IA idea · Statistics & hypothesis testing

Do TV series really get worse over time?

AI SLAI HL Accessible Also in: Art & music

Research question

Across [several long-running series], is there a significant negative correlation between season number and mean episode rating, and are final seasons rated significantly differently from earlier ones?

Adapt it: change the place, the data or the comparison until the question is yours.

Why it makes a good exploration

Fans often say shows “go downhill”. IMDb's public datasets let you test that claim across many series rather than arguing about one.

The mathematics you'll need

Spearman's rank correlation
Linear regression
Two-sample t-test for final vs. other seasons
Weighted means (by number of votes)

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

Use IMDb's non-commercial datasets (title.episode and title.ratings); credit IMDb as required.

IMDb non-commercial datasets — Ratings, votes, runtimes and episodes for every title (non-commercial use, credit IMDb).

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 a fair sample of series (e.g., all with 6+ seasons in a genre).
Compute season means.
Test for a trend within and across series.
Compare final seasons.
Reflect on who votes and review-bombing.

Pitfalls that cost marks

Choosing only famous flops (selection bias).
Ignoring the number of votes.
Treating episodes as independent.

Showing personal engagement

Include a show you love and one you gave up on.
Predict the result before analysing.
Compare genres.

See Criterion C: personal engagement for what examiners look for.

Taking it further

Model how vote counts decline by season and test whether rating decline is linked to audience decline.

Turn this idea into your IA

Plan this IAOpens the planner with this question, the maths and data sources filled in. Desmos workspacePlot your data, fit models and keep your graphs with your plan. AI feedback on a draftCriterion-by-criterion feedback with evidence from your own text. Coaching, not rewriting. Examiner reviewPete, a practising IA examiner, reads your draft and talks it through with you (€100).