ArticlePDF Available

Transferring control from the backend to the frontend: A comparison of the discourse architectures of comment sections on news websites across the post-Soviet world

Authors:

Abstract and Figures

This study compares how comment sections (CSs) were implemented, as of summer 2016, on the 179 leading national news websites across the 15 post-Soviet countries. In order to pursue this aim, a novel coding scheme is developed that facilitates assessment of the degree to which the discourse architectures of CSs transfer control over the content published from the backend to the frontend of a website. Accordingly, each CS is assigned a value on a ‘control transfer index’ (CTI). The study identifies the level of press freedom/democracy of a country as the only significant predictor for whether, and how openly, CSs were implemented. The popularity of CSs and their CTIs decreased with decreasing levels of press freedom/democracy. However, even in the most closed regimes, CSs were still a relatively commonly observed phenomenon. We interpret this latter finding by drawing on theories of citizen participation under authoritarian rule.
Content may be subject to copyright.
Transferring control from the backend to the frontend:
A comparison of the discourse architectures of
comment sections on news websites across the post-
Soviet world
This is the accepted version of an article forthcoming in the journal New Media &
Society. The final, definitive version of this paper will be available at
http://journals.sagepub.com/home/nms (SAGE Publications Ltd, All rights reserved. ©
Florian Toepfl & Anna Litvinenko)
Florian Toepfl
Free University of Berlin, Germany
Anna Litvinenko
Free University of Berlin, Germany
Corresponding author: Florian Toepfl, Institute for Media and Communication Studies,
Free University of Berlin, Garystrasse 55, 14195 Berlin, Germany. Email: f.toepfl@fu-
berlin.de
Acknowledgements
This research was supported by an Emmy Noether Fellowship, sponsored by the German
Research Foundation DFG (awarded to Florian Toepfl). We owe deep thanks to Daria
Kravets and 14 country experts for their invaluable help in collecting the data for this
project, as well as to Andrei Zavadski for his insightful comments, which greatly helped to
develop the coding scheme.
Running head: TRANSFERRING CONTROL 2
Transferring control from the backend to the frontend: A comparison of the
discourse architectures of comment sections on news websites across the
post-Soviet world
Abstract
This study compares how comment sections were implemented, as of summer 2016, on the
179 leading national news websites across the 15 post-Soviet countries. In order to pursue
this aim, a novel coding scheme is developed that facilitates assessment of the degree to
which the discourse architectures of comment sections transfer control over the content
published from the backend to the frontend of a website. Accordingly, each comment
section is assigned a value on a control transfer index (CTI). The study identifies the level
of press freedom/democracy of a country as the only significant predictor for whether, and
how openly, comment sections were implemented. The popularity of comment sections
and their CTIs decreased with the openness of a regime. However, even in the most closed
regimes, comment sections were still a relatively commonly observed phenomenon. We
interpret this latter finding by drawing on theories of citizen participation under
authoritarian rule.
Keywords: comparative communication research, political communication, comment
sections, authoritarian institutions, participation, content analysis, press freedom,
participatory journalism, authoritarianism
Running head: TRANSFERRING CONTROL 3
With the rise of the Internet over the past two decades, news websites around the
globe have implemented on their platforms a broad range of features aimed at facilitating
their audiences’ participation in publishing content. Among the most frequently adopted
features have been readers’ blogs, reader polls, tools to share news via social networks, and
comment sections beneath news articles (De Keyser and Sehl, 2011; Thurman, 2008;
Netzer et al., 2014). Comparing the adoption of these features across national contexts,
scholars have deployed a range of theoretical and methodological approaches (De Keyser
and Sehl, 2011; Suau and Masip, 2014; Himelboim and McCreery, 2012; Netzer et al.,
2014). Extant cross-national comparative research, however, has pursued largely
exploratory goals and arrived partly at contradictory conclusions. For instance, with regard
to the question of whether the level of economic development in a country affected the
adoption of participatory features by news websites, Himelboim and McCreery’s (2012)
exploratory study indicated a positive correlation. By contrast, Bachmann and Harlow
(2012) were not able to provide empirical evidence of a potential association of these two
variables. In the present study, we aim to build upon, and advance, this strand of literature
by sidestepping what we see as three key methodological weaknesses of extant cross-
national comparative research on the implementation of participatory features on news
websites.
Firstly, previous studies have typically drawn on either very small samples or
convenience samples of countries (De Keyser and Sehl, 2011; Himelboim and McCreery,
2012; Humprecht and Esser, 2016; Suau and Masip, 2014; for an exception in this regard,
consider Bachmann and Harlow, 2012). By contrast, this study investigates all the
countries within an entire socio-political region, the post-Soviet world. At the most abstract
Running head: TRANSFERRING CONTROL 4
level, the study thus adopts a most similar systems design, akin to comparative
communication research that has focused on member countries of the European Union or
social science research that has focused on member countries of the OECD (Organization
for Economic Co-operation and Development; see Berk et al., 1995). The 15 countries
included in this study share, most importantly, a common history of at least 45 years of
Soviet communist rule, which ended in 1991, as well as the Russian language as a still
widely used lingua franca. By contrast, they show great differences with regard to key
explanatory factors tested in this study, that is, their current levels of press freedom,
democratic governance, Internet penetration, and socio-economic development. The logic
of this type of research design, which is referred to as regional statistical analysis in
comparative politics, is to initially make a generalisation only about that one region, and,
if successful there to extend that analysis later to be a proposition about politics [or
political communication] more generally (Peters, 1998: 18).
Secondly, extant research has been based on very small samples of news websites
within each country, typically considering not more than one or two news outlets per
country (Bachmann and Harlow, 2012; Singer et al., 2011; Suau and Masip, 2014; for an
exception in this regard, consider Humprecht and Esser, 2016). By contrast, this study
seeks to consider a census of the leading national news websites in each of the 15 post-
Soviet countries. Thirdly, previous research has deployed rather coarse-grained coding
schemes, typically aimed at capturing only the presence or absence of a wide range of
participatory features incorporated into news websites, from the provision of contact email
addresses to multiple-choice opinion polls (see, for instance, Domingo et al., 2008; Suau
and Masip, 2014; Himelboim and McCreery, 2012). By contrast, this study places under
Running head: TRANSFERRING CONTROL 5
scrutiny only one such participatory feature, albeit one of the most widely adopted and
most popular around the globe: the comment sections linked to individual news articles
(Reich, 2011; Singer, 2014; for a discussion of the recent loss of popularity of this feature
in Sweden, consider Karlsson et al., 2015). Focusing on only one participatory feature, this
study is able to develop, and apply, a relatively fine-grained coding scheme. The latter has
been designed to facilitate answering the following two overarching research questions: To
what degree do comment sections on news websites transfer control over the content
published from those working at the backend (media professionals) to those entering data
at the frontend? And, secondly, what are the key factors at the level of (a) the news
organization and (b) the country that predict whether, and how, comment sections are
implemented?
Accordingly, the research goals pursued are partly descriptive (providing an
overview of how comment sections were implemented, as of summer 2016, across the
post-Soviet world), partly exploratory (developing a coding tool and hypotheses that can
be deployed in future research in or across other regions), and partly explanatory
(identifying factors that predict whether and how comment sections were implemented). In
order to pursue these goals, the remainder of the article is structured as follows. In the first
section, we develop a coding scheme that facilitates the ranking of the discourse
architectures of comment sections according to the degree of control that is transferred
from the backend to the frontend of the website, that is, according to what we establish as
the control transfer index (CTI) of a comment section. Subsequently, we derive from
extant literature two hypotheses and four research questions. In the third section, we
provide an overview of the methods adopted, while we dedicate a fourth section to the
Running head: TRANSFERRING CONTROL 6
presentation of results. As one of the study’s key findings, the level of press
freedom/democracy is identified as the only significant predictor for how likely a website
was to implement a commenting feature, as well as for how openly the discourse
architecture of a comment section was designed. By contrast, state ownership, parent
medium, and the political orientation of a news website did not show to play a significant
role. In a concluding section, we discuss how these and other findings of the study advance
extant comparative research on the adoption of participatory features by news websites, as
well as research on citizen participation under authoritarian rule.
