🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
flawed_comparison
Flawed comparison occurs when media coverage juxtaposes two entities, events, statistics, or claims in a way that implies equivalence or contrast where none meaningfully exists. The comparison may involve different scales (comparing absolute numbers with rates), different time periods (cherry-picked baselines), different populations (demographics that differ on confounding variables), or different definitions of the same term across contexts.
A business story compares a company's quarterly profit margin to a competitor's annual margin, presenting both as 'profit margin' without clarifying the time period difference — making the company appear more or less profitable than the comparison actually supports.
A political story compares crime rates across cities using absolute numbers rather than per-capita rates. A city of 5 million is shown to have 'more crime' than a city of 500,000 — a comparison that tells us nothing about safety and actively misleads readers about relative risk.
A healthcare story compares mortality rates from a new treatment tested on elderly patients to rates from an old treatment tested on younger patients, attributing the difference entirely to the treatment — without acknowledging that the structurally different populations make the comparison invalid.
Binary (yes/no) questions an LLM must answer to identify this aspect:
Does the coverage compare two or more things to make a point about relative size, quality, risk, or value?
Type: binaryAre the compared items substantively different in ways that make the comparison misleading — different scales, time periods, populations, definitions, or contexts?
Type: binaryWould an accurate comparison require qualifications, context, or additional data that is absent?
Type: binaryDoes the flawed comparison serve a rhetorical purpose — making one option look better or worse — rather than informing understanding?
Type: binaryFlawed comparison occurs when media coverage juxtaposes two entities, events, statistics, or claims in a way that implies equivalence or contrast where none meaningfully exists. The comparison may involve different scales (comparing absolute numbers with rates), different time periods (cherry-picked baselines), different populations (demographics that differ on confounding variables), or different definitions of the same term across contexts.
Audiences process comparisons quickly as evidence of relative worth or risk. Specifying that a comparison is invalid requires work: checking definitions, time periods, and populations. The flawed comparison exploits the cognitive shortcut of 'A vs. B' presentation.
Ask: are the compared items using the same definitions, time periods, populations, and measurement methodologies? What would be a fair comparison — and does that comparison produce the same conclusion? What context is missing that would change the impression the comparison creates?
Common in economic reporting (GDP comparisons across countries without PPP adjustment), health coverage (risk comparisons using absolute vs. relative risk without clarification), and political coverage (comparing records from different time periods or governing conditions).
Use these tools to detect, analyze, or train this aspect.