Case 4 · correlation · lurking variables · Core Maths
Flu tests and maths answers
1Ask
Every week the UK Health Security Agency publishes the percentage of flu tests in England that come back positive. This autumn it rose almost every week. So did the number of questions answered on MaffsGames. Do they move together, and if they do, why?
2Collect
| Week beginning | Answers | Flu tests positive (%) |
|---|---|---|
| 7 Sep | 1,141 | 3.4 |
| 14 Sep | 875 | 3.8 |
| 21 Sep | 2,291 | 5.0 |
| 28 Sep | 2,846 | 6.5 |
| 5 Oct | 2,169 | 6.2 |
| 12 Oct | 1,887 | 6.4 |
| 19 Oct | 4,368 | 6.6 |
| 26 Oct | 1,890 | 7.8 |
| 2 Nov | 5,326 | 8.1 |
Flu: UK Health Security Agency, national flu and COVID-19 surveillance reports, 2026 to 2027 season (laboratory flu positivity, England, the 7 days to each Sunday, latest published figure). Answers: our own anonymous log.
3Picture it
Answers in a week against flu tests positive
4Decide
The correlation coefficient, r, measures how closely two sets of numbers follow a straight line. It runs from −1 (a perfect line going down) through 0 (no straight-line pattern) to +1 (a perfect line going up). For these two, r = 0.70. Does one cause the other? Neither: both follow the school term. So we don't read a rising line as growth until we can compare the same weeks next year.
What this data can't tell us
- Nine weeks is very few points.
- Flu positivity is the share of tests that are positive, not the number of people with flu.
- Our own testing is in the numbers.