How to Evaluate Sources: Reliability, Usefulness, and Bias

📊 How to Evaluate Sources: Reliability, Usefulness, and Bias

Every source in the pre-release booklet has strengths and limitations. One of the most heavily tested skills in the short-answer questions (4-mark questions) is source evaluation — being able to explain why a source is reliable or unreliable, useful or limited, and whether it shows signs of bias.

These three terms are related but distinct:

Reliability

Reliability is about how trustworthy the data is. A reliable source collects data in a rigorous, repeatable way; is produced by an organisation with relevant expertise; is corroborated by other independent sources; and is recent enough to reflect current conditions. Ask:

• Who produced it? Government agencies (e.g., Environment Agency, ONS) follow strict data collection protocols. A developer's marketing brochure does not.
• When was it produced? Data from 2010 may not reflect the situation in 2024. Coastal erosion rates, population figures, and land use all change.
• How was it collected? A scientific survey using standardised equipment is more reliable than a questionnaire where respondents select their own answers.
• Can it be corroborated? If two independent sources show the same erosion rate, that is more reliable than one source claiming it.

Example evaluation sentence: "Source A, produced by the Environment Agency, is likely to be highly reliable as it is a government agency with statutory responsibility for monitoring coastal erosion and uses standardised measurement methods. However, the data was collected in 2018 and may not reflect the acceleration in erosion rates that Source B suggests has occurred since 2020."

Usefulness

Usefulness is about how relevant the source is to the specific enquiry. A source can be entirely reliable but still not useful — air quality data collected with scientific rigour is unhelpful when the question is about pedestrian safety. Ask:

• Does it address the right scale? A national-level flood risk map may not show enough detail to evaluate risk to a specific street.
• Does it address the right time period? Long-term trends (50-year erosion data) may be useful for predicting future change but less useful for showing current conditions.
• Does it answer the question? If the issue is about biodiversity impacts, a source about economic costs is not directly useful — even if it is accurate.

Example evaluation sentence: "Source C (a photograph of the cliff face) is useful for showing the scale and visual evidence of erosion, but it is limited because it shows conditions at one moment in time and does not reveal the rate of erosion or how this has changed over time. Source D (a graph of annual erosion rates 2010–2024) is more useful for quantifying the extent of the problem and supporting a decision about urgency."

Bias

Bias means a source systematically presents evidence in a way that favours one viewpoint. Bias does not necessarily mean the information is wrong — a biased source can contain accurate facts — but it means the selection of evidence, the language used, and what is omitted are shaped by the producer's interests or values. Ask:

• Who benefits if people believe what this source says?
• What has been left out? A developer's report on a new road scheme will emphasise economic benefits and probably downplay environmental costs.
• What language is used? Emotive language ("devastating loss of habitat", "vital economic lifeline") signals the producer's viewpoint.
• Does it acknowledge counter-evidence? A balanced, objective source addresses the strongest objections to its conclusions; a biased source does not.

Example evaluation sentence: "Source E, produced by the Campaign to Save Our Coast, is likely to show significant bias in favour of hard engineering — the organisation was founded by cliff-top homeowners facing property loss, so it has a strong financial interest in presenting evidence that supports sea wall construction. Even if the statistics it cites are accurate, they have been selectively chosen to support one outcome, and the source omits evidence about beach loss from wave reflection that would weaken its case."