YouTube users find it difficult to focus on the productive task or are quick to forget the initial purpose of their visit to the platform. The user needs to practise continuous control to stay on relevant topics - this mental load is inefficient, draining, and demoralizing.
- We hypothesize that this phenomenon's root cause is the perceived bias in the YouTube recommendation algorithm towards increasing the user's watch time by recommending popular, clickbait, and interestingly irrelevant videos. This leads the user down the YouTube Rabbit Hole.
- We conduct a qualitative study on the relevance of videos that are recommended on YouTube to the original video that the user searched for, herein after called the root video.
- This study addresses how rapidly the topic of the video shifts with respect to the topic of the root video as we repeatedly follow the recommendation links, thereby simulating the behaviour of a user going down the rabbit hole.
- Analysing the conceptual relevance of YouTube’s recommended videos to prove the existence of this phenomenon (YouTube Rabbit hole).
- We developed a web crawler and scraper for modelling the video context and list the recommended links in an incognito environment.
