Evaluating experiments: errors, anomalies and improvementsSpec S5
In short
Evaluating an experiment means judging how good the method and data are, then suggesting specific improvements. Identify any anomalous result, a result that does not fit the pattern, and leave it out of the mean. Name sources of error and exact fixes, such as using a syringe instead of a beaker. Human error or being more accurate scores nothing.
Evaluating means judging how good the method and the data are, then suggesting specific improvements. Paper 6 often describes an experiment with weaknesses for you to spot.
- Anomalous result
- A result that does not fit the pattern of the other results.
- Repeats
- Doing the same test again, so that anomalous results can be spotted and a mean calculated.
- Replicates
- Several identical set-ups tested at the same time, such as five seeds in each dish.
- Source of error
- Something in the method or apparatus that makes the measured values differ from the true values.
Dealing with anomalous results
- Identify the anomalous result by comparing it with the repeats and the trend.
- Repeat that test if you can.
- Leave it out when you calculate the mean.
- Ignore it when drawing a best-fit line.
Calculating a mean with an anomalous result
A student counts bubbles from pondweed in three trials: 21, 9 and 23 bubbles per minute. Calculate the mean.
- 9 is much lower than 21 and 23, so it is anomalous and is left out.
- Add the remaining values: 21 + 23 = 44
- Divide by the number of values used: 44 ÷ 2 = 22
Answer: 22 bubbles per minute
Sources of error and improvements
| Source of error | Improvement |
|---|---|
| Only one result for each value | repeat each test at least three times and calculate a mean |
| Temperature not controlled, or a lamp heats the water | use a thermostatically controlled water bath, or put a heat shield (a beaker of water) between the lamp and the plant |
| Colour end-point judged by eye | compare against a colour standard, or test more often (every 10 s instead of every 30 s) |
| Volumes measured with a beaker | use a syringe or measuring cylinder |
| Bubbles of different sizes counted | collect the gas in a gas syringe or measuring cylinder and measure its volume |
| Pieces of tissue different sizes | cut pieces of the same size with a scalpel and ruler, or measure their mass with a balance |
| Too few values or too wide a range | use at least five values, with smaller intervals near the peak |
Also check the control. If there is no control, suggest one. A conclusion must be justified with data: state the pattern, quote values, and say whether the results support the prediction.
'Human error' and 'be more accurate' score nothing. Name the exact problem and the exact fix, such as 'use a 10 cm³ syringe instead of a beaker'.
Written and checked against the Cambridge IGCSE Combined Science (0653) specification · Updated October 2026