Lecture: Research & Supporting Materials
Lecture script · about 15–20 minutes spoken · Week 4
Opening
Picture three people at work. A nurse at a shift handoff points to the unit's posted protocol. A resident at a neighborhood meeting reads out a traffic count and says who took it. A job candidate backs up a claim with the name of the project behind it. All three win trust the same way. They say where their evidence came from, and it holds up when someone checks.
Here's where we are. In Week 1 you learned that an ethical speaker credits sources out loud and never fabricates. In Week 2 you learned to read an audience. In Week 3 the running speech got its plan: an informative speech about food waste in a campus dining hall, with a central idea that makes three promises. This week you find the proof for those promises, and you learn to say that proof out loud.
In this lecture I'll take you through five ideas. The three kinds of supporting material. How to judge a source. How to read a statistic in plain words. The oral citation. And what to do when a chatbot hands you citations. Next week, in Organizing the Speech, you'll arrange this evidence into main points and a pattern. Let's start with a teammate who said "studies show."
1. Three kinds of supporting material
You're in a group-project presentation. Your teammate says, "Studies show most students waste food." A hand goes up. "Which study?" Nobody knows. The claim might even be true, but it lost the room, because nobody could tell where it came from.
So what does good proof look like? You already know more than you think. When you talk a friend into trying a restaurant, you might tell one story about a great meal there. You might mention the long line every night. You might say a cousin who cooks for a living loves it. That's a picture, a number, and a witness. Three different kinds of proof.
The course names them. Supporting material is the evidence a speaker uses to back up a point; this course sorts it into three kinds: examples, statistics, and testimony.
Examples come in three forms. A brief example is a real case told in a sentence or two to make a point concrete. An extended example is a real case told as a short story with a beginning, a middle, and an end. And a hypothetical example is an invented case that asks listeners to imagine a situation; an honest speaker always flags it as invented, with a cue such as "picture this." That flag matters. If you tell an invented story as if it happened, you've passed fiction off as fact.
A statistic is a number that summarizes facts about a group or an event, used as evidence only when it comes from a credible source, is used accurately, is dated, and is made meaningful to listeners.
And testimony is another person's words, opinions, or findings used as evidence, quoted exactly or paraphrased, and always credited to that person aloud. Testimony has two kinds that people mix up constantly. Expert testimony comes from a person qualified in the field the claim belongs to, such as a researcher reporting a study's findings. Lay testimony comes from an ordinary witness speaking from firsthand experience. Lay testimony tells you what something was like. Expert testimony tells you what a qualified person concluded.
Here's how the running speech uses all three. It opens with a hypothetical example: "Picture the dish-return conveyor at closing time." It uses a statistic from the U.S. Department of Agriculture: between 30 and 40 percent of the food supply in the United States goes to waste. And it uses expert testimony from researchers Wenhao Zhang and Junehee Kwon, who studied trays in a university dining hall.
Pause and try this. At a neighborhood meeting about a crossing near campus, a crossing guard describes waving cars down at that corner, and another speaker cites a traffic engineer's review. Which is expert testimony, and which is lay?
Hold both people in mind for a moment.
Here's the answer. The crossing guard gives lay testimony, a firsthand account of what the corner is like, and the engineer gives expert testimony, a qualified judgment about what would make it safer. A strong speech uses each for the job it can do.
2. Judging a source
Would you take directions from a stranger who won't say where the road goes? Sources deserve the same suspicion.
Credible sources tend to live in a few places. Library databases. Government and institutional pages. Quality news outlets with editorial standards, which means they check facts and correct their mistakes in public. And interviews with experts. Two things to steer clear of: anonymous posts, where you can't tell who's talking, and a chatbot's output cited as if it were a source.
The course gives you a tool. The credibility checklist is five questions asked of a source before a speaker uses it: currency, relevance, authority, accuracy, and purpose. Is it recent enough? Does it fit the point? Who made it, and are they qualified? Is it sourced and checkable? And why does it exist? Academic librarians built this checklist, and Blakeslee described it in 2004. It's one common checklist among several, and the habit matters more than the name.
