Page 1 of 65
Reader Comments on News Articles
Critical or Complementary
Monique Hiller
Master’s Thesis
15 June, 2016
Supervisor: Oscar Hemer
Master’s Programme in Media and Communication Studies: Culture,
Collaborative Media, and Creative Industries
One-Year Masters – 15 credits
Page 2 of 65
Abstract
Guardian and Zeit online articles attract a large number of reader comments in general
and those that include links specifically. Using both quantitative and qualitative content
analysis, the reader comments that include links in 80 articles on four topics – The
Refugee Crisis, the Syrian Conflict, Elections and Artificial Intelligence – from both
newspapers were classified according to type and pro or contra stance to the article they
were posted in response to.
The aim was to discover whether a majority of these comments agreed with or opposed
the main premises of the articles they were responding to and to determine whether
different topics results in different reactions. In addition to that the goal was to try and
ascertain whether these findings indicated falling trust in journalists or not, and what
readers sought to achieve by commenting and posting links, if in fact they had a concrete
motivation.
Findings demonstrated that 61.1% of reader comments with links disagreed with the
articles they were posted to, and controversial topics were more likely to attract links by
readers and greater disagreement. Previous research and this study could not clearly
conclude whether audience trust in journalists is falling or whether that was a mere
extension of falling trust in authorities and organizations for example. It could also not
determine whether audiences comment and share links to effect some form of change or
whether they simply seek a platform to express their opinion without any conscious goal
in mind.
Page 3 of 65
Contents
1- Introduction ... 5
2- A Map of the Research Paper ... 8
3- Context ... 8
3.1- Falling Trust in Journalists and the Media as a Whole? ... 8
3.2 - Article Reader Comment Functions – Yes or No: ... 9
3.3 - Topic Background: ... 10
3.3.1 - Refugee Crisis: ... 10
3.3.2 - Syrian Conflict: ... 11
3.3.3 - Elections: ... 12
3.3.4 - Artificial Intelligence: ... 13
4- Definitions ... 14
4.1- Article: ... 14
4.2 - Pro/Contra:... 14
4.3 - Mainstream media: ... 15
4.4 - Background on Zeit: ... 15
4.5 - Background on the Guardian: ... 16
5- Methods ... 17
5.1 - Research Questions: ... 17
5.2 - Methodology and Researcher Bias: ... 17
5.3 - Sampling: ... 18
5.4 - Coding Unit: ... 19
5.5 - Language: ... 20
5.6 - Links:... 20
5.7 - Coding Scheme: ... 20
5.8 - Analysis: ... 21
5.9 - Ethics: ... 21
5.10 - Validity and Limitations:... 22
6- Theory and Previous Research ... 22
7- Findings ... 30
7.1 - Quantitative Data Analysis: ... 30
7.1.1 - Comments and Links – Totals and Ratios ... 30
7.1.2 - Language ... 33
7.1.3 - Link Types ... 34
7.1.4 - Pro or Contra Articles: ... 37
Page 4 of 65
7.2.1 - Artificial Intelligence Recurring Themes ... 40
7.2.2 - Refugee Crisis Recurring Themes ... 42
7.2.3 – Syria Conflict Recurring Themes ... 44
7.2.4 – Elections Recurring Themes ... 47
7.2.5 - Overall Note: ... 50
8- Discussion ... 50
9- Conclusion ... 51
10-
References ... 54
APPENDIX ... 65
Table of Tables:
Table 1 - Comments and Links - Totals and Ratios ... 30
Table 2 - Link Totals by Newspaper ... 32
Table 3 - Language Distribution and Frequency of Links... 33
Table 4 - Distribution and Frequency of Link Types ... 35
Table 5 - Links to Mainstream Media and Same Newspaper ... 36
Table 6 - Agreement and Disagreement with articles by Newspaper and Topic ... 38
Table 7 - Combined Agreement and Disagreement with Articles by Topic ... 38
Table 8 - Artificial Intelligence Topic - Themes and Sub-themes ... 41
Table 9 - Refugee Crisis Topic - Themes and Sub-themes ... 44
Table 10 - Syria Conflict Topic - Themes and Sub-themes ... 46
Page 5 of 65
1- Introduction
For a number of years, I have had the subjective impression that reader comments on
news articles, particularly those concerning controversial political topics, have
contradicted or disapproved of the premises or perceived agenda of the articles or authors
they were responding to.
To me it has seemed that a very specific viewpoint was being maintained in the
mainstream media in many different countries around the world, while any opposing
views are being left for more independent news outlets. The speed and globalization of
media these days makes it easier to access information and news from around the world
quickly but also makes it easier for journalists to simply take reports as is and repost them
without adding any insights or details of their own. The consolidation of media
corporations also seemingly adds to the problem of a lack of individuality of what
constitutes news or objectivity. I have felt that this impression of mine was reflected in
the reader comments I read that were posted in reaction to articles.
My personal impression has also been that I learn more from comments than the articles
themselves, and that comment sections are often more balanced than these articles, and
provided links to interesting background information or differing viewpoints.
Therefore – and due to the large volume of comments per article and limited timeframe
of this research - I restricted my research in this study to comments that included links.
These comments were posted to 80 articles, 40 in the Zeit newspaper and 40 in the
Guardian newspaper, with these articles covering four different topics: The Syrian
Conflict, the Refugee Crisis, Elections and Artificial Intelligence.
My aim was to try and objectively determine whether actual data confirmed the
aforementioned subjective observations or not, whether there was any difference in this
regard depending on topic and newspaper, what that implies about the current state of
journalism and readers’ trust in what they read and what they seek to achieve by
commenting and including links in their comments.
Page 6 of 65
This is an entirely subjective observation I have made over these two years and had not
tested for accuracy in any way other than by reading that others have had similar reactions
before I began my work on this thesis.
Therefore, my four interconnected research questions were as follows:
Q1: Do a majority of readers posting comments that include links agree or disagree
with article and author premises?
Q2: Does this opinion vary depending on the topic in question?
Q3: Does this potentially imply something about a generalized lack of trust in the
media and journalists?
