Content uploaded by Araceli Castelló-Martínez
Author content
All content in this area was uploaded by Araceli Castelló-Martínez on Sep 21, 2019
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
$
€
£¥
social sciences
Article
Commenting on Top Spanish YouTubers:
“No Comment”
Victoria Tur-Viñes * and Araceli Castelló-Martínez
Communication and Social Psychology department, University of Alicante, 03690 Alicante, Spain;
araceli.castello@ua.es
*Correspondence: victoria.tur@ua.es; Tel.: +34-636-000-250
Received: 22 August 2019; Accepted: 16 September 2019; Published: 20 September 2019
Abstract:
The aim of this paper was to analyze commenting activity and sentiment (polarity
and subjectivity) in interactions in response to videos by Spain’s most-subscribed YouTubers.
An exploratory study was conducted on the content of the comments, their relationship with other
social media actions, subjectivity, and polarity, as well as from the perspective of the participatory
culture. The results show that commenting is a potential option for interaction that is underused by
the communities of users. Replies to comments are found to be limited to the user–user level, while
YouTubers themselves and the moderators that YouTube allows them to designate rarely comment or
reply on social networks. However, creators do monitor comments and provide feedback to a limited
selection thereof in subsequent videos. There thus appears to be a strategic, exploitative use of
comments, marked by a delayed response aimed at attracting audiences to new content.
Keywords: YouTubers; sentiment analysis; interaction; influencers; commenting; Facebook
1. Introduction
1.1. YouTube: Broadcast Yourself
YouTube has been the biggest video viewing platform since 2005. It has been described by scholars
as “post-modern television” (Kavoori 2015;Feixas et al. 2014;Lavado 2013;Murolo 2010) and is the
preferred platform for audiovisual consumption among teens (Garc
í
a-Jim
é
nez et al. 2016). More than
1.9 billion users sign into YouTube (hereinafter, YT) each month, 300 h of content are uploaded to the
platform every minute, and more than 30 million users visit the website every day, with an average
visit time of 8 min 51 s.
In Spain, YT is the third most widely used social network (69% of users), after Facebook
and WhatsApp, and the second highest rated (8.1 out of 10), behind WhatsApp (IAB Spain 2018).
The platform possesses a remarkable capacity for generating a strong sense of community among
users (Boyd 2014, p. 47; Chau 2010, p. 65) who share interests and exhibit a high level of loyalty to
YouTubers. Young people constitute the demographically predominant group on YT, both in terms of
absolute audience and of the volume of feedback actions and interactions (Chau 2010, p. 65).
The mechanisms for interaction are what differentiate YT from television, as they offer additional
spaces for enhancing the YouTuber–community relationship and a source of useful information for
YouTubers to gauge the reactions of their followers and to learn about what they do. Specifically,
commenting offers a productive forum for interaction where followers can express themselves verbally.
This research will focus on comments in order to examine what motivates their content, what sentiment
they reflect (polarity and subjectivity), and how YouTubers make use of this interaction.
Soc. Sci. 2019,8, 266; doi:10.3390/socsci8100266 www.mdpi.com/journal/socsci
Soc. Sci. 2019,8, 266 2 of 15
1.2. YouTubers
A YouTuber is a person who has a channel on the YT social network and uses it to publish videos,
with the aim of generating as many views as possible (Lange 2007;Hidalgo-Mar
í
and Segarra-Saavedra
2017, p. 45) and securing potential revenues through the monetization of their audience (Rull 2014,
p. 1). Some audiovisual creators have become icons in the youth entertainment world, which represents
an alternative to the traditional audiovisual industry (Ramos-Serrano and Herrero-Diz 2016).
A YouTuber may be an influencer, but not all influencers are YouTubers. As a result of the videos
they post on YT, YouTubers become media figures who build their identities through the content
they broadcast (Scolari and Fraticelli 2016, p. 1672). According to Scolari and Fraticelli, another
distinctive feature of YouTubers is the individualization of the viewer. The resources offered, like
the visitor counter, the number of subscribers or the likes and dislikes, and the spaces provided for
users to share their comments, demonstrate this: “This possibility of feedback [
. . .
] is enhanced
and expanded through the interconnection of the YouTuber’s accounts on hypermedia platforms like
Twitter, Facebook, and Instagram, where they receive messages that they often respond to in their
videos” (Scolari and Fraticelli 2016, p. 1680). Based on these considerations, we posited the following
research questions:
Q1: Is the volume of comments generated by a video on YT the same as that generated on the
YouTuber’s Facebook profile in response to the same content?
Q2: How do YouTubers manage the social conversation?
According to the Social Media Marketing Glossary of Argentina’s Direct and Interactive Marketing
Association (AMDIA), an influencer is a person who makes others do or think what he or she wants
them to, thereby changing the behavior of groups or societies (AMDIA 2015).
The leadership role played by YouTubers activates some significant mechanisms for influencing
millions of followers. In addition to the payments offered by YouTube based on visitor numbers and
Google AdSense advertising, YouTubers can also obtain profits through agreements with different
brands (S
á
ez Barneto and Camacho 2017, p. 51). As media opinion leaders, YouTubers establish
commercial relationships with advertisers for the promotion of their products, services, and/or brands,
thereby cultivating an extraordinary power of influence and suggestion over their audience (Del
Pino-Romero and Castell
ó
-Mart
í
nez 2017;Ramos-Serrano and Herrero-Diz 2016). The importance of
YouTubers lies in their power to create and maintain massive audiences of young followers and to
trigger interaction in order to increase the chances of the natural expansion of the video.
In social psychology, this phenomenon is explained by Cialdini’s theory of influence (Cialdini
2001) and its six principles: Commitment and consistency; reciprocity; social proof; authority; liking;
and scarcity. Reciprocity is highlighted by Cialdini as one of the most powerful elements for eliciting
acquiescence from others. Evidence of this property of influence can be found in the comments posted
on YouTube.
The videos broadcast by YouTubers are characterized by a marked aesthetic sense tending towards
professionalism (Sabich and Steinberg 2017, p. 184), with certain rules and tactics that organize the
discourse and lend consistency to their essentially viral nature (Rotman and Preece 2010, p. 323).
Common patterns include the strategy of set introductory and closing phrases in the video (uniquely
identifying each YouTuber) and the use of a personal design intended to promote brand recognition
for the channel (Tur-Viñes et al. 2018, p. 1226).
YouTubers offer young people a new form of monologue-based communication to engage and
attract viewers (Frobenius 2014). Rego and Romero-Rodr
í
guez (2016, p. 219) analyzed the language of
the three YouTubers with the most subscribers in Spain (El RubiusOMG, TheWillyRex and Vegetta777)
and concluded that they all use a colloquial language mainly targeted at millennials. Research by
Gallardo-Camacho and Alonso (2010) shows that internet users who consume videos online, and
specifically on YouTube, adopt a passive attitude, apparently inheriting the behavior of spectators of
traditional one-way media.
Soc. Sci. 2019,8, 266 3 of 15
Dynel (2014) identifies three different levels of communication on YouTuber channels: The level
of the speaker and hearer in the video interaction; the level of the sender and recipient of a YT video;
and the level of speakers and hearers who post and read comments, respectively. Participating in all
these levels is not only the YouTuber but also the members of the YouTuber’s production team and the
hearers themselves, who are able to comment, reply, or post their own videos. Collaboration is another
of the qualities that define the space of YouTuber channels as a collective phenomenon characteristic of
the participatory culture in which we are immersed today.
