IA idea · Statistics & hypothesis testing
Is crime seasonal? Testing monthly crime counts
Research question
Are recorded [one crime type] offences in [one police force area] spread evenly across the months of the year once month length is allowed for, and is the pattern the same for a second crime type?
Adapt it: change the place, the data or the comparison until the question is yours.
Why it makes a good exploration
“Crime by day of the week” is a popular idea, but the UK's open street-level data is published by month, not by day — so check what your data can support before you plan. Testing seasonality with monthly counts is the honest, answerable version, and comparing two crime types gives the chi-squared test real meaning.
The mathematics you'll need
- Chi-squared goodness-of-fit test with expected counts proportional to days in each month
- Chi-squared test of independence (month × crime type)
- Degrees of freedom and expected frequencies
- p-values and significance levels
Course labels show where a technique sits; using maths from outside your course is fine if you explain it clearly and say it is new to you.
Where the data comes from
Download street-level crime CSVs from data.police.uk for at least two full years for one force area.
- data.police.uk — Street-level crime and outcomes for England, Wales and Northern Ireland by month (OGL v3).
Cite every source in a footnote where you use it and in your bibliography. Check the licence of any dataset you download.
A possible outline
- Choose the crime types and area and explain why.
- Check the data's date resolution and reporting notes.
- Test one crime type against a days-in-month expectation.
- Test independence between month and crime type.
- Reflect on recording practices, holidays, daylight and weather.
Pitfalls that cost marks
- Assuming exact dates are available — the open data is monthly.
- Using equal expected counts for 28- and 31-day months.
- Treating recorded crime as all crime.
Showing personal engagement
- Choose an area you know and explain local factors (tourism, universities, nightlife).
- Predict which months will stand out and why, then test.
- Relate your finding to daylight hours or temperature data.
See Criterion C: personal engagement for what examiners look for.
Taking it further
Correlate monthly counts with monthly temperature from the Met Office and discuss confounding (holidays, daylight).