Q4: What do readers hope to achieve by commenting and including corroborating
links?
My choice of topics was inspired by my own interests, topics that seemed to trigger
controversy and topics that were less value-laden but also interesting to me, to serve as a
comparison. The Syrian conflict in particular was what made me notice a sense of
discontent in reader comments in the first place, and notice a certain frustration and a
perception that the media reporting was not reflecting much of the reality on the ground
for whatever reasons and that it was up to readers themselves to correct or redress this
imbalance using their comments, posting links to background information or alternative
viewpoints and offering more insight and balanced views than the actual reports.
This topic is something I felt needed to be analyzed and potentially addressed. It should
not be so difficult for the public to access objective information. Not everyone has the
time, skills or patience to troll through endless repetitive articles to find reliable reporting
or a neutral observer or go through the comment section of each article in the hope of
finding useful information or links to more balanced or opposing viewpoints and those
who do usually have specific aims and motivations.
Page 7 of 65
Do they wish to influence other readers or the journalists themselves or neither and just
seek an outlet to express an opinion without any specific goal in mind? Do they wish to
complement coverage of a certain issue or contribute a completely opposing view?
Several official surveys have shown that readers no longer trust journalists, or at least the
level of trust has decreased dramatically due to a number of factors. My observation has
been that one of the main factors (though this aspect is not always mentioned) has been
the streamlining of coverage on particularly divisive or controversial topics, where
readers in the comments on articles share their scepticism regarding the neutrality and
objectivity of journalists, researching their backgrounds and noting their work with think
tanks for example. In some cases, this has also been happening for non-controversial
topics, where perhaps it is not a matter of distrust but a wish to tangibly contribute to the
discussion and expand its horizon.
In this research study the task was to conduct both a qualitative and quantitative content
analysis on the comments that included links and classify them based on their type, their
language, whether they were links to mainstream media or to the same newspaper and
also according to whether they were pro or contra the original article’s premise or stance.
The objective was to measure whether readers do in fact have a largely negative stance
towards journalism in general and/or specific articles or whether that depended on the
topic in question or not by tabulating the results of this analysis, breaking them down into
topic- and newspaper specific categories.
The research on the motivations and opinions of readers who comment on articles has not
yet received as much focus as it could, however some related research exists such as
Tenenboim and Cohen (2013).
The matter of reader trust can be closely linked to Bourdieu (1977)’s concept of doxa and
whether readers still believe that journalists are still adhering to their profession’s main
behaviour- and principle-related tenets. A loss of trust can lead to a loss of cultural capital
– Bourdieu (1977), and a vicious cycle of ever-decreasing ability to influence public
opinion. There is no doubt that readers’ dispositions – Bourdieu (1990) – play a role in
the fact that they comment and in the formation of their opinions.
Page 8 of 65
2- A Map of the Research Paper
First of all, I will provide some context and background for the paper and the topics
covered in the articles selected for analysis. Next I will provide definitions of some of the
classifications and expressions used in this study and their scope within its context. This
will be followed by a detailed breakdown of the methods used to develop and analyse the
data collected. Next I will introduce and discuss relevant theory material and previous
research that can shed light or help explain and respond to the research questions central
to this paper. This will be followed by a detailed quantitative and qualitative analysis of
the collected data itself. I will then share my ideas regarding potential further discussion
and research related to this research followed by a conclusion to sum it all up.
3- Context
3.1- Falling Trust in Journalists and the Media as a Whole?
Despite a high level of news consumption and readership, trust in journalists has
seemingly seen a sharp decline in the UK, the US and Germany over the past few years
– the countries this research is mainly focused on. A Gallup poll conducted in 2015
showed that readers had very little faith in the “honesty and ethical standards” of
journalists (Gallup, 2015). An Ipsos-Mori poll for 2015 lists journalists as among the 5
least trusted professions, in the illustrious company of politicians, government ministers,
real estate agents and business leaders - Ipsos MORI Veracity Index: Trust in Professions
(2015).
According to a poll that the Zeit newspaper itself commissioned and which was conducted
by infratest dimap on its behalf in 2015, 60% of Germans either have little or no trust in
the media - Zeit Online Report (2015c).
Page 9 of 65
However, these surveys are not conclusive evidence of growing distrust and may not be
applicable to all media outlets across the board in the countries in question. In addition to
that, there are surveys such as Newman (2015) that imply that trust in the media remains
relatively high.
Time and further research will be necessary to reach a definitive conclusion.
3.2 - Article Reader Comment Functions – Yes or No:
In this research I am focusing on two newspaper websites that still allow and even
welcome reader comments on most of their articles. This is not the case for many other
British or German newspapers. Some never offered this option to their readers, while
others have recently banned comment sections or relegated any reader-opinion expression
to forums, something that reduces its immediacy. It seems logical to conclude that
commenting about a specific article in a forum is less convenient than directly below an
article and it is also more inconvenient for those who seek to read these comments.
Forums develop their own dynamic and will more likely lead to comments being about
the topic in general more than about a specific article.
“[M]any newspapers have abandoned the practice of allowing comments.
Online news sites have adopted a variety of strategies to deal with offensive
comments, including turning “comments off,” not archiving comments, and
adopting aggressive comment moderation policies.” (Hughey and Daniels,
2013, p.332)
The issue of abuse or intolerance expressed in comments and the liability of newspapers
for such comments has been a matter of contention over the past few years, and has often
been cited as a reason to ban or not introduce a reader comment section at all.
In 2015 the European Court of Human Rights (ECHR) ruled that websites were liable for
reader comments (Moody, 2015), however the ECHR reversed the ruling in 2016 when
it decided that a Hungarian news portal could not be held liable for third party comments
(The European Court of Human Rights rules again on liability for third party comments,
2016).
Page 10 of 65
This confusion does not of course help matters, but the Zeit and Guardian reader comment
sections have strict guidelines and are carefully and rapidly moderated to ensure that
offensive comments and content is removed as quickly as possible.