The discourse architectures of comment sections: Devising a control
transfer index
As Singer (2014) has argued, many of the recent changes of the news environment
can be understood as occurring along a continuum of journalistic control over content
decisions (Singer, 2014: 67; see also Ihlebæk and Krumsvik, 2015; Scott et al., 2015). At
one end of this continuum is the traditional process of news production, where journalists
make virtually all decisions concerning the content published on a mass media website. At
the other end of the spectrum is a news environment in which users make all the
decisions (Singer, 2014: 67). Drawing upon this conceptualization of a continuum of
journalistic control, in this study, we seek to investigate the degree to which differently
implemented comment sections transfer control over what content is allowed [...] to enter
public circulation (Singer, 2014: 56) from those working at the backend of a news website
(backenders) to those entering comments at the frontend (frontenders). We deploy the
terms backenders and frontenders in order to refer to the two basic types of social roles
Running head: TRANSFERRING CONTROL 7
that commenting features on news websites as technological artifacts commonly create.
The role of backenders is typically filled by media professionals, including journalists,
editors, or outsourced content management teams. The role of frontenders is expected,
from a normative perspective, to be taken on by ordinary citizens who enter comments to
participate in the public life of a society by expressing their genuine political will. The
frontender role may, however, also be exploited by automated bots or organized collectives
of commenters sponsored by governments or corporate interests that seek to manipulate
public opinion by simulating genuine political participation (Woolley and Howard, 2016).
In order to empirically capture the degree to which control over the content published
on a news website is transferred from the backend to the frontend, we deploy the concept
of discourse architecture (Freelon, 2015: 772; see also Jannsen and Kies, 2005). By
discourse architecture, we mean the socio-technological setting within which backenders
and frontenders interact on a news platform: the technical and organizational architecture
of the discussion space (Jannsen and Kies, 2005: 321). Technical characteristics of the
discourse architecture of a comment section include, for instance, whether frontenders
have the opportunity to like, rate, or reply to comments. Important organizational
characteristics are, for example, whether backenders routinely check all comments before
publication (pre-moderation), and whether all or only selected articles are typically opened
up for commenting.
At this point, it is important to emphasize that this study does not subscribe to
technological determinism: We do not assume that specific discourse architectures will
deterministically result in specific discursive patterns emerging in comment sections. By
contrast, following a line of argument widely adopted in the literature (Jannsen and Kies,
Running head: TRANSFERRING CONTROL 8
2005; Freelon, 2011, 2015; Wright and Street, 2007), we presume that discourse
architectures render certain social outcomes more likely than others (Freelon, 2011: 201),
because they assign the roles of backenders and frontenders with different sets of
affordances and constraints. Affordances, then, are understood as communicative actions
that a platform enables or makes easier to pursue. Constraints, by contrast, refer to
communicative actions that a discourse architecture precludes or makes more difficult to
pursue (Freelon, 2011, 2015). To illustrate these lines of thinking with an example, a
discourse architecture that features a pre-moderation regime affords backenders the
checking of every single comment before it is published. A move from pre- to post-
moderation of comments suspends this affordance, and can thus be interpreted as a
significant step away from journalistic [backend] control and toward user [frontend]
control of material published under an organization’s aegis’ (Singer, 2014: 60; for a similar
line of argument, see also, Ihlebæk and Krumsvik, 2015).
In this study, we refer to closed discourse architectures as those architectures that
furnish frontenders with relatively few affordances and many constraints. By contrast, we
call open architectures those that endow frontenders with relatively many affordances
and few constraints. Not subscribing to technological determinism, we associate neither
open nor closed discourse architectures per se with beneficial or detrimental political
outcomes. Openly designed comment sections may, for instance, provide a range of
opportunities to marginalized social groups for engaging in counterpublicity (Toepfl and
Piwoni, 2015). On the other hand, they are also more easily exploited by bots or collectives
of commenters acting on behalf of authoritarian elites or large businesses, covertly
promoting their sponsors’ interests (Woolley and Howard, 2016). In both instances, the
Running head: TRANSFERRING CONTROL 9
content of the comments that ultimately appear on the news websites will depend greatly
on the degree to which backenders leverage their affordances in order to manage, and ban,
user-generated content (Jönsson and Örnebring, 2011). In addition, the comments
published on specific topics will also depend on a range of further context factors,
including the nature of a country’s civil society, the legal environment, the political
system, and historical experiences.
While discourse architectures of comment sections thus cannot be linked
deterministically to specific social outcomes, such as the empowerment of citizens, they
are nonetheless highly important subjects of analysis for at least two reasons. Firstly, in a
metaphorical sense, discourse architectures of news websites constitute the terrain on
which highly visible discursive struggles are staged between different groups of
frontenders (including some that act through sponsored collectives of commenters or bots),
as well as between frontenders and backenders (Toepfl and Piwoni, 2015). In these
struggles, discourse architectures determine not only the tools that frontenders and
backenders have at their disposal (affordances), they also limit the range of actions that
they can pursue (constraints). Being aware of differences between discourse architectures
of comment sections can thus be regarded a prerequisite to any in-depth, comparative study
of the discursive struggles that evolve in these online spaces. Secondly, discourse
architectures of news websites can be interpreted as indicators of the basic approach to
audience participation that a news organization seeks to convey to the wider public (Toepfl
and Piwoni, 2015). Along these lines, Jönsen and Örnbebring (2011), for instance, have
argued that Western news organizations frequently highlight user-generated content as a
democratic tool for purposes of self-legitimization and branding (141). In a similar
Running head: TRANSFERRING CONTROL 10
vein, Karlsson and colleagues (2015) have emphasized that studying the implementation of
participatory features makes it possible to see inclinations (297) towards audience
participation on the behalf of producers. Similarly, we assume in this study that a cross-
national comparison of discourse architectures facilitates the revealing of differences in the
basic approach to audience participation that news organizations in different political
contexts seek to convey to their audiences.
Adopting this framework of thinking, the present study considers the discourse
architecture of a comment section as transferring more control over content decisions from
backenders to frontenders:
(1) the more visible the comment section is on the website,
(2) the more affordances it offers to frontenders, and
(3) the fewer affordances it offers to backenders.
It is along these three dimensions that we outline, in the Methods section, a coding
scheme that facilitates the calculation of the control transfer index (CTI) of a comment
section.
Developing hypotheses: Predicting the discourse architectures of comment
sections
Grounded in relatively small convenience samples, extant cross-national comparative
research has only been able to provide tentative answers to the question of whether factors
can be identified that predict whether news websites implement specific participatory
features. Potentially influential factors have been discussed at the two levels of (1) the
news organization and (b) the country. At the country level, for instance, extant research
Running head: TRANSFERRING CONTROL 11
has found no indication that higher levels of press freedom are correlated with a more
frequent adoption of participatory features (Bachmann and Harlow, 2012; Himelboim and
McCreery, 2012). With regard to the impact of regime type, only Suau and Masip (2014)
have cautiously concluded that in countries governed by democratic systems, the
participatory options do tend to be greater than in countries with autocratic regimes (682).
A positive impact of the level of economic development of a country has likewise only
been adumbrated by one study (Himelboim and McCreery, 2012; with Bachmann and
Harlow, 2012, finding no correlation). Similarly, only one study has indicated a positive
impact of Internet penetration (Bachmann and Harlow, 2012). Based on these cautious
conclusions of the extant exploratory research, we hypothesize in this study that the
discourse architectures of news websites across the post-Soviet region will transfer the
more control to the frontend of comment sections, the higher a country ranks on press
freedom (H1), the more democratic the political regime of a country (H2), the higher the
Internet penetration rate in a country (H3), and the higher the level of economic
development of a country (H4).
