24. CR Logical flaw
Name the flaw family: correlation-causation, sampling, percent vs number, circularity.
Core ideas
- Name the flaw family first: correlation vs causation, unrepresentative sample, percent vs raw number, circular reasoning.
- Correlation-causation: also check for reverse causation and a third factor.
- Percent vs number: a bigger percentage of a small base can be fewer people than a small percentage of a huge base.
- Sampling: ask who was surveyed and whether they resemble the group in the conclusion.
Worked example 1
A columnist writes: Cities with more parks have residents who report better health. Clearly, building parks makes residents healthier, so Grelton should build ten new parks to improve public health.
The reasoning is most vulnerable to which criticism?
Show solution
Name the family: correlation treated as causation. The evidence pairs parks with health but never shows direction or rules out third factors. Reverse causation is plausible: healthier, more active populations may demand and fund more parks. A third factor also fits: wealthier cities can afford both parks and better healthcare. Without eliminating these, the leap to "building parks makes residents healthier" is unsupported. The correct answer states that the argument assumes a causal direction that the correlational evidence does not establish.
Worked example 2
Last year 8 percent of Torvel Airlines' flights were delayed, while 12 percent of tiny regional carrier Wexjet's flights were delayed. A travel blogger concludes that more passengers experienced delays on Wexjet than on Torvel.
The reasoning is flawed because it
Show solution
Name the family: percent versus raw number. The percentages apply to very different bases: Torvel is a major airline with vastly more flights and passengers, Wexjet is tiny. Eight percent of Torvel's huge flight count could easily involve far more delayed passengers than 12 percent of Wexjet's small count. The conclusion about numbers of passengers cannot follow from percentages of flights alone, especially across unequal fleet sizes and plane capacities. The correct answer: it draws a conclusion about absolute numbers from data given only as percentages of groups of unknown and unequal size.
No practice questions banked for this concept yet. Check back soon.