What you'll learn
Evaluating a source means deciding how far it can be trusted for the particular purpose you have in mind, and the answer is never a simple yes or no. A source can be reliable on one matter and unreliable on another, and a biased source can still be the best evidence available if you know what the bias is and account for it. This topic covers the criteria used to evaluate sources, the difference between bias, prejudice and a point of view, the markers that reveal slant in wording and selection, how to judge statistics and how to handle online material where the usual signals of authority are absent. Every source used in your portfolio has to survive this test, and Paper 1 asks you to apply it to unseen material.
Key terms and definitions
Reliability — the degree to which a source can be depended on for accurate information.
Validity — whether a source actually addresses the question you are asking.
Bias — a consistent leaning that shapes what a source includes, omits and emphasises.
Prejudice — a judgement formed in advance of the evidence and held against it.
Point of view — an openly declared position, which is not in itself bias.
Vested interest — a stake in the outcome that gives a source reason to present matters a particular way.
Corroboration — confirmation of a claim by an independent second source.
Provenance — the origin of a source: who produced it, when, for whom and why.
Selection bias — slant produced by what is left out rather than by what is said.
Loaded language — wording that carries evaluation within apparently factual description.
Anecdotal evidence — a claim supported by individual cases rather than systematic data.
Peer review — evaluation of academic work by other specialists before publication.
Core concepts
The evaluation criteria
Five questions cover most of what evaluation requires.
Authority. Who produced this, and what qualifies them on this subject? Credentials are field-specific: an eminent economist has no special standing on epidemiology. For institutions, ask what the organisation exists to do.
Accuracy. Are claims supported, are sources given, and does the material contain errors you can detect? Careless errors in checkable details suggest carelessness elsewhere.
Objectivity. What is the purpose — to inform, to persuade, to sell? Persuasive purpose is not disqualifying, but it tells you the material is selected.
Currency. How recent is it, and does recency matter here? For population figures or membership of a regional body it matters greatly; for an eyewitness account of an event it may not.
Coverage. Does it treat the question in enough depth, and does it address your question rather than an adjacent one? This is validity rather than reliability, and a perfectly reliable source that answers a different question is of no use.
Bias, prejudice and point of view
These three are routinely confused, and the distinction is examined.
A point of view is a declared position. A trade union publication arguing for higher wages has a point of view, states it, and can still present accurate figures. Declaring a position is honest.
Bias is a consistent leaning that shapes the material, often without being declared. It shows in what is selected and omitted, in which voices are quoted, in the adjectives attached to each side, and in what is treated as needing proof.
Prejudice is judgement formed before the evidence and maintained against it. Where bias shapes presentation, prejudice refuses evidence altogether.
The practical consequence: a biased source is still usable if the bias is identified and the claims are corroborated. What is not usable is a source whose bias you have not noticed.
Where bias shows itself
Word choice. Protesters and rioters, regime and government, scheme and initiative describe the same things with the evaluation built in.
Selection and omission. The most effective bias is invisible because it operates on what is absent. Ask which perspectives are missing and who has not been quoted.
Placement and prominence. What appears in the headline and opening paragraph is what most readers take away, whatever appears later.
Attribution patterns. Claims from one side reported as fact and from the other as alleged reveal the slant clearly.
The burden of proof. Which position is treated as needing evidence, and which as obvious?
Unrepresentative sourcing. One articulate interviewee presented as speaking for a whole community.
Vested interest
Ask who benefits if you believe this. Research on a product funded by its manufacturer, a report on an industry's environmental record produced by that industry, a politician's account of their own administration — none is automatically false, but each has a reason for slant that an independent source does not.
Funding should be disclosed and often is, in small print. Where it is not disclosed, that is itself information.
Evaluating statistics
Figures carry an air of objectivity that deserves scrutiny.
Ask what exactly was measured and how the terms were defined, since unemployment and poverty can each be defined in ways that move the figure substantially. Ask who was sampled and how many. Ask whether a percentage rests on a base small enough to make it misleading — a rise from two cases to four is a hundred per cent increase. Ask whether a comparison uses a baseline chosen to flatter. And remember that an association between two variables is not evidence that one caused the other.
Corroboration and triangulation
The strongest single test of a claim is whether an independent source, working separately, reached the same conclusion. Independence is what makes corroboration worth anything, and it is the condition most often missing: a claim carried by a news agency and reprinted by twenty outlets has one origin, however many places it appears.
Triangulation extends the principle by seeking sources of different kinds rather than merely different names — an official statistic, an academic study and an account from people directly affected will each fail in different ways, so agreement between them is more informative than agreement between three of the same kind.
Where sources conflict, do not simply prefer the one that suits your argument. Ask which had better access to the evidence, which has a stake in the answer, which states its method, and whether the disagreement is about facts or about how facts should be interpreted. Reporting an unresolved conflict honestly is stronger than concealing it.
