Run a feed comparison investigation and present the findings

  Two people can open the same platform and see very different information. Personalisation can be useful, but it can also make each person’s online world feel more universal than it really is. You will compare feeds using a structured method, analyse the differences and decide what the evidence can — and cannot — tell […]

Share this challenge
Assign in
Classroom
Microsoft
Teams

Scan to open this challenge on any device

Loading...

Run a feed comparison investigation and present the findings

Your Mission
Run a feed comparison investigation and find out how different information environments can emerge.
You will design a fair comparison, collect structured data and separate what you observed from what you can actually claim about the algorithm.
The Shape of This Challenge
Discover
✏️ ️ Notice & question
Explore
️ Investigate & test
Create
✏️ Make & communicate
Reflect
️ Think & apply

Discover

Start by generating hypotheses without pretending you already know the cause.
Step 1

Two People, Two Internets?

️ Predict + question · 8–10 mins
We are learning to recognise that personalised feeds can expose people to different information.

Feeds are shaped by follows, clicks, watch time, location, popularity, platform rules and other signals. That does not mean every difference is caused by a secret political agenda.

Your Task
  1. Imagine two users with different interests.
  2. Predict five ways their feeds might differ.
  3. For each, mark whether the difference could come from user choice, platform ranking or both.
  4. Write one claim about feeds that would need evidence before you accepted it.
I can explain several reasons personalised feeds can differ.
→ This gives you hypotheses for the comparison investigation.
Critical ThinkingDigital CitizenshipAI Literacy
Step 2

Design a Fair Feed Comparison

✏️ Method builder · 8–10 mins
We are learning to create a comparison method that reduces obvious confounding factors.

A fair investigation needs comparable time periods, categories and sample sizes.

Your Task
  1. Choose what you will compare: teacher-provided screenshots, public feeds, controlled example profiles or another safe method.
  2. Create five categories such as topic, source type, format, emotion/tone and repeated viewpoint.
  3. Choose a fixed sample size, for example 20 items per feed.
  4. Write one limitation your method cannot remove.
I can design a structured and safe feed comparison with clear categories.
→ This becomes the investigation method for Step 3.
ResearchCritical ThinkingData Literacy

Explore

Collect comparable feed data and audit the claims you make from it.
Step 3

Collect and Compare the Feed Data

Choose your route · 15–20 mins
We are learning to collect structured evidence about differences between feeds.

Avoid interpreting too early. First record what appears.

Your Task
  1. Collect the agreed sample from each feed.
  2. Complete one analysis route.
  3. Record counts in a table or spreadsheet.
  4. Identify the three biggest differences.
⚡ Quick Track

Compare two feed samples across your five categories. Calculate simple counts or percentages and identify the largest difference.

Dig Deeper

Compare three feed samples or repeat the comparison at two different times. Look for patterns that persist and patterns that change.

I can use structured data to describe feed differences without overclaiming why they happened.
→ This gives your final investigation its evidence.
ResearchData LiteracyCritical Thinking
Step 4

What Can the Data Actually Prove?

️ Claim audit · 10–12 mins
We are learning to distinguish observation from explanation when analysing algorithms.

Seeing different feeds does not by itself prove why the platform produced them.

Your Task
  1. Write three observations from your data.
  2. For each, write a possible explanation.
  3. Label the explanation supported, plausible or speculative.
  4. Write one extra piece of evidence you would need to be more confident.
I can separate what feed data shows from what I am inferring about the algorithm.
→ This gives the final presentation appropriate caution.
Critical ThinkingEthical ReasoningAI Literacy

Create & Share

Present the findings with data, alternative explanations and clear limitations.
Step 5

Present a Feed Comparison Investigation

✏️ Analyse · Create · 25–35 mins
We are learning to communicate feed differences using data while being clear about limitations.

Your final piece should answer what differed, how you measured it, what might explain it and what you cannot conclude.

Your Task
  1. Include your method and sample size.
  2. Show at least three data comparisons.
  3. Include one plausible explanation and one alternative explanation.
  4. State at least one limitation.
  5. Choose one output below.
Feed Comparison Report
1 page · 250–300 words + chart/table
⏱ 25–30 mins

Summarise method, biggest differences, possible explanations and limitations.

Algorithmic Feed Briefing
5 slides · maximum 30 words per slide
⏱ 30–35 mins

Present method, data, interpretation, alternative explanation and takeaway.

Why Our Feeds Differ
90 seconds final runtime
⏱ 30–35 mins

Explain the investigation using one visual comparison and careful causal language.

I can present a data-based feed comparison without confusing correlation with proof of cause.
→ This is your finished mission output.
ResearchCritical ThinkingCommunication
Step 6

Reflect, Apply, Look Forward

️ Think or discuss · 5–8 mins
We are learning to reflect on how personalised information environments affect awareness and judgement.

The next challenge turns toward digital reputation: how the information other people find about us can shape their first impression.

Your Task
  1. Answer the three prompts.
  2. Name one feed habit that could widen the information you encounter.
  3. Write one conclusion your study could not justify.
Reflect

What was the most surprising feed difference?

Apply

How could you deliberately diversify the sources or viewpoints you encounter?

Look Forward

How might AI agents or personalised assistants make information filtering more powerful in the future?

I can explain how personalisation can influence exposure without assuming it fully controls what people believe.
→ You now have evidence-based language for talking about algorithmic feeds.
ReflectionDigital CitizenshipCritical Thinking
Dex
Dex AI Study Partner
* Click to generate
Hi! I'm Dex, your AI Study Partner. Ask me a question, attach a screenshot, test your skills with Quiz, or click Mastery for feedback!
Attachment Screenshot attached
🔒 Built with Google Gemini as a Study Partner trained strictly on this challenge content. No personal data is stored or shared.

TOOLKIT

Choose Your Tools

Use the tools available to you. The quality of the checking and reasoning matters more than the software.

Compare safely

Use public/teacher-provided screenshots, different browsers or controlled example profiles if personal accounts are not appropriate.

Record

Use a tally sheet or spreadsheet to classify sources, topics, emotions, formats and repeated viewpoints.

Create

Use a feed comparison report, charts/slides or recorded briefing.

Educator / Parent Notes

Age, Stage and Prior Learning: Designed for S1–S3. Learners should not be required to reveal or share personal feeds.
Before You Start: Use public examples, screenshots or controlled fictional profiles where needed. The goal is method and interpretation, not auditing individual learners' beliefs.
How to Open This: Ask: “If two people see different feeds, what could explain the difference?” Keep generating multiple hypotheses.
Scheduling: Steps 1–4 need 40–50 minutes; Step 5 needs 25–35 minutes.
If a Pupil Gets Stuck: Provide two pre-made feed samples of 15–20 items and three fixed categories.
For Fast Finishers: Ask learners to calculate percentages, repeat the sample later or design an experiment that could distinguish two competing explanations.
Marking Guidance: Look for consistent coding, transparent sample/method, data-based comparisons, alternative explanations and careful distinction between correlation and causation.

More like this

Multiple Sessions
60–90 minutes
60–90 minutes