Source Analysis practice questions.

An essay written in 2025 supports a claim about how teenagers use social media with a survey conducted in 2015. What is the main limitation of this evidence?

• surveys are a weaker method than interviews for studying teenage behaviour
• a decade of platform change means the figures may no longer describe the present
• a survey from 2015 is far too recent to count as a proper historical source
• teenagers are too varied a group for any single survey to describe usefully

Answer: a decade of platform change means the figures may no longer describe the present

Currency is a test of its own, separate from origin and purpose: a source can be honest, competently made and simply out of date, and platform use is among the fastest-ageing evidence there is. The typical error is checking who made a source while never checking when.

A news photograph shows a crowded square during a protest against a facial recognition rollout. What can the photograph, by itself, establish?

• that most residents of the city oppose the facial recognition rollout
• that the protest was the largest that the city had seen that year
• that at that moment, from that angle, the square appeared crowded
• that the protesters' claims about the technology itself are accurate

Answer: that at that moment, from that angle, the square appeared crowded

An image evidences one framed moment: crowd size across the day, typicality of the turnout and the truth of the cause all need other sources. The typical error is reading a photograph as a measurement of public opinion, which is precisely why campaigns choose their angles.

A briefing on smart home devices was published in 2023 by an industry association whose members manufacture those devices. Its only evidence is a survey of 2 000 device owners recruited from the manufacturers' own customer mailing lists. Using the provenance alone, which claim can the briefing most safely support?

• What a group of device owners said in 2023 about their own use
• The share of households in that country that owned a device in 2023
• Whether smart home devices reduce the electricity use of a household
• Why households that own no smart home device decided against buying one

Answer: What a group of device owners said in 2023 about their own use

A sample drawn from a manufacturer's customer list contains only buyers, so it can describe those buyers and nobody beyond them. Origin and purpose do not make a source worthless; they fix the range of claims it can carry.

A 2024 report on the effects of ride-hailing apps in a large city was commissioned by an association of ride-hailing operators. It draws on the operators' booking records, a survey of app passengers and interviews with forty drivers. Whose perspective is missing from its evidence base?

• Passengers, whose views were gathered through the satisfaction survey
• Drivers who work through the apps, forty of whom were interviewed
• Residents of the busiest pick-up streets who never use the apps
• Operators, whose booking records supply the report's quantitative basis

Answer: Residents of the busiest pick-up streets who never use the apps

Non-users can carry the costs of a system without ever appearing in its data, because booking records and user surveys cannot reach them. Naming the stakeholder the evidence cannot see is a sharper limitation than saying the report is biased.

A city council reports that 91 per cent of respondents were satisfied with its new online permit service. The survey was a link shown on the confirmation screen after an application had been submitted successfully. Which flaw does that way of distributing the survey introduce?

• Respondents may have forgotten the details of their application by then
• Applicants who could not finish an application never saw the survey
• The number of applicants surveyed was probably too small to be reliable
• Satisfaction was recorded as one figure instead of a range of scores

Answer: Applicants who could not finish an application never saw the survey

The sampling frame contains only successful applicants, so the people most likely to be dissatisfied are excluded by the design. A larger sample collected the same way would repeat the distortion rather than correct it.

A 2023 study analysed four million public posts on a single microblogging platform and reported how quickly unverified claims spread there. A candidate cites the study as evidence for how quickly unverified claims spread online. Which correction does the citation need?

• Four million posts is far too small a sample to support any conclusion
• Research based on written posts can say nothing about how quickly material spreads
• The study should have been repeated on a more recent sample before being cited at all
• The finding covers one platform's public posts, not online communication generally

Answer: The finding covers one platform's public posts, not online communication generally

The population studied was one platform's public posts, so the conclusion has to be scoped to it; users, features and moderation rules differ from platform to platform. The problem is the reach of the claim, not the size of the sample.