Sphere layout final-051415

This report was prepared by three CLALS Research Fellows—principal investigators Michael S. Dan-
ielson, Michael M. McCarthy, and Paula Orlando—with advice from CLALS Director Eric Hershberg.
The work was supported by research assistants Peter Bolton, Olivia Bullock, Adeline Hite, Constanza
Martínez, and Marcie Neil, as well as Center administrators Jacquelyn Dolezal, Inés Luengo de Krom,
and Dennis Stinchcomb. Patrick Breslin provided editorial assistance with several sections.
Bias or Neutrality?
An Assessment of Television News Coverage in Venezuela by Globovisión
A study prepared by
American University’s Center for Latin American & Latino Studies (CLALS)
A P R I L 2 0 1 5

PullquotePullquote
Table of Contents
Executive Summary......................................................................................2
I. Introduction..............................................................................................3
Objectives.............................................................................................3
Brief History of Globovisión..................................................................4
Data and Method.................................................................................5
Summary of Findings...........................................................................5
II. The Venezuelan Context: Elections, Constitutions, and Media.............7
Not One Step Back: The Coup and Deepened Polarization, 2002............8
A Fateful Face-Off: The Chávez Government and the Media...................8
The Oil Strike and Referendum, 2002–2004..........................................8
Reshaping the Media Landscape...........................................................10
Expanded Executive Powers, 2008–2010.............................................10
Chávez’s Death and Maduro’s Election..................................................11
A New Media Environment.................................................................12
2014: A Year of Difficulties..................................................................13
III. Approaches to Determining Media Bias.............................................14
Manifestations of Bias..........................................................................15
A Comparative Framework for Researching Bias....................................15
Research on Media Bias and Political Communication..........................15
Two Frameworks: News Production and News Reception.......................16
Three Types of Bias: Gatekeeping, Coverage, and Statement....................17
Live Coverage......................................................................................18
IV. Content Analysis of Globovisión’s News and Opinion Coverage.......20
Measuring Political Bias......................................................................20
A. Access to Media: Visibility.....................................................20
B. Favorability of Actor Portrayal...............................................21
C. Favorability of Topic Presentation and Framing....................21
D. Gatekeeping Bias..................................................................22
Case Selection......................................................................................23
A. Selection of Critical Juncture Periods....................................23
B. Identification of News and Opinion Programs for Study......25
C. Sampling of Broadcasts.........................................................26
D. Data Collection....................................................................26
V. Empirical Analysis and Findings...........................................................28
Indicators of Bias: Descriptive Statistics and Comparisons......................28
A. Access to Media: Visibility....................................................28
B. Favorability of Actors Portrayed...........................................32
C. Topics of Coverage and Gatekeeping...................................36
D. Favorability of Topic Presentation and Framing...................38
Measuring Overall Bias and What Explains It.....................................40
VI. Conclusion...........................................................................................42
Neutrality and Balance.......................................................................42
A Closing Note on Breadth and Depth: Coverage of Issues and Actors....45

1
CONTENTS
References......................................................................................................46
VII. Appendices..........................................................................................49
Appendix 1. Codebook..............................................................................49
Appendix 2. Codebook Supplementary File 1: Topics and Subtopics.............53
Appendix 3. Codebook Supplementary File 2: CJ Period Summary..............54
Appendix 4. Index of Bias..........................................................................56
Appendix 5. Supplementary Tables and Figures...........................................56
Figures and Tables
Figures
Figure 1: Visibility Scale by Actor Partisanship
Figure 2: Minutes Visible by Actor Partisanship
Figure 3: Nonpartisan, Pro-Government, and Pro-Opposition Actors
Figure 4: Net-Favorability of Actors at the Story/Segment Level of Analysis by Street
Demonstrations, Others, and All Periods
Figure 5: Percent of Broadcasts Covering Target Topic by Critical Juncture Period
Figure 6: Subtopics Covered by Critical Juncture Period
Figure 7: Verbal Favorability of Topic Presentation
Figure 8: Non-Verbal Favorability of Topic Presentation
Figure 9: Comprehensive Index of Bias: Frequency Distribution
Figure A-1: Net Government Visibility Quality: Frequency Distribution
Figure A-2: Net Government Minutes Visible: Frequency Distribution
Figure A-3: Net Government Actor Favorability: Frequency Distribution
Figure A-4: Net Government Actor Non-Verbal Favorability: Frequency Distribution
Figure A-5: Net Government Topic Favorability: Frequency Distribution
Figure A-6: Net Government Topic Non-Verbal Favorability: Frequency Distribution
Figure A-7: Pro-Government Bias and Story Prioritization: Scatter Plots by Critical Juncture
Tables
Table 1: Total Number of Minutes Visible
Table 2: Favorability of Individuals toward the Government (Verbal)
Table 3: Favorability of Individuals toward the Government (Non-Verbal)
Table 4: Net-Favorability of Actors in Stories
Table 5: Net-Favorability of Actors in Broadcasts
Table A-1: Final Content Analysis Reliability Indicators
Table A-2: Topics and Subtopics Covered by Critical Juncture Period
Table A-3: Coverage of Target Topic
Table A-4: Dependent and Independent Variables Used in OLS Regressions
Table A-5: Explaining Variations in the Index of Bias
Table A-6: Factor Analysis of Measures of Bias

