Scientific Method
What the domain tests
The scientific method tests how claims are tested rather than merely asserted. You need to design fair tests, distinguish evidence from anecdote and recognise the limits of induction and models.
Core rules and concepts
- A hypothesis is a testable, falsifiable claim. A prediction follows from it; an experiment produces data that can support or refute it.
- Identify independent, dependent and controlled variables. Control confounds, randomise where possible, use adequate samples and operationalise terms so others can repeat the work.
- Correlation is not causation; a third variable may explain both. Peer review and replication are social checks on error and bias.
- Popper stressed falsifiability; Kuhn described paradigm shifts when anomalies accumulate. Induction (Hume) cannot be justified by simply assuming nature is uniform.
Worked example. Hypothesis: More light makes plants grow taller. Reasoning: measure height (dependent variable), manipulate light (independent variable), hold water, soil, temperature and species constant. Without controlling temperature, a warm bright room may be the real cause. Randomise pots, include a control group and report means with variation. A single tall plant is an anecdote, not evidence.
Traps that cost marks
- Accepting a claim because it is popular or told by an authority outside their field.
- Treating an unfalsifiable claim as scientific.
- Confusing a model with reality; models simplify.
- Ignoring sample size, confounds or selective reporting.