Did you know that five headlines can sometimes amount to one piece of evidence? A student researching a school event might see the same impressive statistic on a news site, a community blog, a social post and two school newsletters. The number of pages feels persuasive. Yet every page could be repeating one original announcement.
Source independence means that a report’s evidence was obtained or verified separately, rather than copied from another report. To decide whether several articles independently support a claim, trace each one backwards to the first record, interview, observation or dataset it relied on. Count independent evidence trails, not URLs. If four stories quote one announcement, they supply four routes to the same claim, not four independent confirmations.
The useful question is not ‘How many websites agree?’
The more useful question is: What did each website actually check? One reporter may have viewed the signed attendance sheet. Another may have interviewed a participant and independently checked photographs. A third may simply have reprinted a press release. These are different relationships to evidence even if all three ultimately use the same sentence.
A source can be useful without being independent. A well-written summary may give background, explain terminology or point readers towards the primary record. But a summary does not become an independent measurement merely because it appears on a separate domain. It is especially easy to miss this when a later article omits the phrase ‘according to’ and makes the original statement sound like its own finding.
The Digital Inquiry Group’s Civic Online Reasoning teaching materials teach students to investigate who is behind information, what evidence supports it and what other sources say. Our additional practical step here is to draw a small evidence family tree. It makes repeated reporting visible.
A five-minute provenance map
- Write the claim exactly. Note the number, period, group and qualifying words, not merely the topic.
- Identify the immediate source. Look for ‘according to’, hyperlinks, credited reports, named interviews and dataset references.
- Follow the citation backwards. Ask whether Site B quotes Site A, whether both use Press Release P, or whether B performed new verification.
- Record what was actually checked. Distinguish direct observation, a document seen, an interview with a knowledgeable witness and a restatement.
- Mark uncertainty honestly. An unattributed number is not proof that the writer performed a separate check.
A neat way to draw this is P → A, P → B and B → C. Three final articles appear on the screen, but they lead to one announcement P. If D independently verifies the original attendance register, add Register → D as a new trail. Even then, independence must be judged against the actual records: the ‘register’ might be the very document on which P was based.
Worked example: three reports and one school reading challenge
Completely invented classroom example; no actual school or news outlet is described. A fictional school has 30 registered pupils in a voluntary reading challenge. A school announcement states that 18 pupils completed the challenge. Website A republishes the statement as ’18 of 30 completed’. Website B says ‘60% succeeded’ and cites the announcement. Website C cites B and writes ‘a majority finished’.
Calculate the percentage by hand: 18 ÷ 30 = 0.6, so 18/30 × 100% = 60%. The arithmetic is correct if the denominator really is 30 registered pupils and the numerator really is 18 pupils meeting the same completion rule. The three articles do not establish either condition independently.
The evidence map is Announcement → A, Announcement → B, and B → C. It has one reported numerical origin, not three independent counts. A reader should write: ‘Three reports repeat a school announcement stating that 18 of 30 registered pupils completed the challenge (60%); independent verification has not yet been established.’
Now imagine that a teacher lets the project team inspect an authorised, anonymised completion log. It lists 18 distinct completion records and 30 eligible registrations, with the same deadline and definition of completion. The team can now say the available log is consistent with the announcement. It should still avoid claiming that an independent person audited the log unless someone actually did. The log and announcement may share the same underlying source.
Try the reverse check: 60% of 30 = 0.60 × 30 = 18, exactly the stated count. But agreement between two calculations based on the same inputs verifies arithmetic, not the truth of the inputs. This distinction transfers to charts, science investigations and historical accounts.
Diagnosis: what does the learner think evidence is?
Ask: ‘Four blogs repeat a statistic from the same email. How many independently measured statistics have we identified?’ A learner who answers ‘four’ is counting appearances. The stronger response is ‘one announced statistic, with no independent measurement yet demonstrated’.
Next ask: ‘If a reporter interviews the person who wrote that email, is that automatically independent confirmation?’ No. It may clarify the origin or method, but the writer might only repeat the same calculation. Students should ask which new record or observation the interview provides.
Finally ask: ‘What if two websites do not link to one another?’ Absence of a hyperlink is not proof of independence. They might both quote the same press briefing, database or syndication service. Search for distinctive wording, original dates, figures and attribution.
