Wait, What? A Credible Source Can Still Be Weak Evidence for the Question
A Secondary 3 student sees the phrase “university professor” and assumes the source must be reliable. Another sees a personal story and dismisses it as biased. Both shortcuts are dangerous. Source quality depends on the claim being made, what the source actually knows, how the evidence was produced and whether the evidence is relevant to the question.
A student describing their own experience can be highly credible about what happened to them and weak evidence about what happens to all students. A researcher can be highly credible about a carefully measured pattern and still be unable to explain every individual case.
Do not ask only, “Can I trust this source?” Ask, “Trust this source for which claim, under which conditions, and with what limits?”
Quick Answer
The Secondary 3 source-evaluation engine is:
CLAIM → SOURCE ROLE → ACCESS TO KNOWLEDGE → METHOD → RELEVANCE → SUFFICIENCY → CORROBORATION → LIMITATIONS → JUDGEMENT.
- Claim: what exactly needs support?
- Source role: witness, expert, institution, participant, commentator, advertiser, researcher?
- Access: how could this source know?
- Method: observation, survey, measurement, interview, document, experiment, anecdote?
- Relevance: does the evidence answer this claim?
- Sufficiency: is there enough evidence for the conclusion?
- Corroboration: do independent sources point in the same direction?
- Limitations: what remains uncertain?
- Judgement: what is the strongest defensible conclusion?
Owned Secondary 3 Learning Job
This guide owns one job: helping Secondary 3 students evaluate whether a source and its evidence are fit for a particular claim, rather than treating credibility as a simple trustworthy/untrustworthy label.
It belongs to the Secondary English Learning Hub. Its Batch 6 companions cover modality, certainty and qualification, register and style transformation, and metacognition, self-editing and independent control.
Credibility, Reliability and Relevance Are Different
| Question | What it tests |
|---|---|
| Is the source credible? | Does the source have reason, access or competence to be believed? |
| Is the evidence reliable? | Was it produced in a way likely to be consistent and trustworthy? |
| Is the evidence relevant? | Does it actually bear on the claim being made? |
| Is the evidence sufficient? | Is there enough to justify the conclusion? |
A source can score differently on each dimension.
Source Role: Who Is Speaking?
- Witness: saw or experienced an event.
- Participant: took part in the system being described.
- Expert: has specialised knowledge relevant to the claim.
- Institution: owns official records, policy or operational data.
- Researcher: systematically studies a question.
- Advocate: aims to persuade on behalf of a cause.
- Advertiser: aims to increase interest or sales.
- Commentator: interprets events, often using public information.
Roles are not automatically good or bad. They tell us what kinds of knowledge and incentives may be present.
Access to Knowledge: How Could They Know?
Ask whether the source has direct access to the information required.
A student can report how a timetable change affected their own morning routine. A school administrator may know attendance records. A transport planner may understand network constraints. These are different access routes.
Credibility becomes stronger when the access route matches the claim.
Worked Example 1: One Claim, Three Sources
Claim: “The new school start time reduced late arrivals.”
- Student: “My bus was less crowded.” Useful personal experience, not enough for school-wide lateness.
- Teacher: “My first-period class had fewer late students.” Direct observation of one class, limited scope.
- School attendance records: show total late arrivals before and after the change. More directly relevant to the school-wide claim.
The “best” source depends on the level of the claim.
Expertise Must Match the Topic
A person can be highly qualified in one domain and weakly qualified in another.
A medical researcher may be an expert on sleep and health. That does not automatically make them an expert on bus scheduling, school timetables or educational assessment.
Ask: expert in what, and is that expertise relevant to this specific claim?
Primary and Secondary Sources
A primary source provides direct evidence from the event, participant, original study or document. A secondary source analyses, reports or interprets primary material.
Neither category is automatically better. A secondary source can synthesise many primary sources. A primary source may be narrow or partial.
Method: How Was the Evidence Produced?
- direct observation;
- survey;
- interview;
- administrative record;
- experiment;
- measurement;
- case study;
- personal anecdote;
- document analysis.
Different methods answer different questions well.
Anecdote: Strong for Experience, Weak for Prevalence
“One student became more confident after joining debate club.” This can illustrate a mechanism or lived experience. It cannot establish that debate club improves confidence for all students.
The problem is not that anecdotes are useless. It is that their evidential job is limited.
Survey Evidence: Ask Who, How Many and What Was Asked
Suppose 80% of respondents support a new activity. Ask:
- Who responded?
- How many?
- Was participation voluntary?
- Were answer choices balanced?
- Was the sample representative of the group being discussed?
- Does support predict actual participation?
A percentage without context can create false confidence.
Worked Example 2: Survey Scope
Claim: “Students want longer library hours.”
