難関大向け｜英語長文問題124（統計）706語＋和訳・解答解説

When people are warned about fake news, they are often told to check whether a claim is supported by statistics. Numbers appear to offer protection against deception because they seem more objective than opinions or emotional language. Yet the presence of data does not automatically make a statement trustworthy. A misleading article may quote a real survey, present an accurate percentage, and still create a false impression. The central difficulty, therefore, is not simply deciding whether a number is correct, but understanding what that number actually represents, how it was produced, and what conclusions it can reasonably support.

Consider a headline claiming that a certain habit “doubles the risk” of an undesirable outcome. The statement may be mathematically accurate, but its meaning depends heavily on the original level of risk. If an event occurs in one person out of ten thousand and later occurs in two, the relative increase is one hundred percent, even though the absolute increase remains extremely small. A reader who sees only the phrase “twice as likely” may imagine a dramatic danger that the underlying data do not justify. This does not mean that relative figures are dishonest; in some contexts, they are useful. The problem arises when a statistic is separated from the scale needed to interpret it.

A similar difficulty appears when averages are used to summarize groups. Suppose a report states that the average income in a community has risen substantially. A reader might reasonably assume that most residents have become better off. However, if a small number of very high earners have moved into the area, the average can rise even while the income of the typical resident remains unchanged. A median figure might tell a different story. Neither measure is automatically superior: each answers a different question. What matters is whether the statistic matches the claim being made. Fake or misleading news often succeeds not by inventing data, but by selecting the measure that produces the most persuasive impression. This is why two articles based on the same dataset can lead readers toward sharply different conclusions without either article containing a fabricated number.

Even apparently careful readers can be misled when they focus too narrowly on the source of a number. Checking whether a statistic comes from a university, government agency, or respected research organization is sensible, but authority alone cannot settle the issue. A reliable organization may publish a finding with important limits, while a later article removes those qualifications. A survey, for example, may accurately describe the people who answered it but not the entire population that readers assume it represents. If participation was voluntary, those with especially strong opinions may have been more likely to respond. The original researchers may state this clearly, yet the warning can disappear as the result is repeated across websites and social media.

Another source of confusion is visual presentation. Graphs can distort judgment even when every plotted value is correct. A vertical axis that begins near the lowest value rather than at zero can make a modest change look enormous; a wide time interval can hide short-term instability; and two lines placed on the same graph may encourage readers to infer a relationship that has not been established. Such choices are not always evidence of deliberate manipulation. Designers may shorten an axis to make small differences visible or combine variables to invite comparison. Nevertheless, readers should ask what impression the graphic creates before deciding what the data prove. Statistical literacy therefore requires attention not only to calculation but also to framing.

The most useful response to suspicious statistical claims is not to distrust all numbers. That would replace one form of poor reasoning with another. Instead, readers should develop a habit of asking layered questions: What exactly was measured? Compared with what? Over what period? Who was included or excluded? Is the claim about correlation being presented as if it proved cause? Could another valid measure produce a different interpretation? No single question can identify every misleading article, and even experts sometimes disagree about how evidence should be weighed. Still, statistical skepticism is strongest when it is selective rather than cynical. The goal is not to reject data, but to understand the conditions under which data deserve confidence.

Q1. What does the first paragraph suggest about evaluating numerical claims?

1. The mere presence of numerical evidence makes misleading claims less likely.
2. Readers should focus mainly on whether reported figures are mathematically correct.
3. Reliable interpretation requires examining both data origins and warranted conclusions.
4. Objective-looking figures usually protect readers from emotionally framed misinformation.

Q2. What can be inferred from the fourth paragraph about source reliability?

1. Trusted institutions can prevent later reports from overstating limited findings.
2. Source credibility matters, but later reporting may erase crucial qualifications.
3. Voluntary surveys become dependable when respected organizations publish their results.
4. Researchers’ warnings usually remain intact as findings spread through media.

Q3. Which statement is NOT supported by the passage?

1. Careful skepticism asks several contextual questions before accepting statistical claims.
2. Established authorities generally remove the need to inspect methodological limitations.
3. Different statistical measures can produce contrasting impressions from identical datasets.
4. Large relative changes may correspond to very small absolute increases.