At the second level of the news organization, extant cross-national research offers
only scarce hints at factors that could be of influence. Himelboim and McCreery’s (2012)
data, most importantly, indicates that the type of parent medium of a news website affected
the type of participatory features implemented: While newspapers favoured text-based
features, broadcast media preferred visual features. As Himelboim and McCreery (2012)
argued, Both industries were more likely to adopt features that are more native to their
traditional practices (439). By contrast, Humprecht and Esser (2016) found no support for
their hypothesis that Web-based outlets would include forums and chats most frequently
Running head: TRANSFERRING CONTROL 12
because those outlets seek to be an alternative to mainstream media and highly value user
participation (12). While cross-national research has thus interrogated only the type of
parent medium as a predictor variable at the organizational level, additional clues may be
taken from research conducted within single countries. Two additional factors at the news
organization level, discussed frequently within single-country studies, include the type of
ownership (Ihlebæk and Krumsvik, 2015; Singer, 2014; see also Humprecht and Esser,
2016) and the political orientation of a news website (El Gody, 2015). Given this scarcity
of extant knowledge, we decided to formulate four research questions with regard to
factors at the news organization level: Are news websites more or less likely to transfer
control to commenters that are stand-alone online media (RQ1), that are critical of the
political leader of a country (RQ2), that are not state-owned (RQ3), or that are owned by
foreign corporations (RQ4)?
Methods
Key challenges that this study had to face - just like previous research adopting a
similar design - were constraints related to analysing websites in [a multiplicity of] native
languages (Himelboim and McCreery, 2012: 440). In previous research, these challenges
frequently resulted in the adoption of convenience sampling strategies (Himelboim and
McCreery, 2012) and the absence of intercoder reliability tests (Suau and Masip, 2014).
We addressed these difficulties by hiring at least one expert coder for each country, who
supported a team of two main coders (one author of this study, one research assistant) as
assistant coders in the process of collecting and analysing the data. A list of short
biographies of all the experts involved, a detailed description of the coding procedure, the
Running head: TRANSFERRING CONTROL 13
codebook, as well as additional descriptive statistics of the data set are provided as online
supplementary files (Codebook / supplementary file).
The unit of analysis: Identifying leading national news websites
In order to identify the opinion-leading national news websites in each country, we
started out with a list of the 500 top websites for each country, as provided by the company
Alexa.com, ordered by the websites’ one-month traffic from within each respective
country. We went through these lists from top to bottom with the expert coder for each
country in order to identify the 20 news websites most visited from within each country.
In this list of news websites, we explicitly have not included regional or local news sites,
business and sport news websites, or search engines that feature small bits of news on their
homepages. From the resulting lists of the twenty most visited news websites within each
country, we then deleted all news websites whose main editorial offices were not located
inside the country (with the exception of those that were forced to operate from abroad due
to political pressure, like charter97.by in Belarus). We considered these 15 lists the census
of leading national news websites in each country, even though they frequently contained
considerably fewer than 20 items. In the case of Estonia, for instance, only six domestic
news outlets remained on the list, because all the remaining top-visited news sites operated
from either Russia or Western countries. Based on this selection process, we identified 179
leading national news websites across the 15 post-Soviet countries.
Running head: TRANSFERRING CONTROL 14
Measures: Outcome and predictor variables
Outcome variables. This analysis seeks to predict two outcome variables. The first is
the binary variable of whether a news website features a comment section or not (CS). The
second is the score that a comment section achieves on the control transfer index (CTI)
proposed in this article. In the course of a series of pre-tests, and drawing on extant
research on audience participation in news websites (Ihlebæk and Krumsvik, 2015; Singer,
2014; Scott et al., 2015), we devised the CTI as being comprised of the following three
dimensions (for details, see the full codebook provided as an online supplementary file):
(1) The visibility of a comment section: The more visible the commenting feature on
a news website, the more control is transferred to frontenders. Within this
dimension, we coded, for instance, whether the website had a section that
featured the most commented-upon stories, and whether comments appeared
immediately below the news article or had to be accessed on an additional page
(4 items, 0-10 points).
(2) Opportunities for frontenders: The more opportunities a comment section offers
to frontenders, the more control is transferred to the frontend. We divided this
dimension into three subdimensions concerning the opportunities for frontenders
(a) to conceal their identities, (b) to create content, and (c) to influence the
presentation of content. Within this dimension, we established, for example,
whether frontenders had the opportunity to comment without registering,
whether they could embed photos or videos, and whether they could like, sort, or
report comments (8 items, 0-10 points).
Running head: TRANSFERRING CONTROL 15
(3) Opportunities for backenders: The fewer opportunities the discourse architecture
of a comment section offers for media professionals to influence the content
published, the more control is transferred to the frontend. Within this dimension,
we coded, for instance, whether journalists closed comment sections at the
bottoms of controversial articles, and whether comments were checked by a
journalist before publication (3 items, 0-10 points).
The CTI is thus calculated based on 15 items, with options for each item to be
assigned between -1 and 5 points. Points are summed up to obtain the overall value of the
CTI, which can thus take on values between 0 and 30 points. At the most abstract level, the
meaning of the CTI is that, the greater the number of points that a comment section scores,
the more control its discourse architecture transfers from the backend to the frontend.
While the calculation and the composition of the index may certainly be refined in future
research, we have sought to guarantee the validity of the measure in primarily two ways.
Firstly, we aimed to maximize the index’s ‘content validity and face validity
(Neuendorf, 2002: 115-117) by repeatedly adding and dropping differently worded items
over the course of several rounds of preliminary coding across national contexts. These
incremental changes were accompanied by extensive discussions within the research team
as well as at academic workshops and with external country experts. Secondly, we checked
the index’s ‘construct validity by investigating whether the CTI was related to other
measures (constructs) in a way consistent with hypotheses derived from theory
(Neuendorf, 2002: 117; for details about these validity checks, see online supplementary
file).
Running head: TRANSFERRING CONTROL 16
Predictor variables. This study tests predictor variables at two levels, that of (1) the
news organization and that of (2) the country. At the first level of the news organization,
we decided to deploy relatively simple binary distinctions with regard to all four variables
coded, in order to achieve intercoder reliability across the 15 national contexts. With
regard to the political orientation of a website, for instance, we coded whether the everyday
news coverage of the outlet could feature criticism of the highest-ranking leader of the
respective country or not (critical website, 0: No, 1: Yes). With regard to the ownership
of a news website, we coded whether the state or a foreign corporation owned the majority
of shares in a website (state-owned, 0: No, 1: Yes; foreign-owned, 0: No, 1: Yes). With
regard to the type of parent medium of a website, we coded whether the news website was
a stand-alone online outlet or affiliated with a traditional parent medium (standalone online
news site, 0: No, 1: Yes). At the second level of the country, we used: as a measure of
press freedom, the country scores provided by the NGO Reporters Without Borders in their
2016 report (H1); as a measure of democracy, the country score provided by Freedom
House in the Freedom in the World Report 2016 (H2); as a measure of Internet penetration,
the World Bank’s data on Internet usage for the year 2015 (H3); and, as measure of the
economic development, the gross domestic product per capita provided by the World Bank
for the year 2015 (H4).
Coding and analysing the data: Securing reliability
The coding was conducted between 7 April and 9 August 2016 by the two main
coders (who both had mastery of three languages spoken in the post-Soviet region),
typically in a personal meeting of several hours with an assistant coder. In order to test for
Running head: TRANSFERRING CONTROL 17
intercoder reliability, the two main coders double-coded (assisted by two different experts
as assistant coders if they did not speak the language of the country) all websites from five
purposefully selected countries from across the geographical area that covered the full
range from fully democratic to entirely authoritarian regime types: Belarus, Estonia,
Kazakhstan, Russia, and Ukraine (n = 65 websites). More than 30% of the data set was
thus double-coded. The values achieved for Scott’s pi ranged between .81 and 1.0,
corresponding to agreement rates between 91% and 100%. The lowest, yet still
satisfactory, values were achieved for the variables of political orientation (91%
agreement, π = .81), ownership (95%, π = .88), and registration for commenting (92%, π =
.90). In order to analyse the data set, we used logistic and linear regression analysis. Even
though our data set can be viewed, from a statistician’s perspective, as the population of
leading news websites across the post-Soviet world, we report inferential statistics such as
p-values and confidence intervals at some points of our argument. We do so based on the
assumption, widely adopted in cross-national social science research, that our data set
represents an apparent population, in the sense that it can be regarded as only one random
‘‘realization’ of some set of social process that could have in principle produced a very
large number of other realizations (Berk et al., 1995: 423; see also Peters, 1998).