1.3. Commenting on YouTuber Videos
The analysis of user interaction with the content broadcast on social networks is known as
natural language processing (NLP) or opinion mining (OM). The possibility of posting comments on
articles, posts, videos, or other content broadcast on social networks is one of the distinctive features of
collaborative websites. The desire to express an emotion or an opinion and to supplement or clarify
information constitute the main motivations behind commenting on social network content (Stroud et
al. 2016).
There are various studies that explore the influence of user comments on the perception of the
content broadcast on social networks. In the area of digital journalism, Von Von Sikorski and Hänelt
(2016) point out that a consensus among user comments affects the perception of journalistic quality,
reliability, and persuasion of the content broadcast. People believe that the comments of others on
online news stories are a representative reflection of what the general public think and this directly
affects their own evaluations of the stories (Kim 2015). This idea was also confirmed by Lee and Jang
(2010), who demonstrated that user opinion about certain information broadcast on online channels
was influenced by comments previously posted by other users.
Documenting a hybrid interaction between journalists and readers, Maniou and Bantimaroudis
(2018) proposed a theory of hybrid salience. Their research suggests that word-of-mouth salience,
as a horizontal influence, exerts a great deal of influence on public salience, demonstrating that people
form opinions by looking at newspaper articles as well as readers’ comments. In some cases, the
latter seem to be more important than the former (p. 15). This perspective can be extrapolated to the
narratives developed by YouTubers on their channels and can facilitate an understanding of how their
public salience is constructed.
There is a huge potential for the public discourse associated with this form of computer-mediated
communication with users, according to Weber (2013). However, this potential is present only
when several users participate in the comments and when their communication becomes interactive.
Weber adapts the news theory of Galtung and Galtung and Ruge (1973) and assumes that the factors
shaping the news in an article affect both the participation and interactivity levels in the comments
section. Therefore, the type of content and how it is narrated will affect the participation of commenters
and their interaction with one another. Lee (2012) posits the concept of a hostile media perception
(HMP) arising from a type of defensive cognitive processing, suggesting that people with high ego
involvement perceive the news as hostile and biased if they read negative comments on it. These studies
confirm that people may erroneously attribute the opinions expressed by others in the comments to the
news article itself. All of this demonstrates that, like the content that generates them, user comments
also have the power to influence and propel the conversation.
Based on the above, we posited the following research questions:
Q3: What percentage of comments generate replies from other users and what is the average
number of replies per comment generated on each video/channel?
Q4: What topics predominate in the comments on each video?
Madden et al. (2013) stress the heterogeneity that characterizes user comments on the content
published on digital platforms. On the question of what motivates users to comment or reply, Chang
et al. (2018) analyzed commenting on Facebook and suggested that relational closeness is the first and
most significant determinant of likelihood to respond. When relational closeness was high, replies
Soc. Sci. 2019,8, 266 4 of 15
were direct and immediate. In the absence of relational closeness between the comment poster and
respondent, the likelihood of responding depends on (1) the perceived acuity and seriousness of the
content, (2) consistency in posting patterns, (3) perceived capacity to provide efficacious support,
(4) history of reciprocity, (5) perceived resonance with posted content, (6) perceived motivations of
the poster of the original comment, and (7) perceptions of other users. Users tend to read comments
posted by others in their interaction with videos on YT with two main motives: Information seeking
and entertainment (Khan 2017).
Focusing on content broadcast in video format, Ksiazek et al. (2014) demonstrated a positive
relationship in news videos between popularity (defined in terms of the number of views and
recommendations) and user–content interaction (comments without replies from others). However,
videos with fewer views generated more user–user interaction (comments with replies by other users).
Siersdorfer et al. (2010) studied comments on YT (specifically, the likes that comments received),
and concluded that positive comments are associated with high levels of popularity defined in terms
of the number of views. Jamali and Rangwala (2009) also provide evidence of a relationship between
interactivity and number of views: The age of the comment and the number of words it contained
were associated with high viewing levels. Lee et al. (2010) proposed a predictive model of views in
which the number of comments in a conversation thread and the lifetime of the comments thread can
predict a high number of views.
The above led us to posit the following research questions:
Q5: What relationship exists between the comments received on videos and other interaction
variables (views, likes, and dislikes)?
Q6: What relationship exists between a channel’s number of subscribers and the polarity and
subjectivity of comments?
Q7: What relationship exists between interactions (views, comments, likes, dislikes) that videos
receive and the polarity and subjectivity of the comments?
In addition to being able to reply to user comments with another comment, the YT platform offers
content creators two interaction options: Giving a red heart to a favorite comment, and “pinning”
a user comment to the top of a thread (“pinned by creator” appears beside the profile of the user
who made the comment). However, the YouTube API could not provide data on these at the time of
this study.
1.4. Sentiment Analysis
Opinion mining involves what is known as sentiment analysis, which refers to the different
methods of computational linguistics that help identify and extract subjective information from content
in the digital world. Sentiment analysis makes it possible to extract a tangible and direct value, such as
determining whether a text published online contains positive or negative connotations.
Sentiment analysis of conversations generally includes two values: Subjectivity and polarity.
Subjectivity relates to whether the comment is objective or subjective. Polarity refers to whether the
comment is positive, negative, or neutral (Pang and Lee 2008). This methodology is therefore focused
on automatically determining whether or not an opinion is included in a text, on identifying whether
the polarity or sentiment expressed is positive, negative, or neutral, and on extracting an author’s
perception of specific aspects of a topic (Vilares et al. 2017, p. 126).
A diverse range of studies have engaged in sentiment analysis of social networks like Twitter
or YouTube (Cheong and Cheong 2011;Siersdorfer et al. 2010;Sureka et al. 2010). Krishna (2014)
demonstrated that trends in user sentiments are directly related to real world events, on the basis of
certain key words.
Some authors (Choi 2003;Tannen 1999) suggest that the anonymity offered by the internet tends
to favor antagonism and conflict in interactions. Lange (2007, p. 11) studied hostile behavior on YT and
confirmed that the presence of a personal image on a profile does not guarantee courteous interaction.
Soc. Sci. 2019,8, 266 5 of 15
Moreover, the motivations of users who post hostile comments are complex and varied, making their
control or regulation rather complicated (Lange 2007, p. 27).
Malicious practices in interaction have been confirmed by Benevenuto et al. (2010) in a study
identifying the six most recurrent actions of YT users (views, list of a user, top videos or related videos,
interactions, search, and others). The study found that some users signed into YT and rated videos
without watching them first. This is evidence that data on interactions can be falsified.
The emotional charge is a determining factor for content expansion. Positive messages get
disseminated more often than negative ones, but emotional intensity in both cases increases the
likelihood of content going viral or provoking changes of attitude, as has been shown in the case of
advertising by (Kirby 2004;Phelps et al. 2004;Eckler and Bolls 2011;Hagerstrom et al. 2014). However,
Thelwall et al. (2012) studied YT comments and found that audiences respond on a mass scale to
negative comments while positive comments elicit few responses.