Choosing not to allow readers this option shows a certain amount of fear on the side of
both the journalists themselves - Schreyer (2014) - with some asking for the comment
function on their articles to be disabled, but also from the side of media outlets
themselves, who have absorbed the information and trends discussed above regarding the
falling trust in the media and in journalists.
Another common complaint by journalists and media bosses is that comment sections are
invaded by “paid trolls”, often citing Russia as a source of such paid-for propaganda with
commenters described as “Putinbots” (Soldatov and Borogan, 2015) who try to influence
the direction of discussions and to portray a more diverse and fragmented public opinion.
3.3 - Topic Background:
The four topics covered in this research have been broadly categorized as the Refugee
Crisis, the Syrian Conflict, Artificial Intelligence and Elections but to provide readers
with a more concrete context I will attempt to establish where each of these topics stood
during the period designated for this paper, from October 1
st, 2015 to March 31
st, 2016.
3.3.1 - Refugee Crisis:
The articles covered span the time-frame in the immediate aftermath of what many have
described as German Chancellor Angela Merkel’s “invitation” for Syrian refugees in
particular to make their way to Germany, implicitly suspending the Dublin agreement
that stipulates that refugees must apply for asylum in the EU country they first reach. This
supposed invitation was at first welcomed by German citizens but the opinion has
somewhat shifted in recent months as a result of a number of criminal incidents that
allegedly involved migrants and refugees and due to the large number of people arriving
and the authorities’ difficulties in handling all in a timely and organized manner.
Page 11 of 65
Chancellor Merkel has been praised by some for her empathy and willingness to welcome
so many people in need of refuge, but she has also been criticized for acting unilaterally
and demanding that other European countries follow suit.
Eastern European - countries and slowly but surely a number of Western countries - have
enforced existing borders or introduced border controls even between Schengen Zone
nations.
During the last few weeks covered in this research and with Greece having to bear the
consequences of this chain reaction of border controls the EU has reached an agreement
with Turkey that stipulates that any refugees arriving by boat from Turkey to Greece
would be returned to Turkey with an equivalent number of Syrian refugees who were
returned relocated directly from Turkey to the EU, in addition to a specific amount of
Syrian refugees who would be directly selected from refugees camps in Turkey and
relocated to the EU. Not all have been in agreement with this deal as Turkey itself has
been criticized for a number of political and human rights violations, and Turkey is still
in the process of implementing a number of criteria that would allow its citizens to travel
to the EU without a visa as part of this deal, which also includes EUR 6 billion in
compensation and to cover the costs of hosting a large number of refugees in Turkey.
3.3.2 - Syrian Conflict:
The research covers the period coinciding with the start of Russian involvement and
support for the Syrian government and Syrian President Bashar Al Assad in a quest to
retake territory controlled either by rebels or by ISIS. At the same time a coalition of
Western and Arab nations had been bombarding ISIS strongholds but not recording
significant progress.
The Paris terrorist attack that took place in November increased the intensity of coalition
involvement in the battle against ISIS with France and the UK in particular upping the
ante.
Page 12 of 65
Tensions between Russia and Turkey escalated during this period, culminating with
Turkey shooting down a Russian jet that had allegedly violated its airspace, with the pilot
of the aircraft being shot and killed by Syrian Turkmen rebels as he descended by
parachute.
During the latter period covered by this research there was significant progress in terms
of reaching a ceasefire between warring factions in Syria, but one that did not include the
groups deemed to be terrorists.
The Syrian government with Russian air support was able to liberate the ancient city of
Palmyra from ISIS control and achieve significant breakthroughs in other areas of the
country.
Towards the end of the period covered, Russia announced that it had achieved what it had
set out to do and would therefore withdraw most of its forces from Syria but remain
present.
3.3.3 - Elections:
The period covered in this research spanned a number of national and regional elections
or the run-up towards them. In December the media focused on the parliamentary
elections in Spain and the rise of two new parties – Podemos and Ciudadanos – that
represented a threat to the usual duopoly of the People's Party (Partido Popular, PP) and
the Spanish Socialist Workers' Party (Partido Socialista Obrero Español, PSOE).
During the same month municipal elections took place in Saudi Arabia which for the first
time allowed women to stand as candidates and which resulted in 17 of them being elected
to public office.
February 2016 saw coverage of the parliamentary elections in Iran with the world
observing whether a moderate alliance would be able to gain a majority.
Page 13 of 65
In March the media focused on regional elections in Germany which were widely seen as
a litmus test for Chancellor Merkel’s refugee policy and which saw her Christian
Democratic Union (Christlich Demokratische Union Deutschlands, CDU), as well as her
grand coalition partner the Social Democratic Party of Germany (Sozialdemokratische
Partei Deutschlands, SPD) party suffer some losses, with the recently established party
Alternative for Germany (Alternative für Deutschland, AfD) gaining significant ground
and the Alliance '90/The Greens (Bündnis 90/Die Grünen, GRÜNE) stepping up to fill
some of the void left by the two main parties.
The lead-up to the US Presidential elections – the Primaries - have also been covered
comprehensively by the media, monitoring the surprising results achieved by Bernie
Sanders for the Democrats and Donald Trump for the Republican party.
3.3.4 - Artificial Intelligence:
The main bulk of coverage concerning Artificial Intelligence during this period was
triggered by AlphaGo - a computer program developed by Google DeepMind – defeating
the World Go Champion Lee Seedol in a best of five competition held in March.
Go is an ancient Chinese board game that involves a 19x19 grid and players taking turns
to place stones on empty square intersections on the board. The game is described as
having simple rules but its exceedingly high number of potential moves means that in
practice it is more difficult to play and master than Chess. Much of its strategy is
considered intuitive and not based on advance calculations.
This was always seen as an obstacle for artificial intelligence’s capacity to outplay human
beings but AlphaGo was able to learn from playing others and itself - after it had been
programmed with three different types of networks that help it do so – and
comprehensively defeat the reigning champion. This ability to learn or teach itself beyond
its programmed parameters triggered a fear in many that we were on the verge of seeing
the development of Artificial General Intelligence – where AIs are able to perform any
task a human being is capable of – or even Superintelligence – where AIs can easily
outsmart the most intelligent of human beings.