Findings
The following findings are based on what we consider the population of the leading
national news websites across the 15 successor states of the Soviet Union (N = 179), as of
summer 2016. The number of websites identified within each of the 15 countries ranged
from four (TJ) or five (UZ) in smaller and less developed countries to 18 (UA) or 20 (RU)
Running head: TRANSFERRING CONTROL 18
in larger and more developed countries (M = 11.33; SD = 5.31). Of the 179 news websites,
73.3% (132) had adopted a commenting feature. The 132 news websites offering a
commenting feature obtained on average 18.12 points on our CTI (SD = 3.87). For further
descriptive statistics of our data set, see online supplementary file.
Predicting the discourse architectures of news websites across the post-
Soviet world
In the theory part, we hypothesized that four factors at the country level would affect
how much control the discourse architectures of news websites transferred to commenters:
the levels of press freedom (H1), democracy (H2), Internet access (H3), and economic
development (H4). However, in testing whether our data set complied with the
assumptions of regression analysis, we found that two pairs of predictor variables were
highly correlated: the levels of press freedom and democracy, r(13) = .92, and the levels of
Internet penetration and economic development, r(13) = .70. Multicollinearity of
independent variables can seriously bias parameter values in regression analysis. In order
to address this issue, we combined the first pair of predictor variables into a new variable
that we termed press freedom/democracy (PFD) and the second pair of predictor
variables into a new variable that we labelled Internet/economic development (IED). In
the following analysis, we thus must limit ourselves to testing in combination H1 and H2
(press freedom/democracy) and H3 and H4 (Internet/economic development).
At the level of the news organization, we conjectured that four characteristics might
affect how the discourse architectures of websites are implemented (RQ1-RQ4). In the
following, we present two multiple regression analyses that evaluate how well these
Running head: TRANSFERRING CONTROL 19
variables predicted our two outcome variables: (1) the binary variable of whether a news
website featured a comment section or not (CS) and (2) the score that a comment section
achieved on our control transfer index (CTI).
Predicting whether a news website implements a comment section
Since the first outcome variable CS is binary, we deployed stepwise logistic
regression analysis. As extant research primarily points to the theoretical importance of
factors at the country level, we entered this set of predictors first into the regression (model
1). In a second step (model 2), we added three of the four predictors at the level of the
news organization. We were not able to include the fourth predictor foreign-owned (1/0,
RQ4), primarily due to the limited role that we found foreign ownership played in the
region. Only 12 out of 179 news websites in our sample (6.7%) were foreign-owned, and
all of these 12 foreign-owned websites operated comment sections. While this observation
clearly points to an association between foreign ownership and the adoption of comment
sections (RQ4), it was impossible to test this relationship in a multiple logistic regression
model, for reasons of incomplete information with regard to several combinations of
predictor variables. As the results presented in Table 1 show, the three remaining
predictors at the country level added only little explanatory power (-2LL change = 3.55, p
= .32). Likewise, the predictor variable IED did not significantly contribute to explaining
the outcome variable CS in either model. By contrast, PFD was positively associated with
whether news sites adopted a commenting feature. An increase in the level of PFD by one
standard deviation increased the likelihood of a news site featuring a comment section by
Running head: TRANSFERRING CONTROL 20
more than a factor of two in both models (Exp(b) in model 1 = 2.14; Exp(b) in model 2 =
2.12).
Table 1. Stepwise Logistic Regression: Predictors of Whether a News Site Implements a
Comment Section
Predictors
Model 1
B
SE
Exp(B)
p
B
SE
Exp(B)
p
Step 1: Country level variables
H1/2: Press freedom/democracy (PFD)
.762
.183
2.14
.000
.751
.249
2.12
.002
H3/4: Internet/economic development
(IED)
.217
.184
1.24
.233
.266
.187
1.31
.147
Step 2: News organization level variables
RQ1: Stand-alone online outlet (1/0)
.269
.406
1.31
.503
RQ2: Critical media outlet (1/0)
-.286
.518
.75
.566
RQ3: State-owned media outlet (1/0)
-.900
.615
.41
.104
Constant
1.183
0.179
3.27
.000
1.355
.440
3.876
.001
-2 log likelihood
187.17
-2 log likelihood change
18.94 (p < .001)
Nagelkerke R2
.147
Note: N news outlets = 179; Results for standard errors (SE) and ps are based on 2,000
bootstrap samples, stratified by country (N = 15); log likelihood of null model = 206.4
Running head: TRANSFERRING CONTROL 21
Predicting the control transfer index of comment sections
Our second outcome variable captured how openly the discourse architecture of
comment sections was designed (CTI). Treating the CTI as a ratio variable, we adopted a
stepwise linear regression analysis. For the reasons discussed above, we entered the two
predictor variables at the country level (PFD, IED) in the first step of the analysis (model
1) and added the four predictors at the news organization level only in the second step
(model 2). Table 2 summarizes the results. As in the previous logistic regression analysis,
the predictor variables at the news organization level, entered in model 2, added very little
explanatory value to model 1 (adjusted R2 change = -0.008). Likewise, the context variable
Internet/economic development (IED) was not significantly associated with the CTI of
websites in either of the two models. By contrast, both models indicate a positive relation
between the level of press freedom and democracy (PFD) of a country and the discourse
architecture of comment sections in that country. All other variables held constant (model
2), for instance, an increase in the level of PFD by one standard deviation resulted in a
predicted value 1.4 points higher for the CTI of a website.
Running head: TRANSFERRING CONTROL 22
Table 2. Stepwise Linear Regression: Predictors of the Control Transfer Index of a
Comment Section
Predictors
Model 1
Model 2
B
b
SE
p
B
b
SE
p
Step 1: Country level variables (z scores)
H1/2: Press freedom/democracy (PFD)
1.648
.422
.274
.000
1.344
.344
.459
.003
H3/4: Internet/economic development
(IED)
-.063
-.016
.284
.827
-.017
-.004
.328
.958
Step 2: News organization level variables
RQ1: Stand-alone online outlet (1/0)
-.516
-.067
.609
.390
RQ2: Critical media Outlet (1/0)
.329
.037
.926
.732
RQ3: State-owned media outlet (1/0)
-1.34
-.090
1.18
.234
RQ4: Foreign-owned media outlet (1/0)
-0.42
-0.03
0.73
Constant
17.82
.267
.000
18.054
.801
.957
Adjusted R2
.161
.146
F Change
F(2, 129) = 13.6
p < .001
F(4, 125) = .428
p = .788
Note: N = 132 news outlets (all news sites featuring comment sections); results for robust
standard errors (SE) and ps (2-tailed) are based on 2,000 bootstrap samples, stratified by
country (N = 15).
Running head: TRANSFERRING CONTROL 23
Taken together, the results of the two regression analyses lend support to H1/H2:
Across the post-Soviet region, higher levels of PFD were consistently associated with
more websites implementing comment sections, as well as with the discourse architectures
of comment sections transferring more control to users. Figure 1 visualizes the association
of the level of PFD with our two outcome variables for data aggregated to the country
level, positioning the 15 post-Soviet countries in two two-dimensional spaces. By contrast,
grounded in our data, we can reject H3/H4: In none of our models was the level of IED
significantly associated with the discourse architectures of comment sections on the
leading news websites of a country. Likewise, we gained little predictive power by
factoring in whether a news website was a stand-alone online outlet (RQ1), whether it
dared to criticize the country’s political leader (RQ2), and whether it was state-owned
(RQ3). Accordingly, we must answer RQ1, RQ2, and RQ3 in the negative. We were not
able to test the impact of foreign ownership (RQ4) on whether a website featured a
comment section (CS), due to the scarcity of the data available and the limited role of
foreign ownership in the region. However, we found no consequences of foreign
ownership for how openly a comment section was designed (CTI).
Running head: TRANSFERRING CONTROL 24
Figure 1. The Impact of Levels of Press Freedom/Democracy on the Discourse
Architectures of News Websites
Discussion
In this study, we have compared how, at the time of research in summer 2016,
comment sections as a participatory feature were implemented on the 179 leading national
news websites across the 15 post-Soviet states (descriptive research goal). In order to do
so, we have developed a novel coding scheme (exploratory research goal), grounded in
Singer’s (2014) conceptualization of a ‘continuum of journalistic control over content
Running head: TRANSFERRING CONTROL 25
decisions (67) and Freelon’s (2015) approach to ‘discourse architecture(s) (772). As the
empirical analysis demonstrated, the discourse architectures of comment sections varied
widely across the post-Soviet region, with average proportions of websites implementing
comment sections ranging from 29% in Azerbaijan to 100% in the Baltic States, and
average CTIs ranging from 12.3 in Tajikistan to 21.7 in Kyrgyzstan (see Figure 1).
