IA idea · Statistics & hypothesis testing
Do students who sleep less react more slowly?
Research question
Is there a significant correlation between hours slept the previous night and simple reaction time among students in my year group, and does the relationship hold after removing students who drank caffeine that morning?
Adapt it: change the place, the data or the comparison until the question is yours.
Why it makes a good exploration
Sleep matters to every student, and reaction time is measurable in seconds rather than by opinion. The confounding variable (caffeine) gives you a genuine methodological decision to reflect on.
The mathematics you'll need
- Scatter diagrams, Pearson's r and Spearman's rank (AI)
- Linear regression and interpretation of the gradient
- Hypothesis test for correlation or a t-test comparing groups (AI)
- Chi-squared test of independence on categorised data (AI)
- Sampling methods and bias
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
With consent, measure each participant's reaction time (mean of 5 trials on the same online test and device) and record self-reported sleep and caffeine; aim for 40+ students.
- phyphox (RWTH Aachen) — Free app that turns your phone's accelerometer, microphone, barometer and gyroscope into data loggers with CSV export.
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
- Define variables precisely (what counts as sleep, how reaction time is measured).
- Justify sampling and sample size; collect data ethically.
- Describe the data, then compute r and r_s; interpret both.
- Test significance; split by caffeine and compare.
- Reflect on self-reporting, the test device and causation.
Pitfalls that cost marks
- Claiming causation from correlation.
- Using different devices (screen lag varies).
- Skipping the conditions for the test you use.
Showing personal engagement
- Test yourself on nights of different sleep as a mini case study.
- Predict the strength of the relationship before collecting.
- Explain why Spearman's and Pearson's r differ in your data.
See Criterion C: personal engagement for what examiners look for.
Taking it further
Fit a regression with a caffeine indicator, or use a t-test to compare reaction times of students above and below 7 hours of sleep.
See it done
Our annotated exemplar Do students who sleep less react more slowly? (AI SL) explores a question like this one, with an examiner's comment on every criterion.