentration. Its About page calls it an independent group devoted to student wellbeing. Read vertically, it looks trustworthy. Now read laterally. You open new tabs and search its name with words like "who runs" and "funding." Several independent sources describe it as the marketing arm of a company that sells lockable phone pouches to schools. Same site. Different picture.
Pause and try this. You read the website's About page carefully. Is that vertical or lateral reading?
Decide before I go on.
Here's the answer. Vertical. The About page is still the site talking about itself. Lateral reading means leaving the site.
7. See bias as a purpose, not a verdict
A car commercial wants you to buy the car. You can still learn the price and the trim levels from it, as long as you remember why it was made.
Bias is a source's point of view and the purpose behind it. Naming it tells you how to read the source, and it does not by itself make the source unusable. This is Week 1's rhetorical situation, turned toward your sources. Every source has a writer, an audience, and a purpose.
Back to the made-up website. Its purpose is to sell pouches. That doesn't make every sentence false, but it explains why the claim is so sweeping. So you cross-check the claim against other sources. The scholarly article found a small average gain concentrated among the lowest-achieving students. That's not ruined concentration for every student. The site went further than the evidence goes.
Pause and try this. Does the website's sales purpose prove its claim is false?
Take a breath on this one.
Here's the answer. No. The purpose tells you why to check, and the cross-check is what shows the claim overreaches.
8. Check the machine
A chatbot can sound like the most confident research assistant you've ever met. Ask it about a website, and it tells you who runs it. Ask it about a study, and it tells you what the study proved.
Those claims are unverified until you check them laterally, the same way you'd check a stranger's word about a website. Ask a chatbot about the made-up website, and it might call it "a respected independent nonprofit." That's a description, not a check. The lateral search is what tells you who runs the site. Ask a chatbot to summarize the scholarly article, and it might say phone bans "raise every student's grades." That stretches a finding about the lowest-achieving students into a claim about everyone.
So the rule is plain. A chatbot can help you brainstorm search words. Its claims about a source are leads to check, never evidence to cite.
Pause and try this. A chatbot says the scholarly article showed phone bans raise every student's grades. What's wrong with that summary?
Think about who the study found gained.
Here's the answer. It overreaches. The article found a small average gain concentrated among the lowest-achieving students, not a gain for everyone.
9. Sort and judge the running draft's four sources
Now we put it all together. The mini-source set is the running draft's four real sources: a news article, a peer-reviewed article, a book, and a government guidance page, each chosen for the job it can do in the essay.
Here's the sort, said out loud. Two are primary: the scholarly article and the guidance page. Two are secondary: the news article and the book. One is scholarly, two are popular, and one is a government document. Two came through the library, and two came from the open web.
And here's the decision. The peer-reviewed article can carry the draft's main grounds. The guidance page shows that governments act on the idea. The news article shows how widespread restrictions have become. The book supplies background on attention. The made-up website earns no place in the essay.
More sources don't make a better argument. Credible, relevant sources do, and each one needs a job.
Pause and try this. Which of the four sources should carry the draft's main grounds?
Pick one before I say it.
Here's the answer. The peer-reviewed article. It reports an original study that passed peer review, so it's the strongest evidence for the draft's claim about who gains.
Closing
Let me pull it together. A research question is focused, answerable with evidence, more than a yes or no, and arguable. Primary is the thing itself; secondary is writing about it, and your use decides which. Scholarly, popular, and trade sources serve different readers. A library database is pre-screened, and the open web isn't, so rank isn't credibility. The checklist asks about currency, relevance, authority, accuracy, and purpose, and a domain ending settles none of them. Lateral reading leaves the site to learn its context. Bias is a purpose to name, not a verdict. A chatbot's claims about a source are leads to check. And every source in the draft now has a job.
Next in this module come the slide deck, the Readings & Resources page, the lecture tutorial, the practice set, and the quiz, where you'll sort and test sources yourself. Next week, in Week 10, Integrating Sources: Quoting, Paraphrasing & Avoiding Plagiarism, you'll bring these sources into your own sentences honestly.
Built with The Course Maker · samples.thecoursemaker.com