2 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | E X E C U T I V E S U M M A R Y
This study assesses the extent to which the Globovisión television station provided
balanced coverage of political and social events in Venezuela during the polarized
period spanning from the fall of 2013 through the summer of 2014. It analyzes
Globovisión’s coverage surrounding the municipal elections of 2013; the wave of
opposition street demonstrations in early 2014; international efforts to convene
talks between the government and opposition in the spring of 2014; and the short-
ages of basic goods during the summer of that year. The study focuses on whether
Globovisión provided unbiased coverage in terms of the access given to representa-
tives from the government and opposition, the favorability of the presentation and
portrayal of prominent individuals and organizations, and the choice of issues and
perspectives receiving coverage.
Drawing on state-of-the-art scholarly methods of media content analysis, the study
analyzes data from 88 hour-long broadcasts, including some 600 news stories or
interview segments featuring more than 1,000 different individuals and discussion
of 36 different topics. The study finds that Globovisión’s framing of the issues
tended to be neutral, and that bias in favor of the government was no more likely
than bias in favor of the opposition.
While overall Globovisión achieved its goal of providing unbiased coverage of
the most critically important social and political events in the country, there were
differences between the periods and issues covered, as well as across particular
measurements of bias. Opposition voices received more total coverage than pro-
government voices during the periods surrounding the municipal elections and the
street demonstrations, and this coverage tended to be unfavorable to the govern-
ment. However, pro-government perspectives received slightly more coverage
when the international dimensions of the crisis or the shortages of basic goods were
the topics of discussion. The analysis also shows that news stories that appeared
earlier in a broadcast tended to be relatively more pro-government, whereas those
appearing toward the end of a broadcast were more favorable to the opposition.
Executive Summary

I. INTRODUCTION | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 3
I. INTRODUCTION
OBJECTIVES
T
his study assesses the extent to which Venezuelan television station Globovisión
provided balanced coverage of critical political and social events in Venezuela dur-
ing the station’s first year under new ownership. It considers this question in light
of the deep divisions between chavismo and the opposition, which led to a highly
polarized political environment marked by competing interpretations of events.
For Globovisión this created the challenge of providing visibility and fair depiction to sharply
different perspectives on enormously controversial events.
To examine Globovisión’s coverage of the most critical political and social developments in
Venezuela during this period, this study analyzes content aired during four “critical junc-
tures.” These include: 1) “The Municipal Elections Period” (November 9–December 15,
2013); 2) “The Street Demonstrations” (February 12–March 21, 2014); 3) “The International
Dimensions of Crisis,” when both the Union of South American Nations (UNASUR) and
the Vatican sponsored mediation talks (March 25–May 14, 2014); and “Shortages of Basic
Goods” (August 7–29, 2014). These four periods of inflection were characterized by high
degrees of polarization, and analysis of how they were handled by the station enables us to as-
sess whether it gave visibility to different interpretations of events, aired balanced portrayals of
government and opposition perspectives, and chose to broadcast coverage of highly controver-
sial events.
The structure of this report is as follows. The first section provides political context, from
Chávez’s election in 1998 to Maduro’s first eighteen months in office (1998–2014). It also
addresses Venezuela’s elections, institutional developments, and shifting media landscape.
The second part defines key terms used in the study and draws on the scholarly literature to
explain why “balance” is best analyzed through the prism of “bias.” The third section provides
an in-depth description of the study’s methodology. The fourth section presents the empiri-
cal analysis of Globovisión’s coverage and details the study’s findings. The concluding section
summarizes those findings and assesses their implications.

