The big idea
Whenever a measurement is made, there is always some uncertainty about the result: the true value could be a little higher or lower. Chemists repeat measurements, look at how spread out the results are, and use that spread to say how certain the mean is.
Why measurements are uncertain
- Resolution: an instrument can only be read to its smallest division, so a reading could be up to half a division away from the true value.
- Random errors: small, unpredictable differences between repeats, such as judging when a colour changes.
The distribution of results
Listing or plotting all the repeat results shows their distribution: how spread out they are. Results that are close together are precise. A result that does not fit the pattern is anomalous and is left out of the mean.
- Mean: add up the results you keep and divide by how many there are.
- Range: the largest value minus the smallest value.
- Uncertainty: half the range, either side of the mean. Write the result as mean ± uncertainty.
\[\text{uncertainty} = \dfrac{\text{range}}{2}\]
Mean and uncertainty
Five students did the same displacement reaction and weighed the copper formed: 1.24 g, 1.28 g, 1.25 g, 1.41 g and 1.27 g. The equation predicts 1.27 g of copper. Give the result as mean ± uncertainty, and say whether it agrees with the prediction.
1.41 g is far above the others, so it is anomalous: leave it out.
Mean: \(\dfrac{1.24 + 1.28 + 1.25 + 1.27}{4} = \dfrac{5.04}{4} = 1.26\) g
Range: \(1.28 - 1.24 = 0.04\) g, so the uncertainty is \(0.04 \div 2 = 0.02\) g
Result: 1.26 ± 0.02 g. The true value is likely to lie between 1.24 g and 1.28 g.
The predicted 1.27 g lies inside this range, so the result agrees with the prediction within its uncertainty.
Using uncertainty
- A smaller uncertainty means a more precise result. An instrument with a higher resolution reduces it.
- Two results are only clearly different if their ranges (mean ± uncertainty) do not overlap.
- Precise is not the same as accurate (close to the true value). A balance that has not been set to zero shifts every reading equally: the mean is wrong but the range is not affected.
Common mistakes
- Giving the whole range as the uncertainty. The uncertainty is half the range.
- Including an anomalous result in the mean.
- Saying repeats make a test "fair". Repeats reduce the effect of random errors and show the spread; a fair test is about controlling variables.
- Writing just "human error". Say what the error is, such as misreading the scale.
Exam tip
When you compare two sets of results, use the numbers: quote each mean ± uncertainty, then say whether the ranges overlap. Give the mean to the same number of decimal places as the data.