This review of the literature on the subject led us to posit the following research questions:
Q8: What is the tone/sentiment of the social conversation in comments generated by the most
viewed content of the top YouTubers?
Q9: What are the characteristics of the videos with the highest levels of polarity and subjectivity
(duration and type of video)?
Q10: What is the time of publishing of the videos with the highest polarity and subjectivity levels?
Q11: What relationship exists between the videos with the highest polarity and subjectivity levels
and the interaction generated on other platforms (Facebook and Twitter)?
2. Method
The main objective of this paper was to analyze commenting activity and sentiment (polarity and
subjectivity) in interactions in response to videos by Spain’s most-subscribed YouTubers. Commenting
activity was considered both on YouTube and on the YouTuber’s official FB page, in relation to the
same video, along with the YouTuber’s participation in the resulting social conversation.
To this end, an exploratory study was conducted, involving a quali-quantitative analysis of the
content of a convenience sample of 8598 comments on YT generated by 10 videos. The samples selected,
covering the period from September 2018 to February 2019, are detailed below:
-
Sample of channels: 10 channels were chosen from a ranking of the 250 accounts with the most
subscribers according to SocialBlade (September 2018). The selection criteria for the channels
were: Spanish YouTuber channels with the most subscribers, together with the presence of
monetization and parallel profiles on other social networks (Facebook and Twitter).
- Sample of videos and comments: The selection was based on two levels:
Level 1 (comment content): From each channel, only the most recent video in the period studied
and with the most views was chosen, resulting in a sample of 10 videos that allowed for the
collection of 8598 comments. The criterion of the most recent video was chosen due to the nature
of the software used to extract details from the comments (NVivo Capture), which allows access
to the last 1000 comments on the video at the time of capture. By choosing the most recent videos,
we could maximize the capture of comments at the beginning of the conversation thread, although
in some cases the volume of comments was very high and it was not possible to capture the first
comments. On three channels the comments did not reach the maximum number of 1000 that
could be captured by NVivo 12. This level was used to answer research questions Q1–Q5. Level 2
(comment polarity and subjectivity): 100 videos were chosen, made up of the 10 videos with the
most views in the study period on each of the 10 previously identified channels. These 100 videos
generated a sample of 1,141,091 comments. This level was used for research questions Q6–Q11.
The variables analyzed in each of the 10 videos of the sample were: YouTuber, title of video,
date and time published, subscribers to channel, views, duration of video, and direct appeals to the
audience. The variables considered in the comment analysis were: Number of comments, replies to
Soc. Sci. 2019,8, 266 6 of 15
comments, motivation of content, users who post replies to comments, polarity, subjectivity, and the
number of posts generated on FB.
The sentiment analysis of the conversation (polarity and subjectivity) was conducted using the
analytical tool TextBlob, a paid software program for analyzing and measuring content that provides
information on the polarity of comments posted by users on the videos. This tool assigns a value to
each word in a sentence in order to calculate the subjectivity and polarity of the comments:
-
Subjectivity of conversation: objective or subjective (+0.0 => +1.0). The value of +1.0 is the
highest level of subjectivity and 0 is the highest level of objectivity.
-
Polarity of conversation sentiment: Negative or positive (
−
1.0 => +1.0). The value of 0
denotes neutrality.
3. Results
3.1. General Interaction Metrics for Channels and Videos Selected (Level 1)
Table 1shows the selection of Spanish YouTuber channels occupying the top positions in
SocialBlade’s ranking in September 2018, based on subscribers and views. These 10 channels have
a collective total of 133,503,699 subscribers, with an average of 13,350,370 subscribers each. There are
three channels that exceed both the average and the median: elrubiusOMG, VEGETTA777, and
TheWillyRex. In total, 80% of the channels belong to the video game category on YT.
Table 1. Ranking of channels and videos in the sample.
Channel
Name Topic Subscribers (M)
(21/02/2019) Video Title Video Type Date and
Time Duration
1
elrubiusOMG
VIDEO
GAMES 33 EL NUEVO GENIO DE ALADDIN Vlog 19/02/2019
14:03 0:10:39
2
VEGETTA777
VIDEO
GAMES 25 FORTNITE - MINIJUEGO *PINBALL
LOCO* (MODO CREATIVO)
Screen-sharing
25/02/2019
11:33 00:14:07
3
TheWillyrex
VIDEO
GAMES 15
AL LIMITE! |PAINT THE TOWN RED Screen-sharing
16/01/2019
16:17 00:09:15
4
ExpCaseros
HOME
MADE
EXPERIMENTS
10
EL INVENTO MÃS ESTÊPIDO Y
ASQUEROSO DE AMAZON -
REVIENTA GRANOS
Sit-down 24/01/2019
12:58 00:13:18
5
Makiman131
VIDEO
GAMES 10
ENTRENANDO COMO UN
MILITAR!! PRACTICA MILITAR
MAKIMAN
Vlog 19/02/2019
12:01 00:11:17
6
luzugames
VIDEO
GAMES 8.6 FINAL INCREIBLE! RESIDENT EVIL
2 REMAKE - LUZU
Screen-sharing
11/02/2019
11:15 00:58:12
7 TheGrefg VIDEO
GAMES 9.6
MI GRAN VICTORIA EN BLACK
OPS 4 *NUEVO CONTENIDO
GRATIS* - THEGREFG
Screen-sharing
24/02/2019
17:43 01:48:38
8
sTaXxCraft
VIDEO
GAMES 7.2
FORTNITE TE DA ESTE CAMUFLAJE
GRATIS!!
Screen-sharing
21/11/2018
20:22 00:10:45
9
gymvirtual
VIRTUAL
GYM 6
CALENDARIO DE EJERCICIOS
PARA ADELGAZAR DICIEMBRE |
GYMVIRTUAL
Sit-down 30/11/2018
10:00 00:05:43
10 elchurches VIDEO
GAMES 5.6
EL NUEVO LADRON
PROFESIONAL! SIMULADOR DE
LADRON - ELCHURCHES
Screen-sharing
06/11/2018
11:00 00:13:07
Source: compiled by authors based on data from SocialBlade and YT.
Table 2presents the data on the interaction with the 10 videos in the sample on level 1 (one for
each YouTuber selected), as well as the polarity and subjectivity values, which will be discussed below.
Soc. Sci. 2019,8, 266 7 of 15
Table 2. Interaction, polarity-subjectivity, and ratios for videos in the sample.