Page 14 of 65
4- Definitions
4.1- Article:
The Cambridge Dictionaries Online (2016) define an article as “a piece of writing on a
particular subject in a newspaper or magazine, or on the internet” and that is the scope of
definition I have used in this research. In most cases – and not unexpectedly - the
classification of articles was not in doubt, however on rare occasion it was a matter of
judgment as to whether to consider a specific link an article or not due to its format.
4.2 - Pro/Contra:
Though it may seem simple to classify any reader comment as in favour of or against the
article it has been posted in response to, in reality this was again a matter of judgment. I
have made the conscious decision to not consider the links posted in isolation, but within
the context of the comment itself. Sometimes readers post links in their comments to point
out that they do not agree with them, and I believe that taking the comment as a whole
including the link is the fairer approach. Again this was not often the case and on most
occasions the links posted matched the opinion of the comment they were included with.
Another complication was the fact that not all the articles covered had a clear direction,
opinion or agenda – which is as it should be of course unless they are opinion pieces – so
it then became difficult to determine if the readers’ comments were in agreement,
disagreement or simply neutral. In many cases it was about what was NOT said in articles
which gave them a somewhat one-sided tone. For example, when it came to Artificial
Intelligence articles most of the time the disagreement with authors was about how
concerned everyone should be regarding the potential future danger this technology
poses. Readers simply had a more intense response to the topic and were not necessarily
in agreement or disagreement with the article itself in an explicit manner. In these cases,
I chose to consider these responses and links as contra as to me it was also about readers
not feeling their concerns reflected in the media they consume.
Page 15 of 65
When it came to the more controversial topics, the articles were often more obviously
one-sided and as a result it was somewhat easier to classify the comments and links as
pro or contra. It was however again a matter of judgment on occasion, particularly when
the comments addressed the original articles but also included a more general opinion on
the topic or on journalism as a whole for example.
4.3 - Mainstream media:
For this research I have chosen to classify mainstream media to include popular media
that is widely read or seen both inside one country and outside it to a relative extent.
I did not focus on the political affiliation of these mainstream media outlets as that would
have complicated this analysis in a way that would not have been feasible to investigate
in such a short span of time allocated for this research. It would however be interesting
to delve into this aspect in future studies.
For the purpose of this research the important aspect to me was whether disagreement
with specific articles reflected a general distrust of mainstream media in general and
therefore the number of links to such mainstream media may point in the direction of
whether that is the case or not.
4.4 - Background on Zeit:
The original weekly newspaper was launched in 1946. It is classified as moderate to
liberal depending on whose definition you adhere to and it continues to be published
weekly, with print circulation figures healthy at 512,075 copies of the newspaper sold and
total readership of over 2 million people (DIE ZEIT rate card, 2015, p.5).
Page 16 of 65
According to the newspaper’s own 2015 rate card it is:
“…Germany‘s premier opinion-leading weekly newspaper. It reports on
topics from the worlds of politics and business, culture and science,
technology and medicine, society and education, travel, lifestyle and sport.
In-depth background, well-researched facts and critical analysis make DIE ZEIT
an important source of up-to-date information. DIE ZEIT lays down the
agenda, takes a stand, discusses issues from different perspectives and shapes
opinions…” (DIE ZEIT rate card no. 60, effective Jan. 1, 2015, p.4)
Though this may be specifically about the newspaper, it applies equally to the newspaper’s
online portal as well. The website is of course more up-to-date, with new articles posted
several times a day, depending on newsworthiness, urgency and interest. According to the
Alexa analytics website Zeit.de is ranked 71
st(as of 12 May, 2016) in Germany. 77.4% of
the website’s visitors are from Germany, 5.3% from Austria, 2.9% from Switzerland,
1.3% from the USA and perhaps surprisingly 1% from Azerbaijan (Alexa, 12 May 2016).
4.5 - Background on the Guardian:
First published on May 5
th, 1821 on a weekly or bi-weekly basis until 1855, the Guardian
daily newspaper has an average circulation of 187,000 daily and an average 1,027,000
readership according to information published on the newspaper’s website. It had a
clearly liberal agenda from the outset and, though it is viewed by many or even most as
having a left-wing leaning, on the newspaper’s website it is stated that the newspaper has
no political affiliation and has in fact supported different parties at different times in its
history (Guardian Political Affiliation, 2008).
According to Alexa statistics the Guardian website is ranked 20
thin the UK and 114
thin
the USA, with 31.9% of visits from the USA, 21.5% from the UK, 5.8% from India, 3.8%
from Australia and 2.9% from Canada (Alexa, as of 12 May, 2016).
Page 17 of 65
5- Methods
5.1 - Research Questions:
Q1: Do a majority of readers posting comments that include links agree or disagree
with article and author premises?
Q2: Does this opinion vary depending on the topic in question?
Q3: Does this potentially imply something about a generalized lack of trust in the
media and journalists?
Q4: What do readers hope to achieve by commenting and including corroborating
links?
5.2 - Methodology and Researcher Bias:
I performed both a quantitative and qualitative content analysis of these links. To classify
the comments with links quantitatively I wanted to determine the proportion of comments
posted in general to the different topics and the ratio of links to comments. I wanted to
establish what types of links were shared and their relative frequency, the language they
are most often in and whether they link back to mainstream media or perhaps to the same
newspapers of the original articles, i.e. the Zeit and Guardian respectively. and how often.
Finally, I wanted to know how many comments could be classified as pro and how many
as contra the articles to which they were posted. All this aimed to establish a pattern or
several that could provide a more comprehensive picture of the readers who take the time
to not just post comments but to also reinforce them perhaps using links.
The qualitative content analysis sought to provide a more in-depth understanding of the
reasons behind readers’ agreement or disagreement with author and article premise.
Page 18 of 65
To achieve that I attempted to discover a number of themes that dominated both the pro
and the contra arguments for each topic and was able to recognize a few such recurring
themes.