Online sources
The web removes the signals print carried. Anyone can produce a professional-looking page, and search ranking reflects popularity rather than accuracy.
Check the domain and who runs the site, look for an About page giving named people and an address, look for dates, and check whether claims are sourced. Be alert to content that exists to generate clicks or to sell, and to sites that imitate news organisations. Encyclopaedias and similar collaborative references are a reasonable starting point for orientation but are not citable evidence; use them to find the sources they cite.
Corroboration matters more online than anywhere else, and it must be genuine — several sites reproducing one original report are one source, not several.
Worked examples
Example 1: Evaluating an unseen source
Question: "Assess the reliability of the passage above as evidence about crime in the region." (10 marks)
Outline. Take the criteria in turn and apply each to the passage rather than defining it. Authority: who wrote it, and what standing do they have on this subject? Purpose: is it informing or persuading, and is a position declared? Evidence: are figures given, and are their sources identified? Currency: is the material recent enough for a fast-changing subject? Look for markers of slant — loaded terms, whose voices are quoted, what is absent. Then reach a judgement that distinguishes matters on which the source is dependable from those on which it is not, and say what corroboration you would seek. An answer concluding simply that the source is unreliable, without this separation, scores poorly; the marks lie in the qualification.
Example 2: Bias and point of view
Question: "Explain the difference between bias and a point of view, using examples." (8 marks)
Outline. A point of view is a declared position; bias is an undeclared leaning shaping selection and emphasis. Example: an environmental organisation arguing openly against a development states its position, and its figures can still be accurate and checkable. A newspaper report presenting itself as neutral while quoting only opponents of the development, calling it a scheme rather than a project, and omitting the employment case is biased, because the reader is not told a position is being taken. Add that bias does not make a source useless — it makes it a source requiring corroboration — and that the greater danger is bias unnoticed rather than bias declared.
Example 3: A contested claim
Question: "A student finds the same statistic on six websites and concludes it is well supported. Evaluate this reasoning." (6 marks)
Outline. Repetition is not corroboration. If all six trace back to one original report, the student has one source repeated six times, and the original's method and sampling remain unexamined. Genuine corroboration requires independent sources that gathered the evidence separately. The correction is to trace each occurrence back to its origin, then evaluate the original on authority, method and definitions. Note also that repetition online is easy and rapid, which makes the appearance of consensus cheap.
Common mistakes and how to avoid them
Treating evaluation as a verdict of reliable or unreliable. Judge a source for a purpose, and separate what it can support from what it cannot.
Confusing bias with falsehood. Biased sources frequently contain accurate information.
Confusing a declared point of view with concealed bias. Declaring a position is honest.
Accepting statistics because they are numerical. Ask what was measured, of whom, and against what baseline.
Counting repetition as corroboration. Trace claims back to their origin.
Defining the criteria instead of applying them. Under an unseen passage, marks come from application.
Ignoring omission. The most effective bias operates through what is absent.
How this links to your School-Based Assessment
Every source in your portfolio has to be defensible. Select material you can evaluate, record provenance with your notes, and prefer sources whose origin and purpose are identifiable.
Where a source has an obvious interest in the theme — an organisation campaigning on it, a company selling a solution to it — you may still use it, but say so in the exposition and corroborate the claims you take from it. Doing this visibly is a strength rather than an admission of weakness; it demonstrates exactly the judgement the syllabus asks for.
In the reflective piece, an account of a source you rejected and why is often more convincing evidence of evaluation than a list of the ones you kept.
Exam technique for evaluating reliability and bias
Apply criteria to the passage; do not define them in the abstract.
Quote the evidence for each judgement — the loaded word, the missing perspective, the unsourced figure.
Qualify your conclusion. Say what the source can be relied on for and what it cannot.
Watch the command word: identify wants the features named, explain wants how they operate, assess and evaluate want a judgement supported throughout.
Look at omission deliberately, since it is easy to overlook and frequently where the marks are.
Say what would strengthen the case — independent corroboration, a stated method, a larger sample — as a well-judged closing point.
Quick revision summary
Evaluate a source on authority, accuracy, objectivity, currency and coverage, always in relation to the purpose you have for it. A point of view is declared and honest; bias is an undeclared leaning shown in word choice, selection and omission, prominence, attribution patterns and where the burden of proof is placed; prejudice refuses evidence altogether. A biased source remains usable when the bias is identified and its claims corroborated. Ask who benefits if you believe a claim, since vested interest gives reason for slant. Treat statistics as claims requiring scrutiny of definitions, sample, base and baseline, and remember that association is not causation. Online, the usual signals of authority are absent, so check provenance directly and remember that many sites reproducing one report constitute a single source. In examination answers, apply the criteria to the material in front of you, cite the evidence for each judgement, and reach a qualified conclusion rather than a verdict.