Evaluating practical results and improving a method
ACE (Paper 3): identify random vs systematic errors, explain limitations, and suggest specific, realistic improvements that score marks.
On this page
A useful evaluation connects an observation to a possible cause and a specific improvement. Ask what the results support, how much they vary and whether the method could bias them.
What the skill involves
Analysis, conclusions and evaluation (ACE) is a practical assessment skill area. Analyse the results, draw a conclusion supported by them and evaluate the method by identifying limitations and possible improvements.
What you need to know
- Random error: unpredictable variation (repeats help).
- Systematic error: a consistent bias (repeats do not fix it).
- Accuracy: closeness to true value. Precision: closeness of repeats.
- Improvements must be specific and linked to the error.
Putting the skill into practice
- Accuracy compares a measurement with the true or accepted reference value; precision compares repeated measurements with each other.
- Repeats reveal the spread. Averaging suitable repeats reduces the influence of random variation, but does not remove a systematic bias.
- A valid investigation uses a method that answers its question and controls other relevant variables.
- Investigate an unexpected result, retain the original reading and repeat the trial. Exclude it only with a justified reason.
Random vs systematic error
| Type | What it looks like | Example | Best improvement |
|---|---|---|---|
| Random | repeats scatter around a value | reaction start time varies | repeat + mean; better timing |
| Systematic | readings biased in a consistent direction | balance reads +0.10 g always | calibrate/zero; check instrument |
Visual example (random scatter vs systematic offset):
Random vs Systematic Error (Example)
Example data showing random scatter around a true value versus systematic error with a consistent offset.
Scroll across the graph to read all labels.
View figure data
| Series | Trial number (unitless) | Trial number uncertainty | Measured mass (g) | Measured mass uncertainty |
|---|---|---|---|---|
| True value (2.00 g) | 1 | 2 | ||
| True value (2.00 g) | 5 | 2 | ||
| Random error (scattered repeats) | 1 | 1.97 | ||
| Random error (scattered repeats) | 2 | 2.03 | ||
| Random error (scattered repeats) | 3 | 1.99 | ||
| Random error (scattered repeats) | 4 | 2.02 | ||
| Random error (scattered repeats) | 5 | 1.98 | ||
| Systematic +0.05 g (precise but inaccurate) | 1 | 2.05 | ||
| Systematic +0.05 g (precise but inaccurate) | 2 | 2.05 | ||
| Systematic +0.05 g (precise but inaccurate) | 3 | 2.05 | ||
| Systematic +0.05 g (precise but inaccurate) | 4 | 2.05 | ||
| Systematic +0.05 g (precise but inaccurate) | 5 | 2.05 |
“limitation” vs “mistake”
- A limitation is a weakness in the method (even if you followed it correctly).
- A mistake is an error in carrying out the method. Explain how it affects the result and how to prevent it, rather than naming “human error” alone.
Matching an improvement to its cause
Bad: “reduce human error”.
Good (specific + linked):
- “Use a gas syringe instead of observing bubbles so volume can be measured accurately.”
- “Repeat the reaction at each temperature, record the spread and calculate a mean of valid trials. Investigate unusual results and repeat any trial with an identified fault.”
- “Use a burette to deliver a variable liquid volume with a smaller reading uncertainty than the available measuring cylinder.”
State which measurement you will repeat, the conditions you will keep the same and how you will use the results. Choose enough repeats to check their agreement; three trials can be a useful starting plan, rather than a rule for every experiment.
A constant offset can cancel when you subtract two readings. For example, if a thermometer reads 2°C too high both before and after mixing, each temperature is wrong but the calculated temperature change is unchanged. This cancellation relies on the offset being the same for both readings.
Avoiding common mistakes
- Writing “human error” with no specific cause.
- Suggesting impossible improvements (e.g., “use a more accurate stopwatch” without explaining timing method).
- Confusing accuracy with precision.
- Not stating the direction of bias when it is obvious (e.g., “gas leaked → volume too low”).
Explaining your method
“A limitation is … therefore the results may be … (too high/too low).”
“To improve accuracy/precision, … because …”
“Repeat the measurement under the same conditions, inspect the spread and
calculate a mean of valid results to reduce random variation.”
Worked examples
Modelled example 1
Leak in Gas Syringe Setup
Problem
Study the worked solution
Trace the lost gas
Method
State that some product gas escapes through the leak.Reason
Escaped gas never enters the syringe barrel.Working
Collected volume is less than the volume actually produced.State the bias and fix
Method
Conclude that readings are too low and make the apparatus airtight.Reason
A tight-fitting bung and checked tubing prevent systematic gas loss.Working
Bias: volume too low. Improvement: secure and leak-test all connections before starting.
Common misconception 2
Accuracy vs Precision
Learner claim
Judge agreement separately from truth
View solution step by step
Judge precision
Method
Call both sets precise.Reason
Each student’s repeated readings agree closely with one another.Working
A: precise; B: also precise.Judge accuracy
Method
Identify A as more accurate and B as systematically high.Reason
Accuracy compares with the true 2.00 g value; B is offset by + 0.05 g every time.Working
A is precise and accurate; B is precise but inaccurate.
Challenge 3
Inconsistent End-Point (Cross Disappears)
Evaluation transfer
Link the subjective observation to improvements
Hints
Hint 1: endpoint
Hint 2: consistency
View solution step by step
Identify the random source
Method
Attribute scatter to subjective endpoint judgement and reaction time.Reason
The precise stopping instant varies unpredictably from trial to trial.Working
Random error in observing and timing the cross disappearance.Improve consistency and reliability
Method
Use the same observer, viewing distance, lighting and depth of solution. Repeat under the same conditions and calculate a mean of valid trials.Reason
Controlled viewing reduces variation, while repeats reveal anomalies and reduce the influence of random error.Working
Standardise observation + repeat and mean.
Try it independently
Mind stretcher 1: Spot the Systematic ErrorExtension
Question: A thermometer reads 2°C too high for every measurement. Are repeats enough to fix this? What should you do?
Show Answer
Answer: Repeats do not fix a systematic error. You should calibrate/replace the thermometer or correct every reading by subtracting 2°C (if instructed/justified).
Mind stretcher 2: Weak ImprovementExtension
Question: A student writes: “Use a different method.” Why is this not a good evaluation answer, and what is a better one?
Show Answer
Answer: It is too vague and does not show understanding. A better answer states a specific method change, e.g., “Use a burette to deliver this variable volume, reducing reading uncertainty compared with the available measuring cylinder.”
Practise and check
See what you know across this topic, then go back to anything you got wrong.
Syllabus and review details
- SEC G3 Pure Chemistry 2027 · 2027
Content structure and subject content, PDF pages 9–24
Last reviewed: