Inductive Reasoning: Analogy, Generalisation and Causation
Inductive strength
Inductive arguments make conclusions probable, not certain. Strength depends on sample size, representativeness, relevance of evidence, and absence of bias. An analogy argues that because two cases share known features, they probably share a further feature; the analogy is stronger when the cases are relevantly similar and dissimilarities are unimportant.
Generalisation and causation
Generalisation moves from observed cases to a population. Random, sufficiently large samples support it; biased or tiny samples do not. Causal reasoning uses Mill’s methods: agreement, difference, joint method, concomitant variation, and residues. Correlation alone does not establish causation.
Worked example
A survey finds 70% of 1,000 randomly selected students favour later library hours. The sample is large and random, so the generalisation to all students is moderately strong. But if only students entering the library were surveyed, the sample is biased and the conclusion is weak.
Traps that cost marks
- Calling a strong inductive argument valid in the deductive sense.
- Hasty generalisation from a few cases.
- Confusing correlation with causation.
- Ignoring a relevant disanalogy in an analogy.
- Treating a biased sample as representative.
- Overlooking alternative causes under Mill’s methods.