Let me run it on a made-up website, invented for teaching. It has a medical-blue logo and a clean layout. There's no author and no date. There's a bold claim that most people are low in key nutrients, with no source. And halfway down, there's a "buy now" button for a supplement. Currency fails, because there's no date. Relevance passes; it's about nutrition. Authority fails; nobody you can look up. Accuracy fails; the claim has no source. Purpose fails; the page exists to sell. That's one pass and four fails. Drop it. Looking official is not a credential.
Now the real sources. The running speech uses three, and the course calls them the mini-source set. There's a U.S. Department of Agriculture page called "Food Waste FAQs." There's a peer-reviewed article by Zhang and Kwon in the journal Sustainability, from 2022, with a DOI, which is a permanent code that lets anyone find that exact article. And there's a PBS News article by Associated Press reporter Dee-Ann Durbin, also from 2022. Three sources, five questions each. That's 15 checks, and all three sources are kept.
One check deserves a closer look. The government page has no dated stamp, and it cites a 2010 estimate. So its currency is the weak spot. The fix isn't to drop it. The fix is to cite it honestly: the speaker names the department and its research service, and doesn't invent a year.
The checklist looks at one source by itself. The second habit looks outward. Before you trust a page, open new tabs and see what other credible sources say about it. An independent source is a source produced by someone other than the original source's author or organization, used to check whether a claim holds up. If a claim shows up only on the page that's selling something, you have your answer.
Pause and try this. The made-up wellness site passed only one of the five questions. Which one, and why isn't that enough?
Run the five questions in your head first.
Here's the answer. It passed relevance, because it's about nutrition, but a page can be on topic and still have no author, no date, no source, and a sales motive, so relevance alone never earns trust.
3. Reading a statistic in plain words
Your roommate says the average shift at the campus café is short. Most shifts are short. One ran past midnight. The average hides that story. Numbers in a speech work the same way. One number stands in for many, and your listeners deserve to know what it's standing in for.
You don't need a formula for this. You need three words and three questions. An average is a single number that stands for the typical value in a set of numbers. Spread is how far the values in a set scatter from one another. And a range is the distance from the lowest value to the highest, often reported as the two end values, like between 30 and 40 percent.
Here's a picture from the chapter. Two food pantries each hand out bags of groceries on four distribution days. Pantry A is steady, with every day close to 40 bags. Pantry B swings wildly, with one day at 10 bags and another at 70. Both average 40 bags a day. If you report only the average, you hide the swing, and the swing is exactly what a volunteer coordinator needs to know to plan who works when.
Now the three questions. Ask them of any number before you say it. Who counted? When? And how many, and of what?
Let's ask them of the running speech's two figures. The Department of Agriculture's page says between 30 and 40 percent of the U.S. food supply goes to waste. Who counted? The department's Economic Research Service. When? The page cites a 2010 estimate. And notice that it's a range, not one number. The range itself tells you the estimate is uncertain.
The PBS News article says about 35 percent of the food available in the United States in 2019 went unsold or uneaten, according to ReFED, a group that studies food waste. Who counted? ReFED. When? 2019. Of what? Food available in the country, counted as unsold or uneaten.
Put them side by side. Different counters, different years, slightly different definitions. The 35 sits inside the range of 30 to 40, so the two broadly agree. But they weren't counted the same way, so they can't tell you whether the problem got better or worse.
Pause and try this. Using the two figures, can the running speech honestly say that food waste went down between the 2010 estimate and 2019?
Ask who counted, when, and of what.
Here's the answer. No, because the two figures come from different counters using different definitions, so the speech can say the counts broadly agree on roughly a third, but not that waste fell.
4. The oral citation
When you write a paper, your reader can flip to the end and check your sources. Your listeners can't. They hear your speech once, in order. The only source they'll ever know is the one you say out loud.
That's the job of the oral citation. It's a spoken credit, placed in the sentence before the evidence. It names who produced the source, with one clause on why they're credible. It names what the source is, its title or venue. And it names when it appeared. Who, what, when. Then the evidence.
Why before? Because then the number arrives with its credentials already attached. Listeners hear "Associated Press reporter Dee-Ann Durbin, writing for PBS News," and they're ready to trust what comes next.
The oral citation has a partner on paper. A written bibliography is the list of sources at the end of a preparation outline, in one consistent style your instructor sets; APA and MLA are the common choices. Your instructor reads it. Your listeners never hear it. So one never replaces the other.
