OECD PISA 2022 ยท 81 countries ยท 690,000 students

690k

students. One question: what is screen time actually doing to us?

Exposition
Distraction in class
Rising action
Scores drop with time
Climax
The sweet spot revealed
Resolution
Connection & belonging

Part 1 โ€” Distraction

Is anyone paying attention?

Researchers asked students in 81 countries if they get distracted in class. Tap the classroom to reveal what they found.

Tap the classroom to reveal โ†’

OECD PISA 2022 ยท Figure II.3.4 ยท OECD average

Explore the world ๐ŸŒ

How distracted is your country?

Search any of the 81 countries and compare to the global average.

๐Ÿ“ฑ Self-distracted
avg 30.5%
๐Ÿ‘€ Distracted by classmates

OECD PISA 2022 ยท Figure II.3.4 ยท 81 countries and economies

Part 2 โ€” Scores

What does screen time do to your marks?

Drag the slider. Pick the activity. Watch the score change. ๐Ÿ”Š Listen too โ€” the pitch reflects the score.

โ‰ค1h
Score
480
Baseline
~490
Up to 1 hour is fine โ€” scores are near their peak here!

OECD PISA 2022 ยท Figure 3 ยท OECD average maths mean scores

Part 3 โ€” The sweet spot

Not all screen time is equal.

Scroll through the story below โ€” each step reveals a new line on the chart, built with D3.js.

01

Start with learning ๐Ÿ“š

When students use devices to study, scores actually rise โ€” peaking around 1โ€“2 hours a day, then plateauing. A little structured screen time helps.

Peak score: ~481 at โ‰ค1hr vs 456 with no device use at all.

02

Add gaming ๐ŸŽฎ

Gaming follows a similar early rise โ€” up to 1โ€“3hrs is actually linked to higher scores than no gaming at all. But the drop after 3hrs is sharp.

Best score: ~496 at โ‰ค1hr. Drops to 441 at 7+hrs.

03

Now social media ๐Ÿ“ฒ

Social media follows a similar arc early on โ€” but the decline is the steepest of all activities. The heaviest users score dramatically lower than moderate users.

Drops ~72 points from the โ‰ค1hr peak (501) to 7+hrs (429).

04

The full picture ๐ŸŒ

All four lines together reveal the pattern: there's a sweet spot around 1 hour, after which more time consistently means lower scores โ€” for every activity type.

The green zone marks the sweet spot. Hover any dot for exact scores.

OECD PISA 2022 ยท Figures 2 & 3 ยท OECD average maths scores ยท Chart built with D3.js v7 + Scrollama

Test yourself!

How much did you pick up?

Five questions, all based on real data. No tricks.

Bonus โ€” how do you feel?

It's not just about grades.

The same study asked students how connected they feel at school. Drag and watch the mood shift.

๐Ÿ˜Š
Feeling connected!
Up to 1 hour of leisure โ€” students still feel pretty connected to their school community.
Belonging index: +0.02 (OECD avg)
๐Ÿ’›
What is belonging?
PISA asked if students feel liked, happy and like they fit in at school. The index goes from negative (lonely) to positive (connected).
๐Ÿ“ฑ
Leisure drops it
Students using devices for leisure 3โ€“5 hrs score โˆ’0.066 on belonging โ€” noticeably lower than the 1-hour group.
๐Ÿ“š
Learning barely moves it
Learning screen time has almost no effect on belonging โ€” the index stays near zero across all brackets.

OECD PISA 2022 ยท Tables 43 & 44 ยท OECD average sense of belonging index

Critical data practice

What this data can't tell us.

Week 9 of this course asks us to audit data visualisations critically. Here's an honest audit of this one.

๐Ÿ‘ค
Who is this data for?
PISA surveys 15-year-olds in schools. It excludes students who have dropped out, are home-schooled, or live in non-participating regions. Results reflect school systems, not all young people.
โ“
Correlation โ‰  causation
Students with lower scores may use screens more because they're disengaged โ€” not the other way round. The data shows association, not cause. This is a key limitation the OECD itself acknowledges.
๐ŸŒ
What's missing?
The dataset measures maths only. It doesn't capture creativity, wellbeing, social skills, or whether screen time content differs by class, gender, or race โ€” all of which would change the story.
๐Ÿ”„
One change I'd suggest
This site should let users filter by country income group. The relationship between screen time and scores may look very different in lower-income countries where device access is a privilege, not a distraction.

Applying feminist data practice principles โ€” D'Ignazio & Klein (2020) ยท Week 9, Hr3: Critical Data Practice