Press freedom/democracy as the only significant predictor of discourse
architectures
Pursuing the third type of explanatory research goal, the study interrogated the power
of six variables, at the two levels of country and news organization, to predict whether and
how the 179 news websites implemented comment sections. Going beyond the primarily
exploratory and descriptive approaches of extant research (Himelboim and McCreery,
2012; Bachmann and Harlow, 2012; Suau and Masip, 2014), the study has been the first to
provide solid empirical evidence of a significant association between political regime type
(i.e., the level of PFD) and the degree to which a country’s leading news websites
implement a participatory feature. In addition to this key positive finding, the negative
findings of the analysis are highly intriguing. Grounded in solid empirical data, we could
reject the impact of a series of factors that are widely assumed to be of potential influence
in the literature (El Gody, 2015; Ihlebæk and Krumsvik, 2015; Humprecht and Esser,
2016; Singer, 2014). Controlling for levels of press freedom/democracy, for instance, our
data shows no significant consequences of the level of IED of a country (H3/H4). One
highly plausible explanation of this finding is that, at the time of research, comment
sections could be implemented with relatively limited technological knowledge and
Running head: TRANSFERRING CONTROL 26
material resources. For instance, in Kyrgyzstan, one of the least developed but relatively
politically open countries included in the study, 5 out of 10 websites featured both a
custom-made commenting tool and the standardized comments plugin provided by
Facebook. The discourse architecture of these comment sections thus offered frontenders
the choice between two differently structured commenting options, which resulted in
exceptionally high CTI values, making Kyrgyzstan’s comment sections, on average, the
most open in the region (see Figure 1).
Contrary to what we expected based on extant research, three predictor variables at
the news organization level showed no significant impact. Across the post-Soviet region,
while controlling for levels of press freedom/democracy and ICT/economic development,
we did not find any consequences of whether or not a news website was (1) state-owned,
(2) a stand-alone online outlet, or (3) critical of the country’s leader. A plausible
explanation for the lack of significance of these factors is that, with regard to each, a
multitude of causal mechanisms appeared to be at work, pushing in different directions
and, arguably, varying greatly in their combinations between the political contexts
included (with the latter ranging from fully democratic to fully authoritarian regime types).
To explicate but one example, critical media in authoritarian contexts may be less reluctant
to invite their readers to comment, as said media can be expected to be less averse than
regime-loyal outlets to a diversity of opinions appearing on their websites. Similarly,
critical media may be assumed to be more likely than regime-loyal media to imitate the
practices of news organizations in Western countries, where comment sections were a
widely popular participatory feature at the time of research (Reich, 2011; Singer, 2014).
Yet, at the same time, critical media in many of the authoritarian regimes included in the
Running head: TRANSFERRING CONTROL 27
study were subject to heavy pressure from national authoritarian elites, facing persistent
threats of their websites being shut down with reference to inappropriate comments
published (Suleymanov, 2014).
Comment section on news websites as authoritarian participatory
institutions
However, as the results also demonstrate, even in the most closed authoritarian
regimes in the region, and even on state-owned and regime-loyal news websites, comment
sections were still a relatively common phenomenon. Consider, for instance, Tajikistan and
Turkmenistan as two of the most closed authoritarian regimes included in the study. In
neither of these two countries did we identify a single news website that dared to publish
criticism of the country’s political leader (i.e., a critical media outlet). Yet, in both
countries, 50% of the invariably regime-loyal news websites operated comment sections.
The discourse architectures of these comment sections transferred relatively little control to
commenters, with average CTIs being 12.3 in Tajikistan and 13.5 in Turkmenistan. In both
countries, comment sections performed relatively well in the first dimension of our index,
indicating high visibility of these comment sections on news websites. By contrast,
comment sections in these countries performed poorly in the third dimension of the index,
indicating that an exceptionally large amount of control over the content published
remained at the backend, in the hands of media professionals (for details, see the
descriptive statistics provided in the online supplementary file).
In order to make sense of these empirical observations, we suggest drawing on
theories of citizen participation in authoritarian regimes (Toepfl, 2016; Stockmann, 2013).
Running head: TRANSFERRING CONTROL 28
As has been observed by political scientists, authoritarian regimes across the globe have
increasingly adopted, over the past three decades, a broad range of participatory
institutions that are typically associated with democracies, including elections, parliaments,
parties, consultative forums, and ‘marketized [news] media’ (Stockmann, 2013: 6; see also
Toepfl, 2016; Toepfl, unpublished manuscript; He and Warren, 2011). Many authoritarian
regimes thus vigorously encourage controlled types of citizen input, while they continue to
preclude regime-level democratization in the form of fair and competitive national
elections. Within this framework of thinking, comment sections of news websites,
particularly those of state-owned and regime-loyal websites, may be considered an
additional input institution facilitating citizen feedback to authoritarian rulers. Just as has
been argued with regard to other input institutions in authoritarian regimes, comment
sections may serve political elites for the purposes of gathering information about society,
credibly increasing the transparency of government, monitoring lower level officials, or
showcasing widespread regime support (Toepfl, 2016, unpublished manuscript;
Stockmann, 2013; He and Warren, 2011). The specific discourse architectures of comment
sections as we coded them in Turkmenistan and Tajikistan (highly visible, but tightly
controlled by media professionals) appear to be ideally poised to perform these types of
goals.
Limitations and promising paths for future research
Aside from contributing to the two strands of literature outlined above, the findings
of this study are valuable because they may serve as starting points for further in-depth
qualitative research. Most importantly, they can help qualitative researchers to develop
Running head: TRANSFERRING CONTROL 29
rationales for case selection. With regard to selecting news organizations for case analysis,
for instance, the CTIs of comment sections can be plausibly interpreted as indicators of the
fundamental approach to audience participation that a news organization seeks to convey
to the public (see our argument in the introduction and Karlsson et al., 2015; Jönsson and
Örnebring, 2011). With regard to selecting country cases, the proportion of a country’s
news websites adopting comment sections (CS), as well as the average CTIs of these
comment sections, may well be considered indicators of the role that comment sections as
an input institution play within a specific authoritarian regime. To illustrate this line of
thinking with but one example, the authors of this study have embarked on a qualitative
follow-up research project, based on in-depth interviews with media professionals and
experts in two countries. This project addresses the research question: Why do Belarus and
Azerbaijan, as two authoritarian regimes with comparable levels of PFD and IED, differ so
greatly with regard to the proportion of news websites implementing comment sections
(89% in Belarus versus 29% in Azerbaijan), as well as with regard to the mean CTIs of
these comment sections (19.4 in Belarus versus 13.4 in Azerbaijan; see Figure 1)?
That said, the present study also has a number of limitations. Firstly, it has focused
on the discourse architectures of only one participatory feature, comment sections. By
contrast, future studies could deploy a similar research design in order to comparatively
investigate other participatory features. Secondly, this study has focused on only one
region, the post-Soviet world. This poses clearly circumscribed limits to generalization.
Future research is required that would interrogate the findings presented by collecting data
within and across other regions. Thirdly, this study collected data at only one point in time.
By contrast, future studies could collect time-series data that facilitate tracing the spread of
Running head: TRANSFERRING CONTROL 30
participatory features as technological innovations across time and space. This type of data
could also permit researchers to tackle the complex issue of sorting out diffusion
mechanisms from other causes of variance in adoption patterns (‘Galton’s Problem’, see
also Peters, 1998: 41-43).