In my research in general and for this study in particular I used a critical realist and
inductive approach to gather my data before drawing any conclusions – Collins (2010)
and Saunders, Lewis and Thornhill (2009). Though I had some initial observations that
triggered my interest in the topic I did not formulate any concrete hypotheses, allowing
the readers “to speak for themselves” using their comments and simply interpreting their
words and apparent intentions as objectively as possible within the context of each article
and each topic.
In terms of axiology and according to Collins (2010) I myself did not enter this research
completely unbiased, as I have for a number of years, particularly since the beginning of
the so-called Arab Spring, begun to realize not only that comment sections are more
balanced than the articles themselves – more often than not at least – but also that I seem
to learn a great deal more from the reader comments than I do from these articles. On
occasion this has been because readers have more insight or background knowledge but
generally agree with the article’s premise, but on others it has been because they provide
links to material that challenges or even destroys the author’s arguments, making it clear
that a hidden – or even not-so-hidden - agenda may be at work.
I also have a very specific viewpoint regarding several of the topics I focused on and
could be guilty of selective interest/comprehension though I try to keep an open mind.
There is however no way to completely exclude that my bias may have impacted how I
perceived comments to be pro or contra the articles analysed, much as I may have tried.
5.3 - Sampling:
For my analysis of links posted in reader comments I selected four topics and two
newspapers, collecting the links from comments posted to 10 articles per topic per
newspaper. This meant that my total selection of material encompassed 80 articles.
Page 19 of 65
The topics I selected were in no particular order: The Syrian Conflict, the European
Refugee Crisis, Elections and Artificial Intelligence.
The first two are controversial and inspire audience and readership investment and often
even extreme opinions and reactions. Comments on articles covering these two topics are
usually significantly higher in number compared to the other two. Elections inspire fewer
comments and less of a personal involvement, while artificial intelligence has an
altogether different type of response.
The two newspapers I selected were the Guardian and the Zeit, newspapers. The goal was
to cover similar yet different commenting environments and cultures, with the main
obvious difference being the language of the articles themselves and the language of
comments. The two newspapers have a similar political outlook but of course a different
focus when it comes to both local and international issues, often depending on how they
potentially affect the citizens of their countries.
I selected articles (news reports and opinion pieces) that had been published between
October 1st 2015 and March 31st, 2016, to secure a wide enough scope to allow for there
to be sufficient articles for each topic and to allow for developments to have occurred to
trigger and inspire conversation with being too outdated.
I attempted to select similar sub-topics for articles from each topic from both newspapers,
so that they were slightly easier to compare, and so that recurring themes and concerns
could be better grouped together if they existed.
5.4 - Coding Unit:
Each comment that included a link was considered as one coding unit with its
characteristics then determined based on type, language, mainstream media or same
newspaper connection and whether it was pro or contra the premise of the article it was
posted to.
Page 20 of 65
Ideally I would have liked to analyse the comments that did not contain links as well to
provide greater context and a more comprehensive understanding of reader opinions, even
if these were not actively involved in guiding the conversation in one direction or another
or at least attempting to do so but time constraints prevented this.
Comments on each article were counted, with the total percentage of comments sharing
links tabulated. Then each link was evaluated as to whether it was in agreement with the
article, in opposition or neutral.
5.5 - Language:
As a German and English speaker of equal capabilities I have conducted my own instant
translations of the comments and links analyzed and reported on throughout this paper. I
have only included paraphrasing of the ideas expressed implicitly or explicitly in reader
comments without using any direct or specific quotes.
5.6 - Links:
The links that have been analyzed in this research are all listed by newspaper and article
in the appendix to this paper. The lists also include the individual coding and classification
for each link.
5.7 - Coding Scheme:
The links were coded as follows:
Link Type:
Article, Video, Image, Infographic, Official Report, Wikipedia
Entry, Website, Audio, Facebook Post, Twitter Post, Research
Paper, Profile, Book, Book Review, Country Report, Quote, Flyer,
Election Programme, Bad link, N/A
Page 21 of 65
Language:
English, German, Arabic, French, Spanish, Czech, Aramaic,
Hebrew, Bad link, N/A
Mainstream Media: Yes (or none)
Same Newspaper:
Yes (or none)
Pro Article:
Yes and short description (or none)
Contra Article:
Yes and short description (or none)
5.8 - Analysis:
After all 80 articles - and a total of 1860 links - were coded, frequency counts were used
to describe the overall and topic- or newspaper-specific trends regarding the
characteristics of the links and their overall agreement or disagreement with the premises
of articles covered.
5.9 - Ethics:
The ethical concerns that are applicable to this research paper mainly involve myself as a
researcher who has a great interest in politics and therefore well-formed opinions that
may have coloured my perception of both the articles and the comments and links posted
to them.
In some cases, my perception of language nuances may have been influenced by these
opinions, making it difficult for me to make completely objective classifications.
The timeframe covered in this research that may have included specific incidents
triggering a shift in opinion and the choice of some controversial alongside
non-controversial topics may also have influenced results and must also be considered.
It is important to note that these comments and links cannot be used to extrapolate the
data to a more generalized theory. The comments and links may not be representative,
and given the popularity of these two newspapers it cannot be excluded that some or many
of the comments have not been planted in some way.
Page 22 of 65
5.10 - Validity and Limitations:
In terms of validity of the study, confining my research to only two newspapers and to
only the comments that included links does not of course provide a true reflection of
reader comment opinion. The totality of comments may indeed provide a completely
different result.
The selection of articles was as random as possible but also attempting to balance out
between some articles with a large number of comments and others with less, as well as
trying to cover similar sub-topics from both newspapers. The selection of news reports or
opinion pieces was also random and may not have been balanced either within one topic
or within each newspaper and this may also have impacted the difference in the levels of
agreement or disagreement in the comments and links posted for each topic and in
combination.
The choice of two left-leaning newspapers also represents an ethical concern as well as a
limitation in terms of the validity of the research. My conclusions should therefore be
considered in that light and as only applicable to these two newspapers and their readers.