4 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | I. INTRODUCTION
BRIEF HISTORY OF GLOBOVISIÓN
In December 1994 Venezuelan citizens Luis Teófilo Núñez Arismendi, Guillermo Zuloaga Núñez,
Nelson Mezerhane Gosen, and Alberto Federico Ravell Arreaza founded a national television station,
Globovisión. The channel became the first 24-hour news network in Venezuela, with headquarters
and an operational base in the capital city of Caracas. The channel began broadcasting to the country’s
main urban areas—the Federal District, Aragua, Carabobo, Miranda, and Zulia—and soon formed
domestic and international partnerships to disseminate its programming via third parties and abroad.
Globovisión could also be found on the dial of virtually all Venezuela’s cable subscribers.
During the Chávez Presidency (1999–2014), the channel gained increased public visibility. Alberto
Federico Ravell, who became the channel’s principal director, was directly involved in the daily news-
room, and his open criticism of the government made him and Globovisión prominent regime an-
tagonists. In April 2002, Globovisión and other privately owned channels broadcast news that favored
the opposition’s version of events during an aborted coup attempt. Globovisión’s coverage was broadly
consistent with the opposition narrative, which claimed that a civil society uprising, not a military
coup, had prompted Chávez to leave office. After the failure of the coup and the president’s return,
Globovisión maintained its hard stance against the Chávez government.
The difficult relationship between Globovisión’s initial ownership group and the Chávez government is
illustrative of a broader pattern of conflict and change in the Venezuelan media landscape. As the me-
dia’s role in politics grew during the president’s first seven years in office, private media outlets moved
to the center of attention in the highly polarized environment. Some adopted pro-opposition editorial
lines, arguably fueling polarization. This set the stage for spiraling confrontation between privately
owned media and the government.
In 2004 the government began to craft a new legal framework for regulating private media. That same
year, Venevisión, another prominent privately owned channel, began to soften its editorial line. Then, in
2007, the country’s other longstanding private station, RCTV, was stripped of its license to broadcast.
RCTV’s principal owner had been an outspoken critic of the Chávez government, and observers widely
interpreted the non-renewal of its license as a sign of an increasingly aggressive government approach to
the media. These developments at Venevisión and RCTV left Globovisión as the lone channel with an
editorial linecritical of the government. Globovisión thus became the media outlet carrying the flag for
the opposition and “independent private media.”
The channel’s leadership insisted relentlessly that station programming provided an independent voice
in a media landscape subject to politically motivated pressures from the state regulatory commission for
telecommunications, CONATEL. In 2009, CONATEL opened investigations into Globovisión’s news
coverage and accounting practices. These led to heavy fines being levied against the station.
Chávez’s criticism of Globovisión became personal, as he directly identified Ravell as the source of the
channel’s putatively unfair coverage. Elements of the state security forces carried out a publicly tele-
vised search of owner Zuloaga’s house, before temporarily arresting him. Subsequently, the government
obtained a minority stake in the station (approximately 25 percent) after the state nationalized a bank
owned by Globovisión’s founder and still part-owner Mezerhane. New owners bought Globovision
shortly after the April 2013 presidential elections, as the previous ownership group claimed that the

I. INTRODUCTION | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 5
channel had become financially unsustainable. They felt forced to sell.
Under new ownership, Globovisión introduced significant changes to programming, and ensuing
changes in personnel saw the departure of its star evening news host Leopoldo Castillo. The channel’s
slogan became “Responsible and Truthful Information,” as it maintained its focus as a news station but
changed its editorial line.1
The new owners claimed they would simply “transmit the news,” make an
effort to “lower the levels of conflict,” and work toward “peace” in Venezuela.2
In this vein, the new
owners signaled that they wanted Globovisión to be a neutral voice in the country’s heated political
conflict.
DATA AND METHOD
This study analyzes data from 88 hour-long broadcasts, including some 600 news stories or interview
segments featuring more than 1,000 different individuals and discussion of 36 different topics. The
data, supplied by Globovisión, consists of footage from three of its television programs, including the
channel’s most widely viewed news program and its two most popular opinion and interview broad-
casts. The research team selected these shows for analysis based on a comprehensive review of the
channel’s programming, and Globovisión delivered footage for a range of dates that were specified by
the research team on grounds that they encompassed the most important coverage aired during the
four critical juncture periods.3
To generate data for the quantitative analysis of potential bias in Globovisión news, information, and
opinion coverage, we conducted a thorough content analysis. Content analysis is a method to sys-
tematically study a broad range of textual, audio, and video content in order to identify patterns and
trends of interest, in this case media bias.4
Drawing on this technique, researchers coded and analyzed
a selected sample of Globovisión most watched news, information, and opinion broadcasts during the
station’s first year under new ownership to identify any patterns of bias and to determine the principal
factors associated with such bias. In this manner, the study developed measurements of bias regarding
the access given to representatives from the government and opposition, how favorably relevant indi-
viduals and organizations were portrayed, and the range of issues and perspectives receiving coverage.
SUMMARY OF FINDINGS
As outlined in detail in the Empirical Analysis and Findings section of this report, the study found
that Globovisión’s framing of the issues tended to be neutral, and that bias in favor of the govern-
ment was no more likely than bias in favor of the opposition. While overall Globovisión achieved
its goal of providing unbiased coverage of the most critically important social and political events in
the country, there were differences between the periods and issues covered as well as across particular
measurements of bias. Opposition voices received more coverage than pro-government voices during
1
Previous slogans for the channel were, from 1994–2010, “24 hours of information,” and then later, 2012–2013, “the
expression of a country.”
2
Quoted from a broadcast on May 23, 2013. Available at https://www.youtube.com/watch?v=GzCmX5Uifu0 (Accessed
24 April 2015).
3
In some instances, interruptions of regularly scheduled program content occurred. Such interruptions, which ranged from
breaking news stories to official state announcements by the president, or from members of his cabinet, involved live news
events. This live content was included in the sample.
4
Data collection entailed following five steps: developing the codebook and coding form, training coders, piloting tests of
validity and reliability, finalizing intercoder reliability tests, and assigning coders to review specific broadcasts.