Two more rules. First, a paraphrase still needs its source. If you put a study's finding in your own words, the words are yours but the idea isn't, so you still name the study. Second, keep Week 1's line clear. Using real words or ideas without credit is plagiarism. Inventing a quotation, a statistic, or a study is fabrication, and fabrication is the more serious breach.
Here's a real case that shows why the source of a quotation matters. Sojourner Truth spoke at a women's rights convention in Akron, Ohio, in 1851. Two versions of that speech survive. Marius Robinson printed his report in 1851, the same year. Frances Gage published a different version in 1863. They differ so much that the famous refrain the speech is now named for appears in Gage's later version and not in Robinson's earlier report. So if you quote "what Sojourner Truth said," you owe your audience an answer: which version? A quotation is only as good as the transcript it comes from.
Pause and try this. A speaker says, "Studies show that skipping breakfast hurts focus." Which of the three parts of an oral citation are missing?
Count who, what, and when.
Here's the answer. All three are missing, because "studies show" names no author or organization, no title or venue, and no date, so the listener has nothing to check.
5. Building the citations and checking the chatbot
Now let's put the week together. The running speech has three real sources, and each becomes one oral citation.
The first sounds like this. "According to the U.S. Department of Agriculture's Food Waste FAQ page, which draws on the department's Economic Research Service, between 30 and 40 percent of the food supply in the United States goes to waste." Who is the department and its research service. What is the FAQ page. And when? The page has no date, so the speaker names the department and the research service instead of making up a year.
The second is expert testimony, paraphrased, and I'll say it in two breaths. In a study published in the journal Sustainability in 2022, researchers Wenhao Zhang and Junehee Kwon compared a week without trays in a university dining hall with a week with trays. Without trays, diners took less food, but the study did not show a significant reduction in food waste. Who, what, and when, all before the finding.
The third names a reporter and the group that counted. As Associated Press reporter Dee-Ann Durbin wrote for PBS News in 2022, about 35 percent of the food available in the United States in 2019 went unsold or uneaten, according to ReFED, a group that studies food waste. The full citation in the chapter gives the exact date.
Three citations, three parts each. That's nine parts named, every one before its evidence. And notice which part is the easiest to drop. It's the when.
Now the chatbot. You'll use one as a rehearsal coach in this course, and it's handy for brainstorming search terms. But here's what happens when you ask it for evidence. In a made-up exchange, a student asks for three statistics about dining-hall food waste, with sources. Back come three tidy citations, each with authors, a title, a journal, and a year. They look complete. They look exactly like the second citation you heard a minute ago.
The student checks each one at the original: the journal's own site, a library database, a search for the exact title. One is found, and it says what the chatbot claimed. The second names a real journal, but no article with that title is in it. The third names authors and a title that turn up nowhere. One kept, two dropped.
That's the sharpest lesson of the week. A chatbot asked for sources can return citations with authors, titles, and years, and some of them don't exist. The original is the source; the chatbot isn't. And even a real article can be attached to a claim it never makes. So the check isn't only "does it exist?" It's also "does it say this?"
Pause and try this. A chatbot hands you a citation with an author, a title, a journal, and a year. What do you do before you say any of it out loud?
Think about where the real source lives.
Here's the answer. Find the article in the journal or a library database, read what it says, and cite the article itself. If you can't find it, drop it, because a chatbot is not a source.
Closing
Let me pull it together in the order we built it. Supporting material comes in three kinds, examples, statistics, and testimony, and expert testimony judges while lay testimony describes what it was like. Every source runs through five questions, currency, relevance, authority, accuracy, and purpose, and then gets checked against independent sources. A statistic gets read in words, average, spread, and range, and gets asked who counted, when, and of what. Every piece of evidence arrives with an oral citation that names who, what, and when before it. And every citation a chatbot hands you gets found at the original or dropped.
Next in this module come the slide deck and the Readings & Resources page, then the lecture tutorial, the practice set, and the quiz. Next week, in Organizing the Speech, the evidence you've gathered gets arranged into main points and a pattern that fits your purpose.
Built with The Course Maker · samples.thecoursemaker.com