6- Theory and Previous Research
According to a reading of Bourdieu (1977) journalism is governed by a set of standards
and ethics that have garnered universal acceptance and he has defined these principles
and behavioural norms as “doxa” (Bourdieu, 1977, pp.164 - 169).
According to the Ethical Journalism Network (2016) these principles could potentially be
broken down into “Truth and Accuracy,” “Independence,” “Fairness and Impartiality,”
“Humanity” and “Accountability.”
Page 23 of 65
The assumption is that if journalists adhere to these principles, they will earn and maintain
the trust of their readers in addition to maintaining a clear conscience for their own sake
if that is of value to them. The latter however is not the emphasis of this paper, and the
focus will be on the reader’s perspective as the profession evolves and as their own access
to information expands with the “affordances” - Gibson (1979) – that technology and the
globalization of knowledge offers them and affects the work of journalists.
Elliott (2008) discusses how these changes in technology have affected journalists’ work
and the application of journalistic norms and guidelines, with the rise of citizen journalism
and the 24-hour news cycle forcing journalists and media organizations to speed up their
processes and which in the end has blurred the lines between journalists and readers and
between the different categories of published material – (Elliott, 2008, pp.29-30)
“Traditional news was once easy to distinguish from other forms of mass
communication. The boundaries between news and opinion, news and
entertainment, news and advertising have softened for a variety of reasons,
but just whether a particular informational product ought to count as news,
reality, analysis, opinion, or parody is sometimes difficult to judge.” (Elliott,
2008, p.30)
“Journalists must choose among the many sides to provide focus for their
stories, but when they choose only two, to give a polarized either/or
perspective, they lose the nuances that citizens need to understand before they
can make educated decisions for self-governance.” (Elliott, 2008, p.35)
It is perhaps this choice alluded to by (Elliott, 2008, p.35) to implicitly but not explicitly
express any nuance that angers readers. Some may know exactly what is being left out of
a story but others may simply sense that not all sides are being given a voice.
When specific stakeholders seem to have greater access to media platforms for whatever
reason and that becomes apparent over a long period of time and across various outlets
readers and audiences feel that they are being intentionally deceived or at the very least
not being allowed to make up their own mind.
The availability and access of information allows citizens to develop their own idea of
what is really happening – even if it is not an entirely objective opinion – and Elliott
(2008) concludes that this is a very good reason for journalists to ensure that they at the
very least attempt to maintain objectivity and avoid being influenced by or beholden to
Page 24 of 65
“powerful individuals and groups that would manipulate them and the pull of public
opinion as well.” (Elliott, 2008, p. 37-38). It is their duty to offer their audiences and
readers the tools they need to make informed decisions but not to make these informed
decisions for them.
Though reader comments may suggest that there is a longing to return to more ethical
principles of journalism, it may perhaps be too late for that. With the technological
developments and the changes they have triggered, perhaps journalism as a profession
has evolved beyond the concept of objectivity. If everyone has access to background
information, perhaps journalists now choose to be controversial or biased to attract an
audience – either in agreement or disagreement. The audience itself may not necessarily
know exactly what it wants, just that it is not what they are reading. It may be a matter of
confirmation or audience bias, where they want what it is written to reflect their own
opinion and not actually be objective, and if it does not they will make sure that the
journalists and fellow readers know it.
Hearns-Branaman, (2016, p.105), however suggests that it is possible that the journalists
are not those focusing heavily on audience bias but their news organizations themselves
who choose to appeal to this perceived bias, possibly for business reasons or out of fear
of losing their “legitimacy.” It seems this legitimacy may no longer be founded on
objectivity.
Media and journalists have never been completely autonomous in their role as
“gatekeepers” but the forces that have influenced them throughout their history have
always existed as discussed by Shoemaker and Vos (2009) but now the audiences and
readers themselves have an influence on what is published and what is not. The question
is whether the audiences can change the newspapers’ agendas or angles - if there are any
– or force the coverage of topics otherwise ignored.
As described in by Craft, Vos and Wolfgang (2015, p.4) there have been a number of
studies that discuss whether the influence of audiences will have an impact on the work
of journalists, potentially leading to more trivial news receiving greater emphasis while
more important news may be ignored to satisfy the seemingly shallow interests of readers
or not.
Page 25 of 65
This paper’s results however imply that it may not be about focusing on less sophisticated
news sought by readers and audiences – or at least not only – but also that it may be about
the manner in which important news is covered.
Another interesting aspect is perhaps whether affecting news coverage is in fact a
motivation for readers commenting on articles or just a potential side-effect or “perk,” or
of no interest at all.
Readers are no longer passive receivers of news but are actively involved in disseminating
it. Commenting on articles is only one form of audience participation that is constantly
expanding and evolving and with it the field of journalism as well is being constantly
reshaped. With members of the public producing their own news and expressing their
unfiltered opinions on blogs for example, reader comments may or may not have a
significant influence in the grand scheme of things, but they are part of a whole system
in which the public’s input is growing and it is worth analysing their motivation and their
potential impact.
According to Domingo and Heinonen (2008, p.4, p.7) a few journalists have themselves
chosen to enter into a more direct form of communication with their readers and
audiences. They have used blogs to more freely express their own privately-held opinions.
This has not pleased all of their respective employers, who perhaps did not wish for these
blogs to influence how objective their journalists truly appeared to be or perhaps they
feared that an expression of a completely different opinion may undermine a specific
political stance that their organization was committed to upholding.
Why do readers comment on news articles?
There are not only different types but also different levels of what is considered reader or
audience participation in or interaction with online news. The first step is to read a news
report, but readers are also welcomed – not always as explained above – to add their
thoughts on what they read in designated comment sections. Sharing particular news
stories to their social network is also considered a form of audience participation
according to Tenenboim and Cohen (2013, p.200)
Page 26 of 65
Tenenboim and Cohen (2013, p.212) analysed exactly what kind of news reports attracted
the greatest amount of participation, particularly in terms of comments posted and
concluded that “a substantial number of the most commented-upon items had elements
that may have aroused controversy. These include references to social/political conflicts,
social cleavages, negative or provocative messages and to prominent figures.”
