Analysing conclusions and evaluating experiments
Show/Hide Sub-topics (Practical Skills | O Level Physics)
Learning outcomes
- Use processed data and physical principles to draw supported conclusions and predictions.
- Identify significant errors, explain their effects and propose specific matched improvements.
1. Definition
Analysis, conclusions and evaluation (ACE) means processing evidence, deciding what relationship it supports, identifying important limitations and proposing improvements that directly address those limitations.
2. Key Ideas
- Describe a relationship with direction and form: increasing, decreasing, linear, proportional or non-linear.
- Support a conclusion using a calculated quantity, gradient, intercept or comparison.
- Do not claim proportionality unless the evidence supports a straight line through the origin.
- Distinguish random scatter from a systematic offset.
- State how an error affects a result when its direction can be deduced.
- Match each improvement to the named source of error.
- Keep predictions within the measured trend unless a physical model justifies extrapolation.
3. Detailed Explanations
A. Move from pattern to conclusion
“Extension increases with force” describes a trend. “Extension is directly proportional to force because the best-fit graph is a straight line through the origin within experimental scatter” is a supported conclusion. If the line has a significant intercept, do not ignore it; consider zero offset, initial extension or whether proportionality is inappropriate.
B. Use evidence, not impressions
Compare repeated values, ratios, gradients or percentage differences. A conclusion should answer the stated question and refer to the evidence that matters. Avoid claiming that a small difference is significant when it is comparable to the measurement scatter or resolution.
C. Evaluate an error as a chain
Use source → effect → improvement:
| Significant source | Effect on evidence | Matched improvement |
|---|---|---|
| hand timing varies | repeated times scatter | time more cycles, repeat and use the mean |
| constant zero error | every reading shifted | measure the zero offset and apply a correction |
| parallax from changing viewpoints | readings may scatter or shift | read perpendicular to the scale using a fixed eye position |
| resistor heats | resistance changes during the run | use smaller current and open the switch between readings |
| heat escapes to surroundings | calculated thermal quantity may be biased | insulate, use a lid and reduce transfer time |
Repeating cannot remove a constant zero error. A higher-resolution instrument cannot correct a flawed alignment. The improvement must target the mechanism.
D. State limitations honestly
An experiment may support a relationship only over the measured range. A single anomalous result does not automatically disprove the trend, but it should be checked. If no accepted value is provided, do not invent one to claim accuracy.
4. Common Mistakes
- Calling any upward graph “directly proportional.”
- Restating raw readings instead of interpreting them.
- Writing “human error” without a specific action or measurement.
- Suggesting repeats for a systematic offset.
- Naming an improvement without explaining how it reduces the error.
- Claiming an effect direction that cannot be deduced.
- Extending a conclusion far beyond the measured range.
5. Exam Tips
For an evaluation mark, write one complete sentence containing all three links: “Because …, the measured … is likely too high/low or scattered; therefore … would reduce this by … .” If the direction is genuinely uncertain, state that it increases scatter rather than guessing high or low.
6. Worked Examples
Example 1: Evaluating an irregular-solid density methodCore
Air bubbles remain attached when an irregular solid is submerged. The displaced volume is then too large, so ρ = m/V is too small. Wet the object, lower it slowly and tap it to release bubbles before reading the meniscus at eye level. This improvement addresses the stated volume bias directly.
7. Mind Stretchers
Mind stretcher 1: Can a precise graph give an inaccurate result?Extension
Yes. Closely grouped readings can produce a neat line while every value has the same calibration offset. Precision concerns agreement; accuracy concerns closeness to the accepted value.
8. Practice and next step
Take three vague improvements—“repeat,” “use better apparatus,” and “avoid human error”—and rewrite each as a source–effect–improvement chain. Then return to the O-Level Physics Practical Hub for applied contexts.
Recommended next step
Analysis, conclusions and evaluation: practice
Why this will help: Use one focused question set to check that you can apply the lesson without prompts.
About 10 minutes
- O Level
- Physics
- Practical Skills
- Evaluation