Scientific Reasoning and Experimental Design
Experimental design questions ask whether a conclusion actually follows from the evidence. You will identify variables, controls and sources of error, and select the graph or statistic that matches the design. The exam favours precise language: "associated with" is not "caused by".
Core rules and concepts
- Independent variable: what the experimenter changes. Dependent variable: what is measured. Controlled variables: held constant.
- A control group receives no treatment or a standard treatment, isolating the treatment effect.
- Replication reduces random error; randomization reduces systematic bias.
- Correlation alone does not establish causation.
- Use a line graph for continuous trends and a bar chart for categories; report uncertainty with each measurement.
Worked example
A student tests a fertilizer by growing one plant with fertilizer and one without, both on the same windowsill. The fertilized plant grows taller. Can the student conclude the fertilizer caused the difference?
Reasoning: light and location are controlled, but there is only one plant per group, so random variation in seed quality, watering or soil is uncontrolled. The conclusion is not justified. A stronger design uses many plants per group, randomly assigns treatments, and standardizes water, soil and pot size.
Traps that cost marks
- Naming the measured outcome as the independent variable.
- Treating replication as a control group.
- Concluding causation from a correlation.
- Ignoring a confounding variable such as temperature or light.