Video YouTuber Views
(28/02/2019)
Comments
(28/02/2019)
Likes
(28/02/2019)
Dis- Likes
(28/02/2019) Polar-Ity Subjec-
Tivity
Comment-
View
Ratio
Like-
View
Ratio
Dislike-
View
Ratio
Comment-
Like Ratio
1.9 elrubiusOMG 6,922,305 46,305 922,619 17,323 3.6 22.62 0.7% 13.3% 0.3% 5.0%
2.1 VEGETTA777 387,239 1595 42,213 786 N/D N/D 0.4% 10.9% 0.2% 3.8%
3.4 theWillyrex 206,277 2117 19,694 1247 2.79 21.15 1.0% 9.5% 0.6% 10.7%
4.5 ExpCaseros 834,523 2557 23,790 1273 1.43 13.47 0.3% 2.9% 0.2% 10.7%
5.3 Makiman131 556,348 2535 28,107 2193 −4.83 24.53 0.5% 5.1% 0.4% 9.0%
6.1 luzugames 167,649 1170 16,853 132 12.15 28.96 0.7% 10.1% 0.1% 6.9%
7.10 TheGrefg 412,418 350 20,675 1120 1.43 18.49 0.1% 5.0% 0.3% 1.7%
8.6 sTaXxCraft 158,551 519 10,612 132 16.08 22.04 0.3% 6.7% 0.1% 4.9%
9.1 gymvirtual 120,144 480 5641 82 2.52 14.06 0.4% 4.7% 0.1% 8.5%
10.4 elChurches 251,897 1023 21,259 295 −0.92 6.47 0.4% 8.4% 0.1% 4.8%
TOTAL 10,017,351 58,651 1,111,463 24,583 0.6% 11.1% 0.2% 5.3%
Source: compiled by authors based on data from TextBlob.
Although the elrubiusOMG video holds first place in all four interaction variables, the relationship
between views and social actions is not repeated for the rest of the channels, as can be seen in the
ranking in Figure 1. The highest ratio between comments and views in this sample belongs to the
TheWillyrex video, with 1%, while the video by TheGrefg has a ratio below 0.1% for this value.
The video with fewest views in this sample is by gymvirtual, with 120,144 views, and three videos have
fewer than 1000 comments: The videos by TheGrefg, sTaXxCraft, and gymvirtual. The videos that
receive the highest number of comments in relation to the “likes” obtained are the ones by TheWillyrex
and ExpCaseros, both with 10.7%.
Soc. Sci. 2019, 8, x FOR PEER REVIEW 7 of 15
Table 2. Interaction, polarity-subjectivity, and ratios for videos in the sample.
Video YouTuber Views
(28/02/2019)
Comments
(28/02/2019)
Likes
(28/02/2019)
Dis-Likes
(28/02/2019)
Polar-
Ity
Subjec-
Tivity
Comment-
View Ratio
Like-
View
Ratio
Dislike-
View
Ratio
Comment-
Like Ratio
1.9 elrubiusOMG 6,922,305 46,305 922,619 17,323 3.6 22.62 0.7% 13.3% 0.3% 5.0%
2.1 VEGETTA777 387,239 1595 42,213 786 N/D N/D 0.4% 10.9% 0.2% 3.8%
3.4 theWillyrex 206,277 2117 19,694 1247 2.79 21.15 1.0% 9.5% 0.6% 10.7%
4.5 ExpCaseros 834,523 2557 23,790 1273 1.43 13.47 0.3% 2.9% 0.2% 10.7%
5.3 Makiman131 556,348 2535 28,107 2193 −4.83 24.53 0.5% 5.1% 0.4% 9.0%
6.1 luzugames 167,649 1170 16,853 132 12.15 28.96 0.7% 10.1% 0.1% 6.9%
7.10 TheGrefg 412,418 350 20,675 1120 1.43 18.49 0.1% 5.0% 0.3% 1.7%
8.6 sTaXxCraft 158,551 519 10,612 132 16.08 22.04 0.3% 6.7% 0.1% 4.9%
9.1 gymvirtual 120,144 480 5641 82 2.52 14.06 0.4% 4.7% 0.1% 8.5%
10.4 elChurches 251,897 1023 21,259 295 −0.92 6.47 0.4% 8.4% 0.1% 4.8%
TOTAL 10,017,351 58,651 1,111,463 24,583 0.6% 11.1% 0.2% 5.3%
Source: compiled by authors based on data from TextBlob.
Although the elrubiusOMG video holds first place in all four interaction variables, the
relationship between views and social actions is not repeated for the rest of the channels, as can be
seen in the ranking in Figure 1. The highest ratio between comments and views in this sample belongs
to the TheWillyrex video, with 1%, while the video by TheGrefg has a ratio below 0.1% for this value.
The video with fewest views in this sample is by gymvirtual, with 120,144 views, and three videos
have fewer than 1000 comments: The videos by TheGrefg, sTaXxCraft, and gymvirtual. The videos
that receive the highest number of comments in relation to the “likes” obtained are the ones by
TheWillyrex and ExpCaseros, both with 10.7%.
Figure 1. Views and comments for videos in the sample. Source: compiled by authors based on data
from TextBlob.
3.2. Analysis of Comments on the 10 Videos in the Sample (Level 1)
For all of the videos, the number of comments captured by NVivo is more than 40% of the total
number of comments, with the exception of the video by elrubiusOMG, for which the number of
comments analyzed represents only 2.4% of the very high number of comments it received; this
reduces the overall average of comments captured per video to 14.7%. In five cases, this value was
above 87%.
In response to Q1, the biggest volume of comments is generated on YT. Activity on FB is much
lower, and in some cases, there are no comments at all. There is no significant relationship between
comments on YT and FB about the same video.
In relation to Q2, the YouTuber’s official profile (ID) was tracked in the comment lists extracted
with NVivo for each video in the sample. The results show that the YouTubers never reply, and thus
in the videos studied, the interaction is strictly between followers. In none of the 10 videos of the
sample is there a comment or reply posted by the YouTuber.
0
10000
20000
30000
40000
50000
0.00
2,000,000.00
4,000,000.00
6,000,000.00
8,000,000.00
Views Comments
Figure 1.
Views and comments for videos in the sample. Source: compiled by authors based on data
from TextBlob.
3.2. Analysis of Comments on the 10 Videos in the Sample (Level 1)
For all of the videos, the number of comments captured by NVivo is more than 40% of the total
number of comments, with the exception of the video by elrubiusOMG, for which the number of
comments analyzed represents only 2.4% of the very high number of comments it received; this reduces
the overall average of comments captured per video to 14.7%. In five cases, this value was above 87%.
In response to Q1, the biggest volume of comments is generated on YT. Activity on FB is much
lower, and in some cases, there are no comments at all. There is no significant relationship between
comments on YT and FB about the same video.
In relation to Q2, the YouTuber’s official profile (
ID
) was tracked in the comment lists extracted
with NVivo for each video in the sample. The results show that the YouTubers never reply, and thus in
the videos studied, the interaction is strictly between followers. In none of the 10 videos of the sample
is there a comment or reply posted by the YouTuber.
Soc. Sci. 2019,8, 266 8 of 15
The ratio of replies to comments was calculated in the following way:
response rate ratio
commentper video =
Preplies to commentsvideo
Total comments video ×100.
The percentage of replies to comments on YT on each channel for the video selected (Q3) shows
that the highest response rate of users to comments made by others is 31% (gymvirtual). The topic of
the channel is not a determining factor in the response rate as channels with different subjects (virtual
gym and video games) obtain the highest ratios. The average comment to reply ratio between users
is 9.9%. Only in the cases of TheGrefg, sTaXxCraft, and gymvirutal is the percentage of replies to
comments above 20%.
Table 3also breaks down the comments captured by NVivo between comments posted by users
on the video and replies to those comments, together with the data related to the different users who
post comments and replies.
Table 3. Comments and replies captured by NVivo.
