Examining algorithmic biases in YouTube’s recommendations of vaccine videos

Abul-Fottouh, Deena; Song, Melodie Yun-Ju; Gruzd, Anatoliy

doi:10.1016/j.ijmedinf.2020.104175

Cited by 42 publications

(42 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For example, YouTube’s search algorithm has historically recommended videos that attracted the most views or clicks. However, the recent heightened concerns regarding harmful misinformation on YouTube has prompted algorithm changes, which have reportedly reduced consumption of borderline content by 70% (24). Further potential action through increasing the ranking and visibility of health content from reputable scientific sources such as universities, hospitals and health charities would improve consumer exposure to high-quality and reliable health information.…”

Section: Discussionmentioning

confidence: 99%

(Mis) Communicating The Gut Microbiome: A Cross-Sectional Assessment of Social Media Video Content

Chidambaram

Maheswaran

Chan

et al. 2022

Preprint

View full text Add to dashboard Cite

Background Social media platforms, such as YouTube, are an integral means of disseminating information about health and wellness to the public. However, anecdotal reports have cited that the human gut microbiome has been a particular focus of dubious, misleading and, on occasion, harmful media content. Despite these claims, there have been no published studies investigating this phenomenon within popular social media platforms. Aim This study aims to (1) evaluate the accuracy and credibility of YouTube videos related to the human gut microbiome and (2) investigate the correlation between content engagement metrics and information quality, as defined by validated criteria. Methods In this cross-sectional study, videos about the human gut microbiome were searched on the UK version of YouTube on 20th September 2021. The 600 most viewed videos were extracted and screened for relevance. Information content and characteristics were extracted and independently rated using the DISCERN quality criteria by two researchers. Results Overall, 319 videos accounting for 62,354,628 views were included. 73.4% of the videos were produced in North America, and over 78.6% were uploaded between 2019-2021. 41.0% of videos were produced by non-profit organizations. 16% of the videos had an advertisement for a product or promoted a health-related intervention for financial purposes. Videos by non-medical educational creators had the highest total and preferred viewership. Daily viewership was highest for videos by internet media sources. The average DISCERN and HONcode scores were 49.5 (0.68) out of 80 and 5.05 (2.52) out of 8, respectively. DISCERN scores for videos by medical professionals (53.2 (0.17)) were significantly higher than videos by independent content creators (39.1 (5.58)), (p<0.01). Videos with promotional material scored significantly lower DISCERN scores than videos without any advertisements or product promotion (p<0.01). There was no correlation between DISCERN scores and total viewership, daily viewership or preferred viewership (number of likes). Conclusions The overall credibility of information about human gut microbiome on YouTube is poor. Moreover, there was no correlation between the video information quality and the level of public engagement. The significant disconnect between reliable sources of information and the public suggests that there is an immediate need for cross-sector initiatives in order safeguard vulnerable viewers from the potentially harmful effects of misinformation and disinformation.

show abstract

Section: Discussionmentioning

confidence: 99%

(Mis) Communicating The Gut Microbiome: A Cross-Sectional Assessment of Social Media Video Content

Chidambaram

Maheswaran

Chan

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Concerns have also been raised by health practitioners and commentators related to miscommunication of risk and subsequent spread of misinformation (Wise, 2021), which is contributing to the rise of anti-vax movements. Research suggests that once people start engaging with anti-vax content, they are likely to be directed towards more of the same, which propagates the spread of misinformation and anti-vax sentiment within such communities (Abul-Fottouh et al., 2020).…”

Section: Covid-19 Vaccine Rolloutmentioning

confidence: 99%

Communicating public health during COVID-19, implications for vaccine rollout

Bloomfield¹,

Magnusson²,

Walsh

et al. 2021

Big Data & Society

View full text Add to dashboard Cite

A large body of information and opinion related to COVID-19 is being shared via social media platforms. Recent reports have raised concerns about the reliability and verifiability of said information being disseminated and the way systems, processes and design of the platforms facilitates such spread. This, alongside other areas of concern, has resulted in several social media platforms taking steps towards tackling the spread of mis- and dis-information. Here we discuss approaches to online public health messaging from a range of sources during COVID-19, with a focus on official and non-official sources in the United Kingdom (UK). We highlight issues for ongoing public health decisions, and the potential impact for the future course of the pandemic.

show abstract

“…The second issue is connected to the recommendation algorithm itself. Due to its proprietary nature, the underlying incentive structures remain opaque [39]. Given the priming possibilities, there is a clear need for increased algorithmic transparency.…”

Section: Hyper-personalisation and Platform Designmentioning

confidence: 99%

Political Influence of Online Platforms: YouTube’s Place in European Politics

Kolotaev

Kollnig

2021

Vestnik SPbSU. International relations

View full text Add to dashboard Cite

Social media platforms are today one of the primary means of political expression and campaigning. The growing entanglement of politics with the online sphere raises interest in how these new types of media shape the social and political reality. While previous research tends to focus on Twitter and Facebook, limited works exists on the political influence of YouTube. This is even though YouTube is the second-most visited website and provides unique ways of engaging users and disseminating political messages through the combination of rich communication functionality with immersive audio-visual media content. Unfortunately, the influence of social media platforms such as YouTube on politics is difficult to analyse, due a lack of transparency and qualitative data. Independent researchers and the public have limited ways to access meaningful statistics about the video service. This article derives a taxonomy to classify YouTube’s ways of political influence from a case-study-driven analysis of YouTube in the European political landscape. Our taxonomy aligns with traditional theoretical concepts from media effects theory, particularly framing, priming and agenda-setting. We provide a brief discussion of the current regulatory landscape in the EU, and highlight gaps that might need improvement. We conclude by considering the question of whether YouTube is a political actor.

show abstract

Examining algorithmic biases in YouTube’s recommendations of vaccine videos

Cited by 42 publications

References 20 publications

(Mis) Communicating The Gut Microbiome: A Cross-Sectional Assessment of Social Media Video Content

(Mis) Communicating The Gut Microbiome: A Cross-Sectional Assessment of Social Media Video Content

Communicating public health during COVID-19, implications for vaccine rollout

Political Influence of Online Platforms: YouTube’s Place in European Politics

Contact Info

Product

Resources

About