Fourthly, partly due to its large n design, this study has had to limit itself to a
comparison of discourse architectures only, that is, to a mere comparison of socio-
technological communication environments. By contrast, in-depth studies of fewer cases
could also explore the rationales of news organizations in different socio-political contexts
for implementing specific discourse architectures, as well as the complex ways by which
the discourse architectures implemented affect the actual content published. Fifthly, the
relatively abstract question remains as to how technological change, over the past decades,
has contributed to shaping practices of audience participation in authoritarian regimes. As,
for instance, Hough (1976) pointed out in a classic essay, the leading official newspaper of
the Communist Party of the Soviet Union, Pravda, received 409,000 letters to the editor in
1972. What are the continuities and key differences between the widespread practice of
writing letters to the editor in the Soviet Union and commenting on the news online in
Russia today? We believe that research of at least these five types could certainly raise
fascinating questions around why, how, and with what consequences news media across
the globe keep encouraging their audiences to participate in the publishing of news.
Running head: TRANSFERRING CONTROL 31
References
Bachmann I and Harlow S (2012) Opening the gates. Interactive and multimedia elements
of newspaper websites in Latin America. Journalism Practice 6(2): 217232.
Berk RA, Western B and Weiss RE (1995) Statistical inference for apparent populations.
Sociological Methodology 25: 421458.
De Keyser J and Sehl A (2011) May they come in? A comparison of German and Flemish
efforts to welcome public participation in the news media. First Monday 16(10).
Domingo D, Quandt T, Heinonen A, et al. (2008) Participatory journalism practices in the
media and beyond. Journalism Practice 2(3): 326342.
El Gody A (2015) Interactivity in Egyptian newspapers. Social Semiotics 25(1): 3356.
Freelon DG (2011) Talking among themselves: Online youth civic communication in
managed and autonomous environments. Information, Communication & Society
14(2): 198218.
Freelon DG (2015) Discourse architecture, ideology, and democratic norms in online
political discussion. New Media & Society 17(5): 772791.
He B and Warren ME (2011) Authoritarian Deliberation: The Deliberative Turn in Chinese
Political Development. Perspectives on Politics 9(02): 269289.
Himelboim I and McCreery S (2012) New technology, old practices: Examining news
websites from a professional perspective. Convergence: The International Journal of
Research into New Media Technologies 18(4): 427444.
Hough JF (1976) Political participation in the Soviet Union. Soviet Studies 28(1): 320.
Running head: TRANSFERRING CONTROL 32
Humprecht E and Esser F (2016) Mapping digital journalism: Comparing 48 news
websites from six countries. Journalism.
Ihlebæk KA and Krumsvik AH (2015) Editorial power and public participation in online
newspapers. Journalism 16(4): 470487.
Jannsen D and Kies R (2005) Online forums and deliberative democracy. Acta Politica 40:
317335.
Jönsson AM and Örnebring H (2011) User-generated content and the news. Empowerment
of citizens or interactive illusion? Journalism Practice 5(2): 127144.
Karlsson M, Bergström A, Clerwall C, et al. (2015) Participatory journalism the
(r)evolution that wasn’t. Content and user behavior in Sweden 2007–2013. Journal
of Computer-Mediated Communication 20(3): 295311.
Netzer Y, Tenenboim-Weinblatt K and Shifman L (2014) The construction of participation
in news websites. A five-dimensional model. Journalism Studies 15(5): 619631.
Neuendorf KA (2002) The content analysis guidebook. London: Sage.
Peters BG (1998) Comparative politics: Theory and methods. Basingstoke: Palgrave
MacMillan.
Reich Z (2011) User comments. The transformation of participatory space. In: Singer JB,
Hermida A, Domingo D, et al. (eds) Participatory journalism: Guarding open gates
at online newspapers. Oxford, UK: Wiley-Blackwell, pp. 96117.
Scott J, Millard D and Leonard P (2015) Citizen participation in news. An analysis of the
landscape of online journalism. Digital Journalism 3(5): 737758.
Suleymanov S (2014, March 19). Bez kommentariyev [Without comments]. TJournal.ru.
Available from https://tjournal.ru/p/no-comments
Running head: TRANSFERRING CONTROL 33
Singer JB (2014) User-generated visibility: Secondary gatekeeping in a shared media
space. New Media & Society 16(1): 5573.
Singer JB, Domingo D, Heinonen A, et al. (2011) Participatory journalism: Guarding
open gates at online newspapers. Oxford, UK: Wiley-Blackwell.
Stockmann D (2013) Media Commercialization and Authoritarian Rule in China.
Cambridge, UK: Cambridge University Press.
Suau J and Masip P (2014) Exploring participatory journalism in Mediterranean countries.
Journalism Practice 8(6): 670687.
Toepfl F and Piwoni E (2015). Public Spheres in Interaction: Comment Sections of News
Websites as Counterpublic Spaces. Journal of Communication 65(3), 465488.
Toepfl F (2016) Innovating consultative authoritarianism: Internet votes as a novel digital
tool to stabilize non-democratic rule in Russia. New Media & Society. Epub ahead of
print 22 November 2016. DOI: 10.1177/1461444816675444
Toepfl F (unpublished manuscript) Comparing authoritarian publics: Three models of the
multiple public-at-large in non-democratic regimes.
Thurman N (2008) Forums for citizen journalists? Adoption of user generated content
initiatives by online news media. New Media & Society 10(1): 139157.
Woolley SC and Howard PN (2016) Political Communication, Computational Propaganda,
and Autonomous Agents [special section]. International Journal of Communication
10.
Wright S and Street J (2007) Democracy, deliberation and design: The case of online
discussion forums. New Media & Society 9(5): 849869.
... 3 Comment sections beneath journalistic articles are currently one of the participatory features most widely adopted by news organizations around the globe (Toepfl & Litvinenko, 2018;Domingo et al., 2008;Singer, 2014). Extant research of this highly popular phenomenon has deployed a wide variety of theoretical perspectives and methodological approaches. ...
... Extant research of this highly popular phenomenon has deployed a wide variety of theoretical perspectives and methodological approaches. Previous studies have examined, for instance: the attitudes of journalists and audiences towards user comments (Heise et al., 2014); the degree to which discussions in comment sections correspond to deliberative norms (Ruiz et al., 2011); how journalists moderate deviant comments (Frischlich, Boberg & Quandt, 2019); and why and how news organizations implement comment sections (Toepfl & Litvinenko, 2018;Singer, 2014). What the majority of these studies have in common, however, is that they develop their arguments against a backdrop of empirical observation of democratic contexts (for exceptions, consider El Gody, 2015; Toepfl & Litvinenko, 2018;Tong, 2015). ...
... Previous studies have examined, for instance: the attitudes of journalists and audiences towards user comments (Heise et al., 2014); the degree to which discussions in comment sections correspond to deliberative norms (Ruiz et al., 2011); how journalists moderate deviant comments (Frischlich, Boberg & Quandt, 2019); and why and how news organizations implement comment sections (Toepfl & Litvinenko, 2018;Singer, 2014). What the majority of these studies have in common, however, is that they develop their arguments against a backdrop of empirical observation of democratic contexts (for exceptions, consider El Gody, 2015; Toepfl & Litvinenko, 2018;Tong, 2015). By contrast, very little is known about how, why, and with what consequences audiences are invited to comment on news in contemporary authoritarian regimes, and about how journalists in repressive contexts evaluate and moderate audience participation. ...
Article
Full-text available
Extant research on how innovations diffuse among news organizations over time has largely focussed on democratic contexts. By contrast, this is the first longitudinal study to investigate the spread of a participatory newsroom innovation under authoritarian rule. Adopting a multiple case study design, the article reconstructs the histories of comment sections on the opinion-leading online media in two authoritarian contexts, which varied maximally with regard to the outcome of the diffusion process. In Belarus, the diffusion process followed a classic S-shaped curve of adoption, whereas in Azerbaijan adoption rates remained low during the studied period. The study identified primarily three factors that obstructed the diffusion process in Azerbaijan: (1) the restrictive policy of the authoritarian leadership specifically towards audience participation on news websites (social system), (2) the low intensity of communicative exchange between local and foreign news organizations (communication channels), (3) the advent of the successor technology in 2010 (time).
... 3 Comment sections beneath journalistic articles are currently one of the participatory features most widely adopted by news organizations around the globe (Toepfl & Litvinenko, 2018;Domingo et al., 2008;Singer, 2014). Extant research of this highly popular phenomenon has deployed a wide variety of theoretical perspectives and methodological approaches. ...