Evidence: 80% of 25 students attending an after-school study programme support longer hours.
Reasonable conclusion: “A strong majority of surveyed students already using after-school study support favour longer library hours.”
Overclaim: “Students want longer library hours.” The evidence may not represent students who leave immediately after lessons or do not use the library.
Correlation Is Not Automatically Cause
If students who read more also score higher in English, several explanations are possible:
- reading contributes to stronger language;
- stronger language makes reading easier and more enjoyable;
- another factor influences both;
- all three may operate together.
Evidence of association does not automatically establish direction of cause.
Before-and-After Evidence
“Attendance improved after the programme began.” Useful, but ask what else changed at the same time. Timing can support a causal hypothesis without proving it by itself.
Relevance: Does This Evidence Answer the Claim?
Claim: “The programme improves academic performance.”
Evidence: “Students said the sessions were enjoyable.”
Enjoyment may matter, but it does not directly establish academic improvement.
Relevance requires claim-evidence fit.
The Evidence Job Test
| Claim job | Evidence that fits |
|---|---|
| How common? | representative frequency data |
| Why? | mechanism, comparison or causal evidence |
| What happened? | records, observation, direct account |
| How experienced? | interviews, testimony, observation |
| What changed? | before/after or comparative data |
| What should be done? | evidence plus criteria, cost, risk and feasibility |
Sufficiency: One Good Piece of Evidence May Still Not Be Enough
Evidence can be relevant and insufficient.
A single class improving after a new teaching method is relevant to the question of effectiveness, but broader claims require more evidence across students, time or settings.
Corroboration: Independent Agreement Matters
If attendance records, teacher observations and student surveys all point towards improved punctuality, confidence increases because different evidence streams converge.
Corroboration is strongest when sources are meaningfully independent rather than repeating the same original claim.
Source Independence
Three websites repeating the same press release are not three independent confirmations. Trace claims back to their origin where possible.
Purpose and Incentive
Why was the source produced?
- inform;
- sell;
- reassure;
- persuade;
- defend;
- criticise;
- report;
- entertain;
- advocate.
Purpose does not automatically invalidate evidence. It tells the reader what kinds of selection or emphasis deserve extra scrutiny.
Bias Is Better Understood as Perspective and Selection
Calling a source “biased” can end analysis too early. Ask instead:
- What interest does the source have?
- What is foregrounded?
- What is omitted?
- Does the selection affect this particular claim?
- Can the factual part still be checked independently?
A campaign can be strongly persuasive and still cite accurate statistics. The persuasive purpose matters, but accuracy must be tested separately.
Worked Example 3: Advertiser Evidence
Claim: “Nine out of ten users preferred Product A.”
Questions:
- How many users?
- Compared with what?
- Who selected the participants?
- What did “preferred” mean?
- Was the study independently conducted?
The advertiser’s incentive does not prove the number is false. It creates a reason to inspect method and wording carefully.
Recency and Context
Some evidence becomes outdated. A transport statistic from ten years ago may be poor evidence for current ridership if routes, population and travel patterns have changed.
Other evidence is historically stable. Recency matters when the underlying system changes.
Geographic and Population Fit
Evidence from one country, age group or school may not transfer directly to another. Ask whether the relevant conditions are similar enough.
Transfer is a judgement, not an automatic rejection.
The Missing-Denominator Problem
“Complaints doubled” sounds dramatic. From how many to how many? From two to four is different from two thousand to four thousand.
Numbers need denominators, baselines and time frames.
The Missing-Comparison Problem
“Students improved by five points.” Compared with what would have happened without the intervention? Without a comparison, improvement may reflect practice, maturation or another change.
Evidence Quality and Modality
Source evaluation should change language strength.
Weak source + limited evidence → “may suggest”.
Strong converging evidence → “indicates”, “shows” or a more confident conclusion within defined scope.
This connects directly to the Modality, Certainty and Qualification Guide.
Evidence Evaluation in Comprehension
When a question asks why a writer includes evidence, identify both the evidence type and its argumentative job.
Statistic → shows scale.
Anecdote → humanises consequence.
Expert quotation → adds specialised authority.
Counterexample → limits a broad claim.
Evidence Evaluation in Writing
Students should choose evidence according to the claim they need to establish. A paragraph about prevalence needs more than a personal story. A paragraph about lived experience may benefit from one.
Evidence Evaluation in Oral Communication
In oral discussion, students can signal source limits naturally:
I’ve seen this work in my own class, although that’s only one example. A broader judgement would need evidence from more students.
This sounds thoughtful because it distinguishes experience from generalisation.
Worked Example 4: Evaluating a Source Pair
Source A: school website reports that a new programme “greatly improved student engagement”.
Source B: participation records show attendance increased from 40 to 65 students, but no learning or engagement measure is provided.