“…items that relate to social/political conflict may be a suitable platform for
users who want to voice their opinions on matters of public concern, as well
as to try to influence others…The motivation to post comments may be high
especially with regard to social/political conflict, in which public opinion or
perceptions about it seem to play a more significant role than in interpersonal
conflict.” (Tenenboim and Cohen, 2013, p. 213 -214)
Commenting on articles therefore seems to be an expression of existing beliefs, an
opinion built up over time – perhaps flexible or not rigidly-held and perhaps not – but
there is a certain investment of a political, ethical or even personal nature involved in
taking the time and making the effort to comment and even more so to share links that
complement or contradict the articles in question.
Bourdieu (1990)’s definition of habitus functions well as a foundation from which to
understand the motivation behind reader comments as a way of expressing the
aforementioned investment is specific topics, even if those commenting are not
necessarily aware of the fact: Habitus is:
"…systems of durable, transposable dispositions, structured structures
predisposed to function as structuring structures, that is, as principles which
generate and organize practices and representations that can be objectively
adapted to their outcomes without presupposing a conscious aiming at ends or
an express mastery of the operations necessary in order to attain them.”
(Bourdieu, 1990, p.53)
Bourdieu (1990) explains that early childhood experiences and knowledge and expertise
gained over the course of an individual’s life and the actions they have taken “predispose”
them to certain attitudes and opinions or “dispositions” that make up their aforementioned
habitus. They act accordingly and not necessarily fully consciously with clear aims and
goals.
Page 27 of 65
If one were to apply this viewpoint to reader comments and links it would imply that
readers share these links and their opinions merely to express them, simply because they
hold them or are reacting to what they are reading but not that they intend to influence
the article author’s or the newspaper’s general stance, but Bourdieu (1990, p.53) qualifies
his definition by adding that the unconscious reactions of the habitus may also be
“accompanied by a strategic calculation tending to perform in a conscious mode the
operation that habitus performs quite differently.”
Another interesting aspect to consider is whether reader comment sections are a mere
form of public sphere (Habermas, Lennox and Lennox (1964) where those who wish to
do so can discuss their opinions with fellow readers, or are the authors the real addressees?
Here the question that raises itself of course is whether readers who comment on articles
and specifically those who post links are truly representative of a wider public audience
or very much a niche or a microcosm of an elite who care enough, are educated enough
and have enough time to contribute to reader comment sections.
McCluskey and Hmielowski (2011, p.304) believe that online reader comment sections
“offer the potential for a range of opinions that more closely matches ideals of the public
sphere.” If McCluskey and Hmielowski (2011, p.307) are correct in their belief that “news
organizations typically pledge traditional news values of objectivity and balance,
potentially attracting a broader range of public views” then it stands to reason that their
comment sections do so as well, therefore creating a greater balance and in my view
representing a more accurate picture of where public opinion truly stands, adding
viewpoints not covered by the articles’ authors, whether for lack of space or interest or
with a clear intention to portray issues in a certain light.
There is also a possibility that active audiences as defined and described by Livingstone
(2015) do not see their role as adversarial in the manner described by Craft, Vos and
Wolfgang (2015, p.4) or even as antagonistic but more as complementary. They offer the
sources of information they have found and even disagree with the authors out of the need
to help them develop a more balanced picture of whatever issue or news item they are
covering.
Page 28 of 65
Perhaps the most important thing about reader comments is that other readers think they
are “interesting”, as was established by Bergström and Wadbring (2015, p.147). Though
the target of readers when posting their comments and links may not be their fellow
readers, or at least not only, it is what drew me to the topic for this research in the first
place, as a reader myself and I arrived at the same conclusion that the comments are truly
more interesting than the articles themselves, at least most of the time.
A reading of Macgilchrist, 2011, suggests that the seemingly one-sided coverage,
particularly of controversial political topics - and particularly those involving Russia in
some way - is not coincidental but that…
“The everyday work of writing and reading the news is part of a complex and
multifaceted set of practices articulating hegemonic relations of power.
Domestic and international news, in this sense, can say as much about the
reporters and their social, political, cultural, historic, practical and technical
contexts as about the reported”. (Macgilchrist, 2011, p.xi)
In my view, readers of the Zeit and Guardian at least seem to be aware of this fact, if not
necessarily of its nuances and it is this that they reject most of all.
Macgilchrist, (2011, p.xi) explains that journalism is significantly influential when it
comes to politics and public-opinion-building, providing readers with a platform that
allows them to take active roles in the democracy of their nations. Their role should not
be as gatekeepers but as a failsafe mechanism against the main sources of power in any
country, who are not working on their behalf, are working against their people or even
aiding in manipulating public opinion.
In many of the links and comments shared in the articles covered there was implicit or
explicit criticism of a seeming deception by the media on behalf of different, special
interest groups, whether governmental or business-related or otherwise, creating a general
suspicion of journalists, with readers perhaps developing a defense mechanism of
doubting all they read first of all, until proven otherwise.
Page 29 of 65
In this context it is important to emphasize that doxa plays a significant role in creating –
and - by nature of the matter - in losing cultural capital for journalists. “Journalistic capital
can be understood as the specific, cultural capital of the journalistic field” (Schultz, 2007
p.194).
If trust and capital are eroded and readers or audiences no longer see doxa being applied,
then journalists’ and journalism’s cultural capital will shrink and the damage may take a
long time and great effort to reverse.
Livingstone (2015, p.440) explains the contradictory attitude towards audiences and
readers when it comes to politics. They are not viewed as knowledgeable or as having the
ability to influence others and/or results, but they are still courted by politicians using the
media as conduits. They are seen as one mass of consumers who should be guided and
educated but who have no real mind of their own, capable of seeing through these
transparent attempts to influence them and unaware of both their own interests and the
interests of those who seek to manipulate them.
Livingstone (2015, p.442) believes that despite this, to deny the power of the media,
particularly large “media conglomerates” would be foolish.