Video YouTuber
NVivo
Comments
(28/02/2019)
Comments Replies
n%
Number of
Different
Users
Different
Users—
Comments
Comments
per User n%
Number of
Different
Users
Different
Users—
Replies
REPLIES
PER USER
1 elrubiusOMG 1107 982
88.7%
943 96.1% 1.0 125
11.3%
80 63.8% 1.6
2 VEGETTA777 1071
1000 93.4%
957 95.7% 1.0 71 6.7% 55 77.2% 1.3
3 theWillyrex 1025
1001 97.7%
961 96.0% 1.0 24 2.3% 21 87.9% 1.1
4 ExpCaseros 1048
1001 95.6%
948 94.7% 1.1 47 4.4% 33 70.9% 1.4
5 Makiman131 1048
1003 95.7%
955 95.2% 1.1 45 4.3% 31 69.0% 1.5
6 luzugames 1034 994
96.1%
964 97.0% 1.0 40 3.9% 22 55.0% 1.8
7 TheGrefg 307 217
70.5%
198 91.4% 1.1 90
29.5%
65 71.8% 1.4
8 sTaXxCraft 464 360
77.6%
328 91.1% 1.1 104
22.4%
77 74.2% 1.3
9 gymvirtual 480 333
69.4%
435 130.6% 0.8 147
30.6%
19 12.9% 7.7
10 elChurches 1014 858
84.6%
951 110.8% 0.9 156
15.4%
11 7.1% 14.2
Source: compiled by authors based on data from NVivo.
The users who decide to comment do not usually post more than one comment, and thus there is
very little difference between the number of comments and the number of different users who post
them. In the replies to comments, it is more common for users to interact more than once.
Q4 relates to the predominant topics in the comments. The most repeated words are the name of
channel, like, ha ha, genius, crack, video, code, cool, YouTuber, hi, free fire, and game. On the gaming
channels the following words also appear very frequently: Upload more free fire, episode, series, I love
it. It is evident that the comments are reactions to elements present in the video that provoke a need
for followers to respond, e.g., a video recorded with a defect, multi-player video games, difficulties
sharing a game, congratulations, and curses when things go badly while playing a video game.
3.3. Sentiment Analysis: Polarity and Subjectivity (Level 2)
The number of subscribers does not determine the polarity and/or subjectivity of the comments
(Q6), as shown in Table 4.
To explore possible correlations between the variables of interaction (views, comments, likes, and
dislikes) on the one hand, and polarity and subjectivity on the other (Q7), the Pearson correlation
coefficient was calculated. We found that there was no statistically significant relationship in this
respect, despite obtaining higher correlation coefficients in the subjectivity of the comments.
Neither polarity nor subjectivity follow a regular pattern in their relationship with the comments
received about the videos. As it was the most outstanding case, the relationship between comments,
polarity, and subjectivity for the channel elrubiusOMG is detailed in Figure 2:
Soc. Sci. 2019,8, 266 9 of 15
Table 4. Average polarity and subjectivity of comments for each channel.
YT Channels and No. of
Subscribers in Millions
Average Polarity of
Comments
Average Subjectivity of
Comments
elrubiusOMG (3.3) 6.6280 22.3090
VEGETTA777 (25) 2.6313 8.8888
TheWillyrex (15) 7.8190 23.1890
ExpCaseros (10) 4.2230 14.3180
Makiman131 (10) 4.5490 16.0040
Luzugames (8.6) 8.8640 26.3890
TheGrefg (9.6) 5.0500 15.2300
sTaXxCraft (7.2) 2.9990 5.1100
Gymvirtual (6) 8.2544 15.0811
ElChurches (5.6) 1.1070 7.1370
OVERALL AVERAGE: 5.28 15.50
Source: compiled by authors based on data from TextBlob.
Soc. Sci. 2019, 8, x FOR PEER REVIEW 9 of 15
TheGrefg (9.6) 5.0500 15.2300
sTaXxCraft (7,2) 2.9990 5.1100
Gymvirtual (6) 8.2544 15.0811
ElChurches (5.6) 1.1070 7.1370
OVERALL AVERAGE: 5.28 15.50
Source: compiled by authors based on data from TextBlob.
To explore possible correlations between the variables of interaction (views, comments, likes,
and dislikes) on the one hand, and polarity and subjectivity on the other (Q7), the Pearson correlation
coefficient was calculated. We found that there was no statistically significant relationship in this
respect, despite obtaining higher correlation coefficients in the subjectivity of the comments.
Neither polarity nor subjectivity follow a regular pattern in their relationship with the comments
received about the videos. As it was the most outstanding case, the relationship between comments,
polarity, and subjectivity for the channel elrubiusOMG is detailed in Figure 2:
Figure 2. Relationship between comments, polarity, and subjectivity in elrubiusOMG videos. Source:
compiled by authors based on TextBlob data.
10.42
8.24
4.99
13.59
5.75 5.22 4.89 3.60 4.75 4.83
23.92
23.23
28.77
28.28
17.32
21.49
17.41
22.62 19.58 20.47
0
5
10
15
20
25
30
35
40
45
0.00
20,000.00
40,000.00
60,000.00
80,000.00
100,000.00
120,000.00
140,000.00
160,000.00
180,000.00
1.10 1.3 1.6 1.5 1.8 1.1 1.4 1.9 1.7 1.2
Comments Average polarity Average subjectivity
Figure 2.
Relationship between comments, polarity, and subjectivity in elrubiusOMG videos. Source:
compiled by authors based on TextBlob data.
Soc. Sci. 2019,8, 266 10 of 15
The top positions in terms of polarity and subjectivity were taken by six videos, posted by the
YouTubers elrubiusOMG (one video), TheWillyrex (one video), Makiman131 (one video), luzugames
(the only channel with two videos in the polarity and subjectivity rankings), and sTaXxCraft (one video).
Although no significant correlation was detected between comments and polarity/subjectivity,
in a few cases, the videos whose comments had higher polarity and subjectivity levels are also the ones
with the most comments and “likes”.
sTaXxCraft and luzugames have the highest polarity levels, with 16.08 and 12.15, respectively.
The highest subjectivity level was found in the video by luzugames (28.96), followed by Makiman131
(24.53). The video by sTaXxCraft has the highest polarity level and the fourth highest subjectivity
level. The video by luzugames has the second highest polarity level and the highest subjectivity
level. However, there are no significant values for these videos in the interaction variables or in the
relationship between them.
The videos on the Makiman131 and elChurches channels scored negative polarity values,
suggesting the presence of negative and hostile comments generating debate and conflict in the
conversation. It is worth noting that the video with the highest negative polarity level (
−
4.83,
Makiman131) also rated the second highest subjectivity level (24.53), one of the lowest like-view ratios,
and the second highest percentage for the dislike-view ratio.
In relation to the tone of the social conversation (Q8), the results reflect low levels of polarity
(5.28 points) and subjectivity (15.5 points). The participants display a low controversy profile with
respect to the polarity and subjectivity of the comments as they barely pass the zero polarity levels,
with a high
score of 19.9 points on a scale from
−
100 to 100, and a maximum subjectivity level of nearly
41 points in only one case and still below the scale average (0–100).
An analysis of the average score for each of the channels does not reveal any atypical or extreme
values (i.e., maximum polarity and subjectivity) to determine any degree of subjectivity or polarity.
There are four YT channels above the average for the whole sample, but the values obtained in these
cases are not sufficient to cause extreme polarization or subjectivity. The values are normalized by
positioning the comments instead on the fringes of neutral tone and relative objectivity.
