Evidence and Research
Evidence is what turns an opinion into an argument. But evidence is not magic: a fact only helps if it is relevant, accurate and explained. This page shows students how to find, test and present support without overclaiming.
Evidence is a relationship
A piece of evidence does not support a claim on its own. It supports a claim when the audience can see the connection. That means the speaker must explain what the evidence shows and why it matters.
| Claim | Evidence | Analysis |
|---|---|---|
| School uniforms do not reduce bullying. | A study found no significant difference between schools with and without uniforms. | The uniform may signal belonging, but it does not change the social dynamics that produce bullying. |
| Social media can amplify misinformation. | Platform algorithms prioritise engagement. | Outrage and false claims generate engagement, so they are promoted regardless of accuracy. |
| Public transport reduces road congestion. | Cities with higher transit use report lower peak-hour traffic volumes. | When people have a viable alternative, fewer cars enter the city at peak times. |
These are illustrative examples, not citations. A real debate should use checked sources with dates and names.
Types of support
| Type | Strength | Limit |
|---|---|---|
| Statistic | Shows scale. | Can be misleading without context. |
| Research finding | Shows systematic evidence. | Quality and applicability vary. |
| Expert opinion | Borrows authority. | Experts disagree; credentials matter. |
| Historical example | Shows precedent. | Context may not transfer. |
| Case study | Shows detail and human impact. | One case is not a trend. |
| Analogy | Makes an abstract idea concrete. | Analogies can mislead if the comparison is weak. |
| Common knowledge | Needs no citation. | What is common to you may not be common to the audience. |
Testing a source
Use a short checklist. The goal is not to find perfect sources but to know how much weight a source can bear.
| Test | Question | Red flag |
|---|---|---|
| Author | Who produced this and what are their credentials? | No author or anonymous post. |
| Publisher | Is it a recognised organisation or a personal blog? | A site with no editorial policy. |
| Date | Is it current enough for the claim? | Old data presented as current. |
| Method | How was the information gathered? | No method described. |
| Sample | Who was studied and how many? | A tiny or unrepresentative sample. |
| Interest | Does the source benefit from the conclusion? | Undisclosed funding or advocacy. |
| Corroboration | Do other sources agree? | A single outlier source. |
| Precision | Does the number mean what the speaker says? | Correlation presented as cause. |
Statistics
Numbers are powerful and easy to misuse. Treat a statistic as a claim that needs context, not as a trump card.
| Problem | Example | Question to ask |
|---|---|---|
| Missing base rate | “Twice as many students ...” | Twice as many as what? |
| Correlation as cause | “Regions with more ...” | Could a third factor explain both? |
| Cherry-picked period | “Since 2020 ...” | What happens if we look at a longer period? |
| Misleading average | “The average student ...” | Is the distribution skewed? |
| Percentage of a small number | “Fifty per cent of ...” | Fifty per cent of how many? |
| Precision without meaning | “37.4 per cent ...” | Is that precision justified by the method? |
Attribution
Name the source in the speech. The audience does not need a full reference, but they need enough to judge the claim.
| Weak attribution | Strong attribution |
|---|---|
| “Studies show ...” | “A 2024 study by the Grattan Institute found ...” |
| “Experts agree ...” | “The Australian Institute of Health and Welfare reports ...” |
| “Everyone knows ...” | “Most Australian adults ...” (with a source) |
| “It has been proven ...” | “One randomised trial found ...” |
| “Research says ...” | “Research published in ... in 2023 suggests ...” |
If you cannot name the source, do not present the claim as established fact. Say “for example” or “imagine” and label it as illustration.
Source record
Keep a simple record for each source so the team can check it quickly during preparation.
| Field | What to record |
|---|---|
| Author or organisation | Who produced it. |
| Title | Exact title. |
| Date | Publication or update date. |
| Link or location | URL or publication details. |
| Claim supported | The exact point it supports. |
| Key detail | The number, finding or quotation. |
| Limits | Sample, method, bias or date. |
| Confidence | High, medium or low. |
Research routine
- Write the claim you need to support.
- List the kinds of source that could support it.
- Search broadly before searching narrowly.
- Open the original source, not only a summary.
- Record the claim, the detail and the limit.
- Find a second source for any key claim.
- Decide how much weight the evidence can bear.
Evidence is not a substitute for argument. A pile of facts with no explanation is just a pile.
Activity: source audit
- Bring three sources on one motion.
- Complete a source record for each.
- Mark each source high, medium or low confidence.
- Identify one limitation in each.
- Write a sentence explaining how you would use it in a speech.
- Remove the weakest source and replace it.
Common errors
- “Studies show” with no study. The audience cannot check it.
- Using a source that does not say what you claim. Misrepresentation destroys trust.
- Ignoring the date. Old evidence may be irrelevant.
- Confusing correlation with cause. The claim overreaches.
- Overloading the speech with numbers. The audience stops listening.
- Inventing a statistic. It is dishonest and easily exposed.
Next, study rebuttal to answer the other side's evidence. The points of information page teaches quick, relevant challenge.