Judgement: the records support increased participation, not necessarily improved engagement. The school website may be accurate, but the available evidence shown does not yet establish the stronger claim.
The Source-Evidence Matrix
| Dimension | Question |
|---|---|
| Access | How could this source know? |
| Expertise | Is the source qualified for this claim? |
| Method | How was the evidence produced? |
| Scope | Who or what does it actually cover? |
| Purpose | Why was the source produced? |
| Relevance | Does it answer the claim? |
| Sufficiency | Is there enough? |
| Corroboration | Do independent sources agree? |
| Limit | What remains unknown? |
The Claim-Source Fit Test
Write the claim above the source. Then ask: “What part of this claim can this source directly support?” Cross out everything beyond that boundary. The exercise teaches students to shrink claims until the evidence can carry them.
The Corroboration Ladder
- one source;
- two sources repeating same origin;
- two independent sources;
- different methods pointing same way;
- multiple independent methods across time or context.
Confidence generally increases as genuinely independent evidence converges, while limitations still matter.
Earliest Weak-Link Diagnosis
| Visible problem | Likely earliest weak link | Repair |
|---|---|---|
| “Expert = true” | expertise matching | ask expert in what |
| Dismisses anecdote entirely | evidence-job model | separate experience from prevalence |
| Accepts percentages immediately | method/scope | ask sample, denominator and wording |
| Confuses relevance and reliability | dimension separation | evaluate each dimension independently |
| Calls everything biased | purpose analysis | identify specific selection effect |
| Uses one source for broad claim | sufficiency | narrow claim or seek corroboration |
| Multiple sources are copies | independence | trace original source |
A Weekly Source-Evaluation Cycle
- Choose one claim.
- Collect three source types supporting or challenging it.
- Identify each source role and access to knowledge.
- Classify the evidence method.
- Judge relevance and sufficiency separately.
- Trace whether sources are independent.
- Write the strongest conclusion the combined evidence supports.
- State one important limitation.
- Return with a different topic.
Student Source-Evaluation Receipt
- What exact claim needs support?
- Who is the source?
- How could they know?
- Does their expertise match the claim?
- How was the evidence produced?
- What population or context does it cover?
- Is it relevant?
- Is it sufficient?
- What purpose or incentive may shape selection?
- Do independent sources corroborate it?
- What remains unknown?
- How confident should my final wording be?
Parent and Tutor Teaching Guide
When a student says “This source is reliable because it is official”, ask what specific claim the official source owns. An institution may be authoritative about its own rules and weak evidence about whether those rules are effective.
Use mixed source packs. Include a statistic, a testimonial, an expert comment and an institutional statement. Ask the student to match each source to the claim it can best support. This prevents source status from replacing reasoning.
Require one limitation in every evidence judgement until the student learns that limits do not destroy evidence; they define its proper range.
Unfamiliar Transfer Challenge
Give the student four unfamiliar sources about a proposed school policy: an official notice, student survey, parent comment and external expert article. Ask the learner to:
- state what each source can know directly;
- identify method and purpose;
- match each to one claim;
- rank evidence strength for different questions;
- identify any dependence between sources;
- write a qualified overall judgement;
- state what extra evidence would most improve confidence.
Common Secondary 3 Source-Evaluation Traps
- Treating credentials as universal expertise.
- Rejecting personal accounts because they are subjective.
- Assuming official sources are neutral on every question.
- Using percentages without denominator or sample.
- Confusing correlation with cause.
- Accepting repeated reports as independent corroboration.
- Using relevant evidence that is still insufficient.
- Calling a source biased without explaining how.
- Ignoring recency or context.
- Allowing one source to support a claim broader than its scope.
Useful eduKate Sengkang Routes
- How Students Judge Whether Evidence Is Relevant and Sufficient
- How Students Separate Fact, Opinion and Interpretation
- Observed, Inferred, Unknown | Marie Curie Series
- Comparative Reading and Multiple-Text Judgement
Evidence Boundary
The source-evaluation matrix in this guide is an eduKate learning scaffold, not an official examination marking formula. Real evidence evaluation depends on the actual source, claim, method and context. Treat credibility judgements as evidence-based conclusions rather than permanent labels.
Continue Secondary 3 English Learning Guide Batch 6
- Modality, Certainty and Qualification
- Register and Style Transformation
- Metacognition, Self-Editing and Independent Control
- Return to the Secondary English Learning Hub
The Quiet Return
Source evaluation becomes calmer when the student stops asking whether a source is simply “good” or “bad”. Evidence has jobs. Sources have access. Claims have scope.
Start with the claim. Ask how the source could know. Inspect the method. Match the evidence to the job. Then state only the conclusion the source can carry.