The videos with the highest polarity and subjectivity levels are found in the screen-sharing/
collaboration, sit-down, and vlog categories (Q9). The polarity and subjectivity levels of the comments
do not reveal any relationship with the duration of the videos (Table 5).
Table 5. Types of videos with highest polarity and subjectivity levels.
By Video Type
Ratio
freq_actions_day/
yout_viewCount
Duration Subjectivity Level Polarity Level YT Channel
Screen-sharing/collab.
17.464% 00:25:02 >8 points *
16 videos/100 =16%
>22 points *
13/100 =13%
VEGETTA777,
luzugames
Sit-down 5.101% 00:07:48 Gymvirtual
Vlog 1.248% 00:11:17 Makiman 131
Source: compiled by authors based on data from TextBlob. * The value corresponds to the third quartile, i.e., only
25% of the videos with the highest subjectivity and polarity levels are above 8 points in subjectivity and none are
above 20 points, and in polarity 25% are above 22 points and only one has a score of 41 points.
No regular trend was found between the polarity and subjectivity levels of the comments and the
time the videos were published (Q10). Finally, no pattern could be identified between the polarity and
subjectivity levels of the comments posted on YT and the comments posted on Facebook (Q11).
4. Discussion and Conclusions
Our analysis reveals a low level of interaction generated by the content of YouTubers in the
sample studied. Comments represent the lowest figure of all. In our research, we were not able to
confirm the assertions of Scolari and Fraticelli (2016), who claim that YouTubers frequently reply to
Soc. Sci. 2019,8, 266 11 of 15
comments on their videos and that the likelihood of responding is greater because the videos are
expanded on hypermedia platforms like Twitter, Facebook, or Instagram. The results show a low level
of interaction on social networks in response to the videos, both on YT and on FB (Q1), with absolutely
no replies by the YouTubers themselves (Q2 and Q3). In the user–user relationship, conversation is also
minimal: On average, only 9.9% of the comments are replies to other comments (Q3). YT’s social media
tools distinguish it from television, yet they are underused by both YouTubers and their audiences.
It would be useful in future research to examine how the low level of participation of YouTubers in
comments influences the activity of their followers and whether there is a cause–effect relationship.
In this preliminary exploration, it was not possible to consider this question.
As Madden et al. (2013) also concluded that the topic matter of user comments is highly
heterogeneous (Q4). In the case studied here, there is a notable number of comments that respond
to a direct question or invitation made by the YouTuber in the video, a strategy to encourage user
participation. The analysis of Weber (2013) is thus corroborated here, as the type of content and how it
is narrated, especially direct appeals, affect participation and interactivity in the comments.
Expressing an emotion or an opinion and supplementing or clarifying information are the main
motives behind commenting on content on social networks (Stroud et al. 2016). According to our
findings, comments were generally made to verbally express emotions, to respond to a direct appeal by
the YouTuber, to praise the YouTuber, or to comment on the most striking or interesting aspects of the
video. These results expand on the motives limited to information seeking and entertainment indicated
in the studies of Khan (2017). However, no direct relationship was found between the volume of
comments received for YT videos and other interaction variables (Q5) like views, likes, or subscribers,
which was confirmed in the studies of Siersdorfer et al. (2010), Jamali and Rangwala (2009), and Lee et
al. (2010). The presence of video games as a topic in 80% of the sample may represent a limitation
of the research, as the criterion chosen (channels with most subscribers and views) inadvertently
resulted in a sample with a prominent presence of a single topic. It would be useful to procure more
heterogeneous samples for future studies.
The polarity and subjectivity levels analyzed were not dependent on the number of subscribers
(Q6) or on any of the other content interaction variables (Q7). The absence of extreme levels of polarity
or subjectivity identified here in response to Q8 coincides with the findings of Lee and Jang (2010) and
Lee (2012), who demonstrated that user opinion was influenced by the comments previously posted by
other users. Thus, the trend in the tone or style of the comments follows the pattern set by comments
posted previously and read by other users before posting their own comment, resulting in a highly
homeostatic and contagious phenomenon, in line with the findings of Von Sikorski and Hänelt (2016).
No significant relationships were revealed between the polarity and subjectivity rates for the
comments on the one hand and the duration of the video, type of video, time of publishing, or
interaction generated on additional platforms like Facebook (Q9, Q10, and Q11) on the other.
According to Cialdini (2001), comments on YouTuber channels exhibit: A medium level of
commitment and consistency; minimal reciprocity; limited social proof; a marked reverence for the
authority of the YouTuber; and contained liking and pronounced scarcity, which increases the value of
the replies chosen by the YouTuber. YouTubers often respond to comments made by their followers
with ad-hoc videos. In the sample of videos selected, responses by YouTubers appear in the new videos
themselves, mixed in with the regular content at random locations within the narration. Users will
thus look out for each new video to see whether the YouTuber has chosen their comments to respond
to; however, the fact that users do not know the exact moment when the YouTuber will make reference
to the followers’ comments gives them another reason to watch the new video in its entirety. In this
way, YouTubers can hold the attention of their audience by means of a carefully designed loyalty
marketing strategy.
In conclusion, interactivity based on commenting is a potential option used by only a small (almost
incidental) proportion of the massive communities of users created around the top YouTuber channels.
Clearly, the interactive potential of YouTuber channels is being underused. Moreover, YouTubers
Soc. Sci. 2019,8, 266 12 of 15
themselves, despite creating parallel profiles on other social networks, rarely participate in them either
personally or through members of their team of collaborators. However, YouTubers do demonstrate
an interest in the social conversation provoked by their videos through three actions: (1) Making
reference to selected comments in subsequent videos (mentioning user names or the content of the
comments identified); (2) giving a “heart” to their favorite comments, facilitating the identification of
their followers’ most read comments; and (3) pinning comments to the top of a comments thread so
that they are more visible and highlighted for other users. In relation to these last two actions, other
users can only like comments to help maintain their visibility in the best positions in the thread. In this
way, YouTubers or their collaborators respond to, manage, and offer feedback on the comments made
by their community.
Although it seems logical to assume that YouTubers would be focused on creating content and
would feel incapable of replying to every comment made in their community, reciprocal interaction
would lend greater authenticity and naturalness to the conversation generated by the content. YouTube
allows creators to designate moderators who can participate in the conversation thread on their behalf,
but this tool is rarely used. Following the social conversation constitutes a very useful source of
information for YouTubers that can help them to secure the loyalty of their audience, correct mistakes,
explore new topics of interest, and adapt their content to the tastes of their community. YouTubers
generate expectations related to the comments they will chose and respond to. This is a widespread
practice that is confirmed by this study. We can therefore conclude that users interact mostly with
each other in the comments section, while also using the opportunity to address and appeal directly to
the YouTuber, but YouTubers interact with their audiences by means of new content. The two-way
exchange is thus delayed in time as the social media response is offered in the form of a new video,
which will in turn generate a new social conversation, feeding the circuit of the virtual community on
the basis of video sharing. Commenting activity is thus exploited and focused to keep the channel
alive with new content. The comments serve a function of linking the different videos together in
temporal succession. They also provide an element of novelty and surprise that keeps the channel
active in the periods between the posting of new videos.
