Experimental Skills and Internal Assessment
This domain tests practical design, data processing, statistical analysis and evaluation, including the internal assessment. The exam expects precise variables, appropriate controls, correct uncertainty, suitable graphs and justified statistical tests.
Core rules and concepts
- Define the independent variable, dependent variable and controlled variables. Include a control group where a treatment effect must be isolated. Use random sampling, replication and standardised conditions to improve validity and reliability.
- Report uncertainty appropriately: for an analogue scale, use half the smallest division; for a digital display, use the smallest digit. Keep units and consistent significant figures.
- Summarise replicate data with a mean and a measure of spread such as standard deviation. Error bars show variation or uncertainty but do not by themselves show statistical significance.
- Choose a test that matches the data: chi-square for frequencies in categories, t-test for comparing two means, and correlation coefficient for the strength of a relationship. A significant result supports rejection of the null hypothesis; it does not prove causation.
- For the IA, justify a focused research question, relevant background, variables, method, ethics, data processing and evaluation of limitations.
Worked example
A genetic cross predicts a 1:1 ratio. Observed counts are 60 and 40, expected counts are 50 and 50. Chi-square = (60−50)²/50 + (40−50)²/50 = 2 + 2 = 4.0. With one degree of freedom, the 5% critical value is 3.84. Since 4.0 > 3.84, reject the null hypothesis at the 5% level: the observed ratio differs significantly from 1:1.
Traps that cost marks
Do not enter percentages into chi-square when counts are required. Do not state that a test “proves” a hypothesis. Avoid missing units, unexplained anomalies or an evaluation that merely lists limitations without saying how each affected the data.