... Extant research of this highly popular phenomenon has deployed a wide variety of theoretical perspectives and methodological approaches. Previous studies have examined, for instance: the attitudes of journalists and audiences towards user comments (Heise et al., 2014); the degree to which discussions in comment sections correspond to deliberative norms (Ruiz et al., 2011); how journalists moderate deviant comments (Frischlich, Boberg & Quandt, 2019); and why and how news organizations implement comment sections (Toepfl & Litvinenko, 2018;Singer, 2014). What the majority of these studies have in common, however, is that they develop their arguments against a backdrop of empirical observation of democratic contexts (for exceptions, consider El Gody, 2015; Toepfl & Litvinenko, 2018;Tong, 2015). ...
... Previous studies have examined, for instance: the attitudes of journalists and audiences towards user comments (Heise et al., 2014); the degree to which discussions in comment sections correspond to deliberative norms (Ruiz et al., 2011); how journalists moderate deviant comments (Frischlich, Boberg & Quandt, 2019); and why and how news organizations implement comment sections (Toepfl & Litvinenko, 2018;Singer, 2014). What the majority of these studies have in common, however, is that they develop their arguments against a backdrop of empirical observation of democratic contexts (for exceptions, consider El Gody, 2015; Toepfl & Litvinenko, 2018;Tong, 2015). By contrast, very little is known about how, why, and with what consequences audiences are invited to comment on news in contemporary authoritarian regimes, and about how journalists in repressive contexts evaluate and moderate audience participation. ...
Article
Full-text available
Little is known presently about how, why, and with what consequences audiences comment on their news in contemporary authoritarian regimes. In order to address this gap, this study leverages recent theorizing about the multiple public sphere under non-democratic rule. Accordingly, critically commenting publics are theorized as “input institutions” that not only create risks but also offer important benefits for autocrats. Grounded in this approach, the study develops a series of hypotheses about the extent of political criticism that should be visible beneath the news in three purposefully selected authoritarian contexts: Azerbaijan, Russia, and Turkmenistan. In order to test these hypotheses, commenting environments facilitated (or not) by 46 leading news organizations on seven platforms were considered (N=322). For each environment, coders established whether comments were published that were (1) critical of the autocrat himself, (2) critical only of lower-level policies or officials of the regime, or (3) entirely uncritical. As the findings show, the extent of readers’ criticism differed systematically between the three contexts, broadly following the patterns hypothesized. Moreover, in line with this study’s key assumptions, critically commenting publics were facilitated not only by opposition media but also by substantial numbers of state-controlled news organizations.
... Other studies confronted these optimistic beliefs revealing incivility, homophily, and polemics in user discussions [1]. Researchers were concerned about the low quality of comments (offensive nature, lack of relevant discussion) on the news websites [43], on textual social media [41], and on video hosting platforms like YouTube [6]. Financial reasons for the media business [19] or negative attitude of journalists to the comment sections [7] led to the closure of the comment sections on the websites of news media. ...
Chapter
Full-text available
Implementation of user comments created a new phase in audience participation. However, high expectations on rational, responsible, and civil user commenting have quickly vanished. Online, people want to interact with others, to express their opinion on the problem, and to teach others. Thus, recent studies criticize the quality of discussions formed in the comment sections on Facebook pages of news media as well as on the news websites. High-quality discussions demand from the participants to actively connect with other participants’ comments. This research aims at exploring the interactivity of discussions that are formed through comments under FB posts published on the news media FB pages. We focus on the media for migrants and, of them, upon those for Russian-speaking migrants in Germany. This group of Internet users tend to be excluded from the public sphere of the host country, but migrant media allow them to join public debate. Thus, two German news FB public pages in Russian with more than 25,000 subscribers in the spring of 2019 were selected for analysis: German News and Germany24. The level of interactivity is measured as the ratio of comments posted in response to another user’s comment to the total number of comments under the post. On average, every third comment to the most commented posts was published as an answer to the comment of another user. Due to the criticism towards comments sections and evidence of absence of a real dialogue, we might seem this average level as a significant one.
... S.S.Gulyamov, who devoted his works to the legal foundations of corporate governance, provides for the consideration of commercial banks as a joint-stock company -a commercial legal entity. He pointed out that according to the banking legislation of the Republic of Uzbekistan [19], banks are created, as a rule, in the form of a jointstock company on the basis of any form of ownership. Consequently, the provisions of the Law of the Republic of Uzbekistan "On Joint-Stock Companies and Protection of Shareholders' Rights" apply to banks in the form of a joint-stock company (Gulyamov S.S., 2018). ...
Article
Full-text available
The banks of Uzbekistan had to go through a long evolutionary path of development, overcome considerable difficulties to become an important system that affects the national economy today. In strengthening the banking system, the legislative framework is important, and therefore the study of development priorities is relevant. The purpose of this article is to develop proposals and recommendations for further liberalization and harmonization of banking legislation, simplification of the procedure for participation of foreign banks in the banking system of the Republic of Uzbekistan, strengthening the legal status of banks. The results of the study show that the banking legislation is not systematized, integration into the international banking system is weak, and the Central Bank of the Republic of Uzbekistan has broad administrative powers. The implication of the research is that strict legal norms and prohibitions against foreign banks make the banking sector of Uzbekistan not attractive. This study can be used as a reference point for the liberalization and harmonization of banking legislation in Uzbekistan.
... In 1999, Estonian press publications opened the door to the involvement of readers by creating online publications. Toepfl and Litvinenko (2018), who compared the online publications of the former Soviet republics, found that compared to other news portals, the Estonian portals are an example of an open and inclusive online architecture. The comments section is an anticipated feature of online publications, which enables the audience to express its opinions, direct the discussion and raise new issues. ...
... This assumption is based on comparative studies using concepts such as "journalism cultures" (Hanitzsch, 2011) or "media systems" (Hallin & Mancini, 2004) to trace the differences between journalistic organisations and routines at the local level. These differences are also known to influence the process of adopting media innovations as shown by multiple studies on media innovation and its adoption (see, for instance, Hanusch et al., 2019;Humprecht & Esser, 2018;Lehtisaari et al., 2018;Nozal Cantarero et al., 2020;Toepfl & Litvinenko, 2018) that suggest that these processes develop differently for specific media systems and media markets. ...
Article
Full-text available
News outlets worldwide increasingly adopt user- and system-driven personalisation to individualise their news delivery. Yet, the technical implementation of news personalisation systems, in particular the one relying on algorithmic news recommenders (ANRs) and tailoring individual news suggestions with the help of user data, often remains opaque. In our article, we examine how news personalisation is used by quality and popular media in three countries with different media accountability infrastructures - Brazil, the Netherlands, and Russia - and investigate how information about personalisation usage is communicated to the news readers via privacy policies. Our findings point out that news personalisation systems are predominantly treated as black boxes that indicate a significant gap between practice and theory of algorithmic transparency, in particular in the non-EU context.
... Thus, readers' comments are not only important sources of knowledge about the audience's perception of existing social problems but also an instrument of social problem formation. It is also important that the agenda-setting, framebuilding, problem-setting, and opinion-forming roles of comments have been found in different societies far beyond simply the "West", including such diverse countries as China (Zhou & Moy, 2007), South Korea (Lee & Jang, 2010), India (Jayachandran, 2015), as well as Post-Soviet countries (Toepfl & Litvinenko, 2018). ...
Article
Full-text available
Media audiences representing a significant portion of the public in any given country may hold opinions on media-generated definitions of social problems which differ from those of media professionals. The proliferation of online reader comments not only makes such opinions available but also alters the process of agenda formation and problem definition in the public space. Based on a dataset of 33,877 news items and 258,121 comments from a sample of regional Russian newspapers we investigate readers’ perceptions of social problems. We find that the volume of attention paid to issues or topics by the media and the importance of those issues for audiences, as judged by the number of their comments, diverge. Further, while the prevalence of general negative sentiment in comments accompanies such topics as disasters and accidents that are not perceived as social problems, a high level of sentiment polarization in comments does suggest issue problematization. It is also positively related to topic importance for the audience. Thus, instead of finding fixed social problem definitions in the reader comments, we observe the process of problem formation, where different points of view clash. These perceptions are not necessarily those expressed in media texts since the latter are predominantly “hard” news covering separate events, rather than trends or issues. As our research suggests, problematization emerges from readers’ background knowledge, external experience, or values.