Comments are written text, and all written text has an emotional tone. Commenting is thus the
richest of all possible forms of interaction on social networks because it includes the emotional expression
inherent in liking/disliking and involves an investment of time and effort (engagement) motivated by
the content viewed, and its ultimate objective is to share, to document a reaction, to express an opinion,
to contribute something, or to request more information. Commenters seek to be answered—by
other users, by the YouTubers themselves, or by someone on their team (moderators)—but they also
seek to leave a record, a declaration that “I was here”, watching this specific video. This particular
objective has a meaning of its own, similar to the visitors’ books of the non-digital world, where people
can express the sensations elicited by what they have experienced, or to the initials in trees or the
padlocks on bridges left by couples as a testimony to their relationship. Although it results in abortive
conversations, commenting constitutes rich and intriguing evidence of the fan phenomenon intrinsic
to YouTuber communities.
Author Contributions:
Conceptualization, V.T.-V.; methodology, V.T.-V.; software, V.T.-V. and A.C.-M.; validation,
V.T.-V. and A.C.-M.; formal analysis, V.T.-V. and A.C.-M.; investigation, V.T.-V. and A.C.-M.; resources, V.T.-V.;
data curation, V.T.-V. and A.C.-M.; writing—original draft preparation, V.T.-V. and A.C.-M.; writing—review and
editing, V.T.-V. and A.C.-M.; visualization, V.T.-V.; supervision, V.T.-V.; project administration, V.T.-V.; funding
acquisition, V.T.-V.
Funding:
This research was funded by the Ministry of Science, Innovation and Universities of Spain (EU), grant
number CSO2016-74980-C2-2-R.
Acknowledgments: Martin Boyd (Text translator).
Conflicts of Interest: The authors declare no conflict of interest.
Soc. Sci. 2019,8, 266 13 of 15
References
AMDIA. 2015. Social Media: La Revoluci
ó
n de las Redes Sociales. Available online: www.amdia.org.ar (accessed
on 9 October 2018).
Benevenuto, Fabr
í
cio, Adriano C. M. Pereira, Tiago Rodrigues, Virgilio Almeida, Jussara Almeida, and Marcos
Andr
é
Gonçalves. 2010. Characterization and analysis of user profiles in online video sharing systems.
Journal of Information and Data Management 1: 261–75.
Boyd, Michael S. 2014. (New) participatory framework on YouTube? Commenter interaction in US political
speeches. Journal of Pragmatics 72: 46–58. [CrossRef]
Chang, Pamara F., Janis Whitlock, and Natalya N. Bazarova. 2018. “To Respond or not to Respond, that is the
Question”: The Decision-Making Process of Providing Social Support to Distressed Posters on Facebook.
Social Media & Society 4: 1–11. [CrossRef]
Chau, Clement. 2010. YouTube as a participatory culture. New Directions for Youth Development 128: 65–74.
[CrossRef] [PubMed]
Cheong, France, and Christopher Cheong. 2011. Social Media Data Mining: A Social Network Analysis
of Tweets During The 2010–2011 Australian Floods. Pacific Asia Conference on Information Systems,
PACIS 2011: Quality Research in Pacific Asia, Brisbane, Queensland, Australia. Available online: http:
//bit.ly/cheong-cheong-2011 (accessed on 23 September 2018).
Choi, Kyungsub Steve. 2003. Imposing computer-mediated communication theories on virtual reality. Paper
presented at International Conference on Information Technology: Research and Education, Newark, NJ,
USA, August 11–13; Newark: IEEE. [CrossRef]
Cialdini, Robert B. 2001. Influence: Science and Practice, 4th ed. Boston: Allyn & Bacon.
Del Pino-Romero, Cristina, and Araceli Castell
ó
-Mart
í
nez, eds. 2017. La estrategia publicitaria basada en
influencers. El caso de SmartGirl by Samsung. In Publicidad y Convergencia medi
á
tica. Nuevas Estrategias de
Comunicaci
ó
n Persuasive. Sevilla: Egregius Ediciones, pp. 116–46. Available online: http://bit.ly/publicidad-
convergencia-mediatica (accessed on 11 December 2018).
Dynel, Marta. 2014. Participation framework underlying YouTube interaction. Journal of Pragmatics 73: 37–52.
[CrossRef]
Eckler, Petya, and Paul Bolls. 2011. Spreading the virus: Emotional tone of viral advertising and its effect on
forwarding intentions and attitudes. Journal of Interactive Advertising 11: 1–11. [CrossRef]
Feixas, Daniel, Ernest Codina, and Roger Carandell. 2014. Cómo triunfar en YouTube. Barcelona: La Galera.
Frobenius, Maximiliane. 2014. Audience design in monologues: How vloggers involve their viewers. Journal of
Pragmatics 72: 59–72. [CrossRef]
Gallardo-Camacho, Jorge, and Ana Jorge Alonso. 2010. La baja interacci
ó
n del espectador de v
í
deos en Internet:
caso YouTube España. Revista Latina de Comunicación Social 65: 421–35. [CrossRef]
Galtung, Johan, and Mari Ruge. 1973. Structuring and selecting news. In The Manufacture of News: Social Problems,
Deviance and the Mass Media. Thousand Oaks: Sage, vol. 1, pp. 62–72.
Garc
í
a-Jim
é
nez, Antonio, Beatriz Catalina Garc
í
a, and Mar
í
a Cruz L
ó
pez de Ayala. 2016. Adolescents and
YouTube creation, participation and consumption. Prisma Social: Revista de Ciencias Sociales 1: 60–89.
Hagerstrom, Amy, Alhabash Saleem, and Kononova Anastasia. 2014. Emotional dimensionality and online ad
virality: Investigating the effects of affective valence and content arousingness on processing and effectiveness
of viral ads. American Academy of Advertising. Paper presented at the Annual Meeting of the American
Academy of Advertising, Atlanta, GA, USA, March 27–30.
Hidalgo-Mar
í
, Tatiana, and Jes
ú
s Segarra-Saavedra. 2017. El fen
ó
meno YouTuber y su expansi
ó
n transmedia.
An
á
lisis del empoderamiento juvenil en redes sociales. Fonseca, Journal of Communication 15: 43–56. [CrossRef]
IAB Spain. 2018. Estudio Anual Redes Sociales 2018. Available online: http://bit.ly/iabspain-2018 (accessed on
9 March 2019).
Jamali, Salman, and Huzefa Rangwala. 2009. Digging Digg: Comment Mining, Popularity Prediction, and Social
Network Analysis. Paper presented at the 2009 International Conference on Web Information Systems and
Mining, Shanghai, China, Novermber 7–8; Shanghai: IEEE, pp. 32–38. [CrossRef]
Kavoori, Anandam. 2015. Making sense of YouTube. Global Media Journal 13: 1–25.
Khan, M. Laeeq. 2017. Social media engagement: What motivates user participation and consumption on
YouTube? Computers in Human Behavior 66: 236–47. [CrossRef]
Soc. Sci. 2019,8, 266 14 of 15
Kim, Youngju. 2015. Exploring the Effects of Source Credibility and Others’ Comments on Online News Evaluation.
Electronic News 9: 160–76. [CrossRef]
Kirby, Justin. 2004. Getting the Bug. Brand Strategy 184: 33.
