top of page

More than just cat videos: YouTube as a platform for science engagement

Title: Interactions between emotional and cognitive engagement with science on YouTube

Author(s) and Year: Ilana Dubovi and Iris Tabak 2021

Journal: Public Understanding of Science (open access)

TL;DR: This study looks at public engagement with scientific content on YouTube, one of the most popular social media platforms worldwide. The researchers conducted quantitative analyses on a sample of top trending YouTube videos, focusing on different types of viewer engagement (behavioral, cognitive and emotional), and how they interact. The results show that science and education is a prominent video category, creating more viewer engagement than other popular categories. The authors also find that emotional stimulation leads to deeper engagement, suggesting that science communicators could focus on using emotional triggers to achieve greater engagement with their audiences, and social media is a promising platform on which to do so. Why I chose this paper: I am interested in how science is communicated in online videos, and I was keen to better understand how audiences engage with them.


Social media is becoming increasingly prevalent as a means through which the public consumes and engages with scientific information. However, compared to other media sources such as print, radio and television, the way the public interacts with science on social media is not well understood. Consequently, it is becoming a target of higher priority for science communication researchers.

This study from Dubovi and Tabak takes a step towards reducing this knowledge gap by focusing on public engagement with science on YouTube, one of the most popular social media platforms in the world with over 2 billion active users. Two research questions were asked:

1) Within the top trending videos on YouTube, what proportion is related to science content?

2) How are users engaging with this science content?

The nature of the second question is much more difficult to answer, but to help define the scope of the study, the researchers looked at three different types or ‘levels’ of engagement:

Behavioral engagement – This is the most basic level of engagement, and covers any kind of interaction with the content, including both passive users (e.g. someone who just watches a YouTube video) and active users (e.g. someone who posts a comment on the video). It can also include other actions such as liking and sharing the video, and reading comments of others.

Cognitive engagement – This type of engagement is characterized by users expending additional cognitive effort than what is required by behavioral engagement. For this study, this type of engagement is measured through the post-video comments section on YouTube, where users may post comments that elaborate on information provided in the video or argue against it. In this case users are going beyond the basic requirement of watching the video, and taking the extra step of posting a comment that requires a cognitive effort.

Emotional engagement – This type of engagement is defined by interactions with the content that have an emotional basis, either positive or negative (e.g. joy, anger, sadness). For this study, emotional engagement is inferred from the text provided in the post-video comments section, through automated sentiment analysis, which can identify the type and intensity of emotional expression.

The extent of science content on YouTube

The first research question above involved evaluating the magnitude of public engagement with science on YouTube. The study focused on ‘trending’ videos, which allowed for systematic collection of the video dataset and automatic organization into different categories using a YouTube data collection feature. Trending videos are also a useful dataset as they can have a high public impact, having the potential to become widely viewed in a short time period and reach the largest possible audience.

A total of 14,700 trending videos were collected from March to July 2019. ‘Science and Education’ was the 11th highest video category, accounting for 3.4% (N=503) of the top trending videos. However, the ‘Science and Education’ category performed even better in regards to engaging the public – being able to generate a relatively high number of average views (6th highest, above the ’Comedy’ and ‘Sports’ categories), as well as a high number of post-video comments (8th highest category).

The varying levels of engagement with science videos

Of the 503 videos classified as ‘Science and Education’, videos that were selected as trending videos multiple times were removed, leaving a sample of 89 videos. Quantitative analyses were used to evaluate these videos against the different levels of engagement highlighted above.

Behavioral engagement

This was assessed by counting the number of views, likes, dislikes and number of comments for each video. This data was collected at two different time points a year apart (May 2019 and May 2020). The authors tested for correlations between these types of interactions, finding moderate to strong correlations between them (e.g. the higher the number of comments, the higher the number of views, likes and dislikes).

Cognitive engagement

A random sample of 1000 comments from each video (total 89,000 comments) were used. To evaluate cognitive engagement within this sample of comments, the authors developed an “automated coding scheme of linguistic argumentation” (i.e., used argumentative language found within the comments as a proxy for cognitive engagement). This detected that at least 16% of comments included argumentative language.

Emotional engagement

The same sample of 89,000 comments was also used to evaluate emotional engagement. All comments were assessed via an automated sentiment analysis, which tested for markers of emotional language (e.g. joy, anger, sadness). The assessment found significantly fewer negative expressions than positive and neutral expressions per video, and that trust was the most commonly expressed emotion.

Interrelationships between the different engagement types

Analysis from the authors showed strong positive correlations between all engagement types: behavioral, cognitive and emotional. For example, more ‘behavioral engagement’ (more likes, comments etc.) lead to more ‘cognitive engagement’ (more argumentative comments), which makes intuitive sense.

More interestingly, detailed analysis of emotional engagement showed that expression of all emotions, whether positive (e.g. joy and surprise) or negative (e.g. anger and fear), is strongly correlated with argumentative conversation (and hence cognitive engagement). Conversely, statements that were neutral from an emotional standpoint were negatively correlated with argumentation (and hence negatively correlated with cognitive engagement).

Implications for Science Communication

The findings from the first part of this study show the prominence of science content within trending YouTube videos, outperforming other popular categories such as ‘sports’ and ‘news and politics’ in regard to average views and average number of comments. This illuminates the importance of YouTube – and social media more broadly – as a platform which has considerable influence on the way the public consume scientific content, and thus is worthy of further attention from researchers in future science communication studies.

The findings also highlight the significance of emotional discourse in public engagement with science on social media, and show that the post-video comments section on YouTube is an important space for this emotional expression. A key finding from the study was that emotional stimulation, irrespective of whether it is positive or negative, leads to more behavioral and cognitive engagement and thus triggers a deeper intellectual involvement from the public. This suggests that when creating online video content, in addition to other important considerations such as accuracy and clarity, science communicators should also focus on generating an emotional response from viewers. They can also be flexible when choosing what emotions to target, as both positive and negative emotions correlate to greater engagement.

Social media creates a platform for the public to participate in emotional discourse, and thus represents a promising means through which to promote and advance public engagement with science.

Edited by Joshua Buchi

Cover image credit: StockSnap from Pixabay



bottom of page