How this can mislead a school project
Suppose a presentation slide says: ‘Three independent sources confirm the participation rate.’ An audience may reasonably infer that three separate teams counted pupils. If the citations all lead back to one announcement, that description is too strong. It turns repetition into invented certainty.
A safer slide distinguishes coverage (‘the claim was repeated by three outlets’) from corroboration (‘the claim was independently checked against additional evidence’). Neither needs exaggeration. The point of education is to help a reader see exactly what the evidence warrants, not to collect the maximum number of footnotes.
Progressive practice: from tracing to independent judgement
Guided: recognise an echo
Article A quotes a city report saying 12 new trees were planted. Article B cites A and repeats 12. Article C reproduces the city report. Draw arrows: Report → A, A → B, Report → C. Write a one-sentence conclusion: three articles repeat one reported count. Do not add 12 + 12 + 12 or claim three independent tree censuses.
Supported: distinguish a second kind of evidence
Imagine a school event announcement reports 42 visitors. A student volunteer independently counts 40 entries at one entrance but knows a side entrance was open. The volunteer’s record is separate, yet it does not cover exactly the same population. The right answer is not ‘the announcement is wrong’; it is ‘the counts measure different possible sets, so we need the event boundary, all entrances and a documented method’.
Independent: check the same claim in a new setting
Give learners four fictional accounts of a sports fundraiser: one firsthand cash ledger, one organiser’s statement summarising that ledger, a newspaper quoting the organiser, and a volunteer’s separate count of donation boxes. Ask them to draw the source graph, specify what each record measures, calculate any supported percentage and compose a claim with one appropriate uncertainty sentence. The task is complete only when they can explain why some evidence trails are linked.
What parents and teachers can do tomorrow
At home, ask a child to choose one harmless factual claim from a school report or an age-appropriate news article. Rather than demanding more links, ask: ‘Which page first supplied this number?’ Encourage opening the original record when publicly available and appropriate. If the source is not accessible, let ‘not independently checked’ be a legitimate final finding.
In class, provide a three-column evidence table: statement, immediate citation, original basis. Keep the exercise short. A five-minute source map before drafting can prevent a paragraph built on mutually dependent citations. Evaluate the explanation of source relationships, not the number of tabs the learner opened.
Observable progress looks like this: a pupil can draw an accurate arrow graph without teacher hints, avoid counting repeated publication as separate measurements, calculate the stated number correctly and preserve caveats when the original source is inaccessible. Repeat with an unfamiliar claim a week later to check transfer.
Limits: dependence is sometimes subtle
Two independent newspapers may both rely on a single government data release, yet each may perform independent calculations or investigations. Their arithmetic can be independently checked even though their base data are not. Conversely, two differently worded websites may use the same undisclosed report. Do not infer the whole evidence relationship from wording alone.
There are also reasons not to open every underlying record: privacy, copyright, restricted school information and safeguarding. Students should never publish attendance names or private family details to prove that a calculation is sound. Use public originals or teacher-provided anonymised records. State the limit instead of attempting access without permission.
Frequently asked questions
Does a fact-checking website automatically count as independent?
Not automatically. Read its method. A fact-checker may independently inspect records, interview knowledgeable people and test calculations, or may simply cite another assessment. Credit what it actually verifies.
Can two secondary sources be helpful?
Yes. They can add context, interpretation and links to primary evidence. ‘Helpful’ and ‘independent numerical confirmation’ are different classifications, not compliments or insults.
Should students always cite only original sources?
No. Cite a trustworthy interpretation when it contributed interpretation; cite the original for what it directly shows. Attribute each kind of contribution accurately. Our guide to quoting a source that quotes another source explains a closely related problem.
What if the original report has been corrected?
Use the corrected version and show which number or conclusion changed. Our guide to original publication and update dates helps separate the event date from the date a claim was revised.
The transferable habit
A thoughtful learner does not need to distrust every publication. They need to trace exactly what each publication knows. The habit is simple: find the claim, follow the evidence backwards, preserve independent records and write no more than those records support. That is critical reading in action.
For wider learning pathways, visit the eduKate Education Hub and the Core Aim of Education learning hub. The Clementi mathematics tutorial reference explains the same diagnosis-first philosophy in a subject-specific setting.