An anomaly is a result that does not fit the general pattern. Handling it transparently means you mark it, explain it carefully and show how much it changes your summary. This skill appears when you evaluate fieldwork results and when an unseen dataset contains one odd value.
It follows selecting a graph in fieldwork analysis and evaluation.
How do you spot and handle an anomaly?
- Look at the pattern first. Decide where most points sit and where the trend runs.
- Mark any point far from the trend with a circle or label on the graph.
- Check the record. Look for a copying error, a wrong unit or a field note about that site.
- Offer a possible reason, worded as a possibility.
- Show the effect. Calculate the mean and median with and without the point.
- Keep the result in your data. Say how it affects confidence in the conclusion.
Worked example
The data below is invented. It repeats the street survey, but site E now shows a much higher count than the sites on either side.
| Site | Distance (km) | Pedestrians in 10 min |
|---|---|---|
| A | 0.1 | 84 |
| B | 0.3 | 71 |
| C | 0.5 | 66 |
| D | 0.8 | 52 |
| E | 1.2 | 74 |
| F | 1.6 | 39 |
| G | 2.0 | 35 |
Spot it: the trend falls from 84 to 35, so a value near 45 would be expected at 1.2 km. The value 74 sits far above the line, so site E is the anomaly.
Mean with site E: 84 + 71 + 66 + 52 + 74 + 39 + 35 = 421, and 421 ÷ 7 = 60.1.
Mean without site E: 421 − 74 = 347, and 347 ÷ 6 = 57.8.
Medians: with all seven, the sorted values are 35, 39, 52, 66, 71, 74, 84, so the median is 66. Without E, the six values are 35, 39, 52, 66, 71, 84, so the median is (52 + 66) ÷ 2 = 59.
Explain it: the high count at E may be due to something near that point, such as a bus stop or a school entrance. This is only a possibility, so the write-up says it was not recorded and should be checked in a repeat visit.
Write the effect: the mean moves by 2.3 and the median by 7, so the anomaly changes the summaries but does not change the downward trend.
What mistake costs marks?
A common slip is to delete the anomaly and say nothing.
Mistaken answer: “Site E was wrong so I left it out.”
This hides evidence and gives no reason. The correction is to keep the point, mark it, state that the cause was not recorded, and report the figures with and without it. That is honest and shows you understand the data.
Check yourself
All data is invented.
1. Five readings of stream speed are 10, 12, 11, 30 and 13 cm per second. Which reading is anomalous?
Show answer
The reading 30 sits far above the others, which are all between 10 and 13.
2. Find the mean of the five readings with and without the anomaly.
Show answer
With: 10 + 12 + 11 + 30 + 13 = 76, and 76 ÷ 5 = 15.2. Without: 46 ÷ 4 = 11.5.
3. Write one honest sentence explaining the anomaly when your notes say nothing about it.
Show answer
For example: “The reading of 30 cm per second is anomalous, and its cause was not recorded; it may be due to a measuring slip or a local narrowing of the channel, which a repeat reading could check.”
Where this leads next
Handling anomalies well makes the conclusion stronger, because you can say how reliable it is. Continue with linking a conclusion to the investigation aim. The statistics and distribution explorer shows how a single extreme value moves a mean.
If you want a teacher to question your explanations the way an examiner would, see our online one-to-one Geography tuition.