6 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | I . I N T R O D U C T I O N
the periods surrounding the municipal election and the street demonstrations, and this coverage tended
to be unfavorable to the government. However, pro-government perspectives received slightly more
coverage when the international dimensions of the crisis or the shortages of basic goods were the focus
of broadcasts. The analysis also shows that news stories that appeared earlier in programs tended to be
relatively pro-government, whereas those appearing toward the end of a broadcast were more favorable
to the opposition.

II. THE VENEZUELAN CONTEXT | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 7
II. THE VENEZUELAN CONTEXT:
ELECTIONS, CONSTITUTIONS, AND MEDIA
T
he 1998 presidential elections marked a watershed in Venezuela’s history. After 40 years
of two-party democratic rule known as the Punto Fijo era—from a 1958 agreement
among political parties to share and rotate power—many Venezuelans became disen-
chanted with established political elites. In the campaign none of the main contenders
ran on the tickets of traditional parties. Hugo Chávez, a former army lieutenant who
had been jailed for staging a failed coup in 1992, distinguished his candidacy by promising to
challenge established political elites and institutions. His message resonated in a time of economic
recession and disillusionment with the establishment parties—Accíon Democrática (AD) and
Comité de Organizacíon Política Electoral Independiente (COPEI). Chávez coasted to victory with
56 percent of the vote.
His first moves were to orchestrate a constitutional reform process. Votes were held on the pro-
posal to convene a constituent assembly, to select the members of the assembly, and to approve the
newly drafted Constitution. That 1999 Constitution, which renamed the country the Bolivarian
Republic of Venezuela, mandated new elections. Chávez won a new term, 2000–2006, and the
Constitution allowed him to stand for consecutive re-election. The stage was set for deeper chang-
es when the 2000–2005 Congress granted Chávez decree powers. In November 2001, Chávez
decreed 49 laws, the most controversial of which increased state control over the oil sector and
called for a land reform that implied a new property rights regime. The expansion of state jurisdic-
tion, coupled with Chávez’s use of decree powers, stirred heated opposition.
Established interest groups, rather than parties, served as the main vehicles for channeling anti-
government sentiment. With the business federation FEDECAMARAS and the country’s main
labor organization Centro de Trabajadores Venezolanos (CTV) speaking for organized anti-Chávez
forces, party politicians moved to the background. The media played a major role in stirring
up opposition sentiment. In the 1998 presidential elections, private media had tended to favor
Chávez editorially once he became the frontrunner. By 2001, though, most had pivoted to a hard
anti-government editorial line and coverage seemed in step with the views of opposition-aligned
civil society groups.

By the end of 2001, instead of a governing party competing with opposition parties, Chávez, a charismatic
leader representing a new coalition, competed with traditional civil society interest groups that represented
the old guard. With communication between government and opposition closed off, politics moved from
institutions to the streets.
NOT ONE STEP BACK: THE COUP AND DEEPENED POLARIZATION, 2002
In early 2002, mass demonstrations, both for and against the government, consumed the country. By
March of that year, rhetoric heightened to the point that coup rumors became highly plausible. On
April 11, CTV President Carlos Ortega headed an opposition march in Caracas to an office of Petróleos
de Venezuela (PDVSA), the state oil company that came to symbolize the battle for political control over
Venezuela’s oil-dependent economy. With the slogan, “Ni un paso atrás” (not one step back), Ortega
called for the march to continue downtown to Miraflores, the presidential palace. With the march
headed for Miraflores, a coup went into motion. In the blocks around Miraflores, a firefight between
government supporters and the opposition left 19 dead. On April 12 at 4 a.m., Chávez was taken
from Miraflores palace to a distant military base. Soon after, the President of FEDECAMARAS,
Pedro Carmona, appeared on television to announce an interim government. It lasted only 24 hours.
Pro-Chávez military leaders, primarily in the Army, reacted when Carmona named a government com-
posed of conservative