It is merely important to be aware of this potential to influence and to be aware of who
exactly is pulling the strings, with the public -or at least some members of it - slowly but
surely able to see the “puppeteers.”
If journalists were to seek to perhaps reverse this trend of falling trust and to capitalize
on the input of readers who comment and share links, instead of blaming them for ever
having put their faith in the concept of objective journalism, they could do worse than to
actually…
“Agree[ing] with some of a potential reader’s assumptions and knowledges,
texts which reframe mainstream stories provide a subject position of a
discerning, intelligent individual…suggesting that not only specialists or
reporters are capable of inspecting more closely, but also the readers
themselves if they have appropriate information. This contrasts with some
other arguments which position the reader as naïve or gullible for having
believed the mainstream view in the first place.” (Macgilchrist, 2011, p.206)
Page 30 of 65
Some of the data discussed below shows that while the relationship between readers and
journalists seems to be strained it has not been irredeemably damaged and more
transparency and an appreciation of readers’ intelligence and capacity of finding their
own alternative sources of information can rebuild trust and return the profession of
journalism to its ethical roots and principles.
7- Findings
7.1 - Quantitative Data Analysis:
Although the data collected provided at least some tentative answers to my questions, it
often raised a number of others, and I have chosen to include some of these questions in
the analysis below to perhaps point out the directions that further research on the topic
could take.
7.1.1 - Comments and Links – Totals and Ratios
Comments and Links -Totals and Ratios
Comments Total Links Total % of Links to Comments Guardian Refugees 11622 411 3.54 Guardian AI 5302 223 4.20 Guardian Elections 5034 180 3.58 Guardian Syria 5526 392 7.09 Guardian Total 27484 1206 4.39% Zeit Refugees 3165 213 6.23 Zeit AI 464 34 7.33 Zeit Elections 1456 106 7.28 Zeit Syria 2290 301 13.14 Zeit Total 7375 654 8.9% Total: 34859 1860 5.3%
Page 31 of 65
As can be seen in Table.1, the total number of comments for all 80 articles analysed in
this research was 34,859, with 27484 of those retrieved from the 40 Guardian articles and
the remaining 7,375 comments originating from the 40 Zeit articles. This is quite a
significant difference, with the Guardian articles nearly quadruple those of the Zeit. It is
of course clear that the Guardian can draw on a wider audience of English speaking
readers. The newspaper also has a longer history than the Zeit. Both these factors could
influence the number of visitors and therefore the number of comments that are left on
each article.
For both newspapers there were significantly higher comment totals for the articles
discussing the Syria conflict, which confirms Tenenboim and Cohen, (2013)’s findings
that controversial – particularly political topics seem to attract more comments.
What is also perhaps worth noting here is that all 10 articles randomly selected from the
Zeit newspaper on the Refugee crisis topic all appeared as general newspaper reports
without individual authors, as did 8 of those on the Syria conflict and the articles covering
the more controversial aspects of the election topic, such as the municipal elections in
Saudi Arabia and the parliamentary elections in Iran.
It seems highly unlikely that it is a coincidence that these controversial topics and
sub-topics were those that had no identifiable journalist responsible, while the
non-controversial topics generally did.
It would be interesting to analyze this matter in more
detail to determine the reasons and who makes the decisions in this regard.
Whyatt (2016) discusses the abuse experienced by journalists covering far-right protests
in Germany, and perhaps this is symptomatic of abuse journalists experience when
covering controversial topics in general, with some even fearing for their safety. This is
a possible explanation for a lack of identifiable authors, but that would imply that media
organizations or journalists are submitting to the abuse in a certain manner, as
understandable as that may be on a human level. There is of course also a possibility that
the newspaper has fewer resources on the ground across the world and must rely on
agency-generated reports, however that cannot be the case for an article covering the
German regional elections which also had no author acknowledged, but could be seen as
controversial due to the rise of the AfD.
Page 32 of 65
Link Totals by Newspaper
Guardian Zeit Link per Newspaper 1206 /1860 654 /1860
Percentage of Total 64.8% 35.2%
Table 2 - Link Totals by Newspaper
Table 2 shows that when it comes to actual links posted the difference is diminished, with
the Guardian total of 1206 links, representing a 64.8% of the total links analyzed in this
research and around double the amount of links posted to Zeit at 654 comments with links
at 35.2%. This means that an average of 8.9% of Zeit articles included a link (this is
described as an average as on occasion there are comments that included multiple links)
while only an average of 4.39% of Guardian reader comments included links. This is true
for all four topics analyzed, however there is one clear exception in terms of link to
comment ratio within these four topics, with both Zeit and Guardian article to link ratio
on the Syrian conflict much higher than for all three other topics, with almost double the
figure.
Throughout the analysis of this data it became clear that the numbers related to the Syrian
conflict stand out compared to the rest. The topic has been more controversial and more
emotion- and value-laden and therefore perhaps it has inspired more participation and
more contribution, but particularly more sharing of links that reinforce a given opinion or
provide access to a differing viewpoint.
Though Tenenboim and Cohen (2013) do not specifically discuss links, it stands to reason
that posting links can be viewed as even greater participation and once more the Syrian
conflict inspires a more “hands-on” approach in readers.
Page 33 of 65
7.1.2 - Language
Language Distribution and Frequency of Links
English German Other (English+) None/ Unavailable Guardian Refugees 387* 13* 6 (3) 2 Guardian AI 218 - - 6 Guardian Elections 140 18 18 (1) 4 Guardian Syria 376 4 9 (1) 3 Guardian Total 1121 /1206 (93%) 35 /1206 (2.9%) 33 /1206 (2.7%) 15 /1206 Zeit Refugees 9 204 - - Zeit AI 13 20 - 1 Zeit Elections 14 88 2 2 Zeit Syria 103 185 5 7 Zeit Total 139 /654 (21.3%) 497 /654 (76%) 7 /654 (1.1%) 10 /654 Total: 1260 /1860 (67.7%) 532 /1860 (28.6%) 40 /1860 (2.2%) 25 /1860
Table 3 - Language Distribution and Frequency of Links