Scientific Inquiry: Experimental Design, Data and Conclusions
What this domain tests
Inquiry questions ask you to judge an experiment rather than recall a fact: identify variables, spot what makes a test unfair, read a table or graph, and decide whether a conclusion actually follows from the data. They carry a large share of marks because they test reasoning that applies to all three sciences.
Designing a fair test
- The independent variable is what you deliberately change; the dependent variable is what you measure.
- Everything else - temperature, mass, light, time - must be held constant, otherwise the result cannot be attributed to the change you made.
- A control group is untreated or set to a standard value, giving a baseline for comparison.
- Repeat each measurement and average to reduce random error; a larger sample reduces the effect of odd individuals.
Interpreting data
Put the independent variable on the x-axis and the dependent on the y-axis, with units on both. Describe patterns with numbers: "as the distance doubled from 10 cm to 20 cm, the count roughly halved". Identify anomalous points and say why they might be excluded. Judge reliability by repeats and consistency, validity by whether the method actually measures what it claims.
Observation and conclusion
An observation is what you see or measure; an inference is the explanation you add. A conclusion must answer the original question, be limited to the range of data collected, and use "supports" rather than "proves".
Worked example
Pondweed is placed at 10, 20 and 30 cm from a lamp and bubbles are counted for one minute. Temperature, carbon dioxide supply and the same shoot are kept constant; each distance is repeated three times and averaged. Bubbles fall as distance increases. Conclusion: within 10–30 cm, lower light intensity reduces the rate of photosynthesis. Not claimed: anything about 50 cm, or about other plants.
Traps that cost marks
Changing two variables at once. Counting bubbles for different lengths of time. Ignoring an outlier with no comment. Writing "proves" from a single trial. Stating a conclusion that goes beyond the measured range.