Krishna, Amar. 2014. Polarity Trend Analysis of Public Sentiment on YouTube. Master’s thesis, Iowa State
University, Ames, IA, USA. [CrossRef]
Ksiazek, Thomas B., Limor Peer, and Kevin Lessard. 2014. User engagement with online news: Conceptualizing
interactivity and exploring the relationship between online news videos and user comments. New Media &
Society 18: 502–20. [CrossRef]
Lange, Patricia G. 2007. Commenting on Comments: Investigating Responses to Antagonism on YouTube. Paper
Presented at Society for Applied Anthropology Conference, Tampa, FL, USA, March 31; Annenberg Center for
Communication, vol. 31, pp. 163–90. Available online: http://bit.ly/lange-2007 (accessed on
15 March 2018
).
Lavado, Antonio. 2013. El consumo de YouTube en España. Global Media Journal México 7: 76–92.
Lee, Eun-Ju. 2012. That’s Not the Way It Is: How User-Generated Comments on the News Affect Perceived Media
Bias. Journal of Computer-Mediated Communication 18: 32–45. [CrossRef]
Lee, Eun-Ju, and Yoon Jae Jang. 2010. What Do Others’ Reactions to News on Internet Portal Sites Tell Us? Effects
of Presentation Format and Readers’ Need for Cognition on Reality Perception. Communication Research 37:
825–46. [CrossRef]
Lee, Jong Gun, Sue Moon, and Kave Salamatian. 2010. An approach to model and predict the popularity of online
contents with explanatory factors. Paper presented at the International Conference on Web Intelligence and
Intelligent Agent Technology, Toronto, ON, Canada, August 31–September 3; New York: IEEE, pp. 623–30.
Madden, Amy, Ian Ruthven, and David Mcmenemy. 2013. A classification scheme for content analyses of YouTube
video comments. Journal of Documentation 69: 693–714. [CrossRef]
Maniou, Theodora A., and Philemon Bantimaroudis. 2018. Hybrid salience: Examining the role of traditional and
digital media in the rise of the Greek radical left. Journalism, 1–18. [CrossRef]
Murolo, Norberto Leonardo. 2010. Post-zapping: Transmite t
ú
mismo. YouTube como la televisi
ó
n posmoderna.
Razón y Palabra 71: 15.
Pang, Bo, and Lillian Lee. 2008. Opinion mining and sentiment analysis. Foundations and Trends in Information
Retrieval 2: 1–90. [CrossRef]
Phelps, Joseph E., Regina Lewis, Lynne Mobilio, David K. Perry, and Niranjan Raman. 2004. Viral Marketing or
Electronic Word-of-Mouth Advertising: Examining Consumer Responses and Motivations to Pass Along
Email. Journal of Advertising Research 44: 333–49. [CrossRef]
Ramos-Serrano, Marina, and Paula Herrero-Diz. 2016. Unboxing and brands: youTubers phenomenon through
the case study of EvanTubeHD. Prisma Social: Revista de Ciencias Sociales 1: 90–120.
Rego, Sandra, and Luis M. Romero-Rodr
í
guez. 2016. Representaci
ó
n discursiva y lenguaje de los youTubers
españoles: estudio de caso de los gamers más populares. Index Comunicación6: 197–224.
Rotman, Dana, and Jennifer Preece. 2010. The ‘WeTube’ in YouTube—Creating an Online Community through
Video Sharing. International Journal of Web Based Communities 6: 317–33. [CrossRef]
Rull, Antonio. 2014. Ser Youtuber es un Infierno Creativo. El Diario. May 14. Available online: http://bit.ly/rull-2014
(accessed on 8 October 2018).
Sabich, Mar
í
a Agustina, and Lorena Steinberg. 2017. Discursividad youTuber: afecto, narrativas y estrategias
de socializaci
ó
n en comunidades de Internet. Revista Mediterr
á
nea de Comunicaci
ó
n/Mediterranean Journal of
Communication 8: 171–88. [CrossRef]
S
á
ez Barneto, Gloria, and Jorge Gallardo Camacho. 2017. La relaci
ó
n de los youTubers con la publicidad y sus
espectadores. El caso de YouTube España. Telos: Cuadernos de Comunicación e Innovación107: 47–57.
Scolari, Carlos, and Dami
á
n Fraticelli. 2016. Nuevos sujetos medi
á
ticos en el ecosistema de medios: el caso de los
youTubers españoles. Paper presented at V Congreso de la Asociaci
ó
n Argentina de Estudios De Cine y
Audiovisual (ASAECA), Bernal, Argentina, August 19; Available online: http://bit.ly/scolari-fraticelli-2016
(accessed on 12 October 2018).
Siersdorfer, Stefan, Sergiu Chelaru, Wolfgang Nejdl, and Jose San Pedro. 2010. How Useful are Your comments?
Analyzing and Predicting YouTube Comments and Comment Ratings. Paper presented at the 19th
International Conference on World Wide Web, Raleigh, NC, USA, April 26–30; Raleigh: ACM, pp. 891–900.
Available online: http://bit.ly/siersdorfer-et-al-2010 (accessed on 23 November 2018).
Soc. Sci. 2019,8, 266 15 of 15
Stroud, Natalie Jomini, Emily Van Duyn, and Cynthia Peacock. 2016. News Commenters and New Comment
Readers. Engaging New Projects 1: 1–21.
Sureka, Ashish, Ponnurangam Kumaraguru, Atul Goyal, and Sidharth Chhabra. 2010. Mining YouTube to
Discover Extremist Videos, Users and Hidden Communities. In Information Retrieval Technology. AIRS 2010.
Lecture Notes in Computer Science. Edited by Pu-Jen Cheng, Min-Yen Kan, Wai Lam and Preslav Nakov. Berlin
and Heidelberg: Springer, vol. 6458, pp. 13–24. [CrossRef]
Tannen, Deborah. 1999. The Argument Culture: Stopping America’s War of Words. New York: Ballantine.
Thelwall, Mike, Kevan Buckley, and Georgios Paltoglou. 2012. Sentiment strength detection for the social web.
Journal of the American Society for Information Science and Technology 63: 163–73. [CrossRef]
Tur-Viñes, Victoria, Patricia N
ú
ñez G
ó
mez, and Mar
í
a-Jos
é
Gonz
á
lez-R
í
o. 2018. Menores influyentes en YouTube.
Un espacio para la responsabilidad. Revista Latina de Comunicación Social 73: 1211–30. [CrossRef]
Vilares, Manuel, Elena S
á
nchez Trigo, Carlos G
ó
mez-Rodr
í
guez, and Miguel A. Alonso. 2017. Tecnolog
í
as de la
lengua para análisis de opiniones en redes sociales. Procesamiento del Lenguaje Natural 59: 125–28.
Von Sikorski, Christian, and Maria Hänelt. 2016. Scandal 2.0: How Valenced Reader Comments Affect Recipients’
Perception of Scandalized Individuals and the Journalistic Quality of Online News. Journalism & Mass
Communication Quarterly 93: 551–71. [CrossRef]
Weber, Patrick. 2013. Discussions in the comments section: Factors influencing participation and interactivity in
online newspapers’ reader comments. New Media & Society 16: 941–57. [CrossRef]
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