... All these diverse types of media, including state-run, maintain their accounts in social networks and permit some form of comment sections there, or even host them on their websites (Toepfl & Litvinenko, 2018), thus becoming actors of the online deliberation. Unlike China, Russia has shown no evidence of developing a large-scale centralized censorship system involved in mass deletion of user messages (which is useless, given the availability of foreign social networks), although it has already introduced a limited blocking of international (LinkedIn) and Russian-language (Telegram) networking platforms. ...
Article
Full-text available
Russian media have recently (re-)gained attention of the scholarly community, mostly due to the rise of cyber-attacking techniques and computational propaganda efforts. A revived conceptualization of the Russian media as a uniform system driven by a well-coordinated propagandistic state effort, though having evidence thereunder, does not allow seeing the public discussion inside Russia as a more diverse and multifaceted process. This is especially true for the Russian-language mediated discussions online, which, in the recent years, have proven to be efficient enough in raising both social issues and waves of political protest, including on-street spillovers. While, in the recent years, several attempts have been made to demonstrate the complexity of the Russian media system at large, the content and structures of the Russian-language online discussions remain seriously understudied. The thematic issue draws attention to various aspects of online public discussions in Runet; it creates a perspective in studying Russian mediated communication at the level of Internet users. The articles are selected in the way that they not only contribute to the systemic knowledge on the Russian media but also add to the respective subdomains of media research, including the studies on social problem construction, news values, political polarization, and affect in communication.
Chapter
Recently, the media sphere has been found to be regularly flooded with misinformation and disinformation. Some scholars refer to such outbreaks as infodemics (Bruns, A., Harrington, S., & Hurcombe, E., Media International Australia, 2020). We are at a point where we must challenge the assumption of online spaces as emblematic of democratic ideals where participation is assumed to solely foster healthy debate in an authentic marketplace of ideas. Prominent in this subverted media arena are bots, inauthentic participants in the exchange of information, and trolls, the “Ghostwriters,” flamethrowers of the internet, whether automated or not. This problem is felt acutely in Lithuania where there are fears that a Russian information war could turn into real war or other political disruption. It is, therefore, imperative to understand the process of this media manipulation. This chapter argues that researchers will have to adapt theories and methodologies to do so by linking the theoretical groundings of mass media influence to the concept of information warfare (Cronin, B., & Crawford, H., Information Society 15:257–263, 1999), where social media chaos can warp civic participation (Zelenkauskaite, A., Creating Chaos Online: Disinformation and Subverted Post-Publics. University of Michigan Press, 2022).
Chapter
In this study, we focus on decrease in public activity of news media audiences on social networks and on possible media strategies of “comfortable involvement” of socially anxious people into commenting. We draw upon cognitive research that describes a vicious circle of user experience. The environment for the anxiety scenarios in relation to user comments creates favorable conditions for the growth of digital escapism. We hypothesize that: 1) Users who are looking for getting rid of anxiety on social networks are drawn into an even greater spiral of anxiety, interacting with other people in the context of news stories that provoke the whipping up of negative emotions; 2) The dynamics of the cascade of messages can depend on the characteristics of the emotions embedded in the message and its context. We took 10 Russian regional media and their content on the VKontakte social network published from November 2020 to November 2021 and processed all commented posts in each media. Frequency analysis of messages demonstrated no specific pattern for anxiety. Number of comments and likes on comments correlated with specific regional issues and local news agenda more than with coronavirus agenda. Our results showed that engaged users comment more than 2 times in a period but drop out of the discussion in 2–3 moths if agenda is not escalated by news outlet itself. Emotional triggers in news stories containing reasons for anxiety depends on a combination of factors from the site’s functionality to the type of news media and local information policy.
Article
Full-text available
Extant research on the consequences of the Internet for non-democratic politics has focused on how oppositional activists leverage new digital tools. By contrast, still, relatively little is known about how authoritarian elites proactively deploy digital technologies to legitimize their rule. This article contributes to filling this gap by scrutinizing one highly innovative tactic that has recently been adopted repeatedly by Russia’s ruling elites: the organization of ‘Internet votes’ to staff advisory bodies to the government. In contrast to online petitions, online votes are aimed at aggregating citizen preferences not on issues but on candidates, that is, on individuals who later act as political representatives. The article presents an in-depth case study of the first such Internet vote conducted in Russia in 2012. It concludes that ruling elites deployed the tool swiftly to (1) disempower oppositional activists and (2) convey to the mass public the image of a transparent, accountable and responsive government.
Article
Full-text available
Media organizations throughout the Western world struggle to adjust their practices to rapidly changing conditions. Initially, online journalism was celebrated for potentially revolutionizing political reporting due to its new technological possibilities: According to this, it is able to (1) increase transparency by providing hyperlink sources, (2) increase understanding by providing further background information, and (3) add to deliberation and follow-up communication by providing a platform for interactive exchange. A comparative content analysis of 48 news websites from six countries (France, Italy, Germany, Switzerland, Great Britain, and United States) examines the degree to which these three potential strengths are fully exploited. By mapping the different news outlets in relation to the digital functions, we identify three models prevalent in different countries and organization types. The first model contains outlets promoting the usage of links to make their sources transparent to the reader (‘transparency model’), outlets focusing on the provision of background information to enable their audiences to gain a wider understanding of the reported topic (‘background model’), and outlets that mainly avoid the adoption of new technologies (‘print-oriented model’). These findings show that different structural developments and professional orientations lead to the adaption of different technologies in digital journalism.
Chapter
Full-text available
Article
In most liberal democracies commercialized media is taken for granted, but in many authoritarian regimes the introduction of market forces in the media represents a radical break from the past with uncertain political and social implications. In Media Commercialization and Authoritarian Rule in China, Daniela Stockmann argues that the consequences of media marketization depend on the institutional design of the state. In one-party regimes such as China, market-based media promote regime stability rather than destabilizing authoritarianism or bringing about democracy. By analyzing the Chinese media, Stockmann ties trends of market liberalism in China to other authoritarian regimes in the Middle East, North Africa, sub-Saharan Africa, and the post-Soviet region. Drawing on in-depth interviews with Chinese journalists and propaganda officials as well as more than 2,000 newspaper articles, experiments, and public opinion data sets, this book links censorship among journalists with patterns of media consumption and media's effects on public opinion.
Article
Studies of political discussions online have been dominated by approaches that focus exclusively on deliberation, ignoring other equally relevant communication norms. This study conducts a normative assessment of discussion spaces in two prominent web platforms—Twitter hashtags and newspaper comment sections devoted to particular political issues—applying the norms of communitarianism, liberal individualism, and deliberation. The platforms’ distinct design features and users’ left/right issue stances emerge as significant predictors of normative differences.
Article
This article explores implications of the transition to an environment in which users have become secondary gatekeepers of the content published on media websites. This expanded user role, facilitated by technology and enabled by digital news editors, includes assessment of contributions by other users; communication of the perceived value or quality of user- and journalist-produced content; and selective re-dissemination of that content. The result is a two-step gatekeeping process, in which initial editorial decisions to make an item part of the news product are followed by user decisions to upgrade or downgrade the visibility of that item for a secondary audience. Preliminary empirical evidence indicates these user gatekeeping capabilities are now pervasive on US newspaper sites.
Article
The utilisation of information and communications technologies (ICTs) in Egypt has irrevocably changed the nature of the traditional Egyptian public sphere. One can see the Egyptian online society as a multiplicity of networks. These networks have developed, transformed and expanded over time, operating across all areas of life. Nonetheless, in essence they are sociopolitical and cultural in origin. This trend changed the way audiences consumed news, with traditional media – especially independent and opposition – started to utilise ICTs to access online information to develop their media content to escape government control. Several media organisations also started to expand their presence online so that, as well as providing news content, they also provided them with a “space” to interact amongst themselves and with media organisations. Audiences started to provide detailed descriptions of Egyptian street politics, posting multimedia material, generating public interest and reinforcing citizen power hence democratic capacity.