Liu et al. (2023)

Question 1

Study’s aim

Answer 1

Investigate how different recommendation algorithms influence user behaviour and interests on online platforms. Focuses on Zhihu and examines how shifting from a content-based filtering algorithm to a social filtering algorithm impacts user engagement

Question 2

Recommendation algorithms

Answer 2

The systems that suggest content to you on platform like YouTube, Netflix, or social media. They shape what we see online. They help personalise content, but are often criticised for creating “filter bubbles” or “echo chambers”

Question 3

Filter bubbles/echo chambers

Answer 3

When people only see information that reinforces their existing beliefs. This can lead to polarisation or narrow interests

Question 4

Zhihu

Answer 4

China’s largest knowledge-sharing platform (similar to Quora). It allows users to ask questions, share answers, and follow topics. It’s a mix of social media and a Q&A site. At first, it used a content-based filtering algorithm. Later, it switched to a social filtering algorithm

Question 5

Content-based filtering algorithm

Answer 5

Recommends content based on the topics a user subscribed to. Work better for experienced users who already know what topics interest them

Question 6

Social filtering algorithm

Answer 6

Recommends content based on what a user’s friends or followers are engaging with. Help new or casual users discover diverse content by showing them what their friends or popular users are engaging with

Question 7

“Rich-get-richer” effect

Answer 7

Where already popular users gain even more followers and visibility, further amplifying their influence on the platform. This happened when Zhihu shifted to the social filtering algorithm

Question 8

Content-oriented behaviour

Answer 8

Seeking and answering questions

Question 9

Socially-oriented behaviour

Answer 9

Following users and engaging with their activities

Question 10

“Cold-start” problem

Answer 10

When it is harder for new users to gain visibility. Happened after implementation of social filtering algorithm

Question 11

Hybrid algorithms

Answer 11

Combine both approaches, adjusting recommendations based on each user’s activity and preferences. This strategy improves user engagement and can help platforms grow by encouraging social interaction and content creation, as seen with Zhihu’s successful expansion and monetisation