Research and Source Quality
An explanation is only as good as the material behind it. Research for a speech is not collecting everything you can find; it is finding the few sources that can carry the weight of your claim. This page offers a practical method for judging sources and keeping track of them.
The four core criteria
| Criterion | Questions | Warning signs |
|---|---|---|
| Authority | Who wrote it? What qualifications or experience do they have? Is the publisher reputable? | No author, no organisation, anonymous claims. |
| Accuracy | Are claims supported by evidence? Can they be checked? Does the source cite others? | Unsourced statistics, emotive language, obvious errors. |
| Currency | When was it published or updated? Has the field changed since? | Old data presented as current; no date. |
| Relevance | Does it answer your question for your audience? | Interesting but tangential; different context. |
No source is perfect on every criterion. A government report may be authoritative and current but broad; a firsthand account may be relevant and vivid but not representative. The skill is matching the source to the job.
Primary and secondary sources
| Type | What it is | Strength | Limitation |
|---|---|---|---|
| Primary | Original data, documents, interviews, observations. | Close to the event or evidence. | Requires interpretation; may be partial. |
| Secondary | Analysis or summary of primary material. | Provides context and synthesis. | Can distort or oversimplify. |
| Tertiary | Reference works, encyclopedias, summaries. | Useful for orientation. | Not usually sufficient for a specific claim. |
For a short speech, a reputable secondary source is often enough. If you are making a specific factual claim, go to the primary source where possible.
Reading laterally
Professional fact-checkers do not stay on one page and inspect it for clues. They open new tabs and read about the source itself. This is called lateral reading. Ask: What do other reputable sources say about this organisation or author? Is the claim repeated by independent sources, or only by sites that copy each other?
- Open the source.
- Identify the author and publisher.
- Open a new tab and search for the author or organisation.
- Look for independent descriptions, criticism or corrections.
- Return to the source and decide whether it is trustworthy enough for your purpose.
Bias and perspective
Every source has a perspective. Bias is not automatically disqualifying; a lobby group may be the best source for its own position, and a personal account may be the best source for lived experience. The problem is using a partisan source as if it were neutral, or accepting a claim without checking it elsewhere.
| Source type | Useful for | Check against |
|---|---|---|
| Government statistics | Population, health, education, economy. | Definitions, collection methods, date. |
| Peer-reviewed research | Evidence about causes and effects. | Sample size, replication, limitations. |
| News reporting | Recent events and public debate. | Primary documents and other outlets. |
| Advocacy groups | Arguments, campaigns and affected voices. | Independent evidence and opposing views. |
| Personal testimony | Experience, emotion and detail. | Generalisability and consent. |
| Social media | Public reaction and leads. | Original sources and verification. |
Statistics and sample size
A statistic is not self-explanatory. Ask: What was measured? Who was included? How many? Compared with what? A survey of twelve students in one class cannot support a claim about all Australian teenagers. A percentage without a base is meaningless.
| Claim | Question to ask |
|---|---|
| “Seventy per cent of students are stressed.” | Seventy per cent of which students, in what survey, with what question? |
| “Crime has doubled.” | Reported crime or actual crime? Over what period? Per capita? |
| “This program works.” | Compared with what? Measured how? For whom? |
| “Scientists say ...” | Which scientists? Is there a consensus? What is the source? |
A source log
Keep a simple log as you research. It saves time later and makes attribution easy.
| Field | What to record |
|---|---|
| Author or organisation | Who produced it? |
| Title | Exact title of the page, report or article. |
| Date | Published or last updated. |
| URL or location | Where you found it. |
| Claim used | The specific fact or idea you are borrowing. |
| Verification | Which independent source confirms it? |
| Notes | Limitations, bias or uncertainty. |
Activity: rank three sources
- Choose a question for an explanation.
- Find three sources: one official, one news or advocacy, one personal or social.
- Score each out of five on authority, accuracy, currency and relevance.
- Write one sentence explaining which source you trust most and why.
- Find one claim that appears in more than one independent source.
- Find one claim you cannot verify and decide whether to drop it.
If you cannot check it, do not claim it. A speech does not need many sources, but every factual claim needs a source you could name if asked.
Common errors
- Using the first search result. Ranking is not evidence of quality.
- Trusting a familiar site. Familiarity is not authority.
- Ignoring dates. Old data can mislead.
- Treating a statistic as an argument. Numbers need interpretation.
- Hiding uncertainty. Audiences respect honest limits.
- Copying a claim without reading the source. The summary may distort the original.
Next, study the structure of an explanation to organise your research for a listener. The examples and analogy page shows how to make evidence concrete.