A good measurement improvement names one specific change and says why it reduces error. It stays at the level of ideas, such as measuring a larger quantity, using a finer scale or timing more cycles, and it never depends on anything unsafe.
This sits within investigations and evidence and builds on knowing which conditions were controlled. All data here is invented.
What kinds of improvement are there?
Three conceptual ideas cover most questions:
- Make the measured quantity larger, so a fixed error becomes a smaller fraction of it.
- Use a finer scale, so the smallest division is smaller.
- Reduce a human delay, for example by timing more events or using an automatic sensor if the school has one.
Always attach a reason, using the words “so that” or “because”.
Worked example
A student times one swing of a pendulum with a stopwatch and records 1.8 s. Human reaction time adds about 0.2 s of uncertainty to the stopwatch reading.
Step 1, find the fraction. 0.2 ÷ 1.8 = 0.111, so the error is about 11% of the reading.
Step 2, apply the idea. Time 10 swings together and divide by 10. Suppose 10 swings take 18.2 s.
Step 3, recompute. The same 0.2 s is now 0.2 ÷ 18.2 = 0.011, about 1.1%. One swing then takes 18.2 ÷ 10 = 1.82 s.
Step 4, write the answer. “Time 10 swings and divide by 10, because the 0.2 s reaction-time error is then about 1.1% of the reading instead of about 11%.”
The mistake to watch for
Mistaken answer: “Use a more accurate stopwatch and be more careful.”
“More accurate” and “more careful” name no change and give no reason. The correction is to say what is different (more swings), and why that shrinks the error (a smaller share of a larger time).
Check yourself
1. A 5 cm length is measured with a ruler marked every 1 cm, giving about 0.5 cm uncertainty. Suggest a conceptual improvement.
Show answer
Use a ruler with millimetre divisions, so the smallest division, and therefore the reading uncertainty, is smaller.
2. If one swing was timed as 1.8 s with a 0.2 s error, what percentage is that error?
Show answer
0.2 ÷ 1.8 × 100 = 11.1%, so about 11%.
3. Why is “heat the substance much hotter to get a clearer result” not an acceptable improvement?
Show answer
It could be unsafe and is not required by the idea. Safe improvements keep conditions within what the school would normally allow.
Where this leads next
Repeating measurements helps with random error, but not with a flaw built into the method. Learn why repeated readings do not fix a biased method, then try the scientific investigation critic.
A teacher can check your wording on real exam-style stems in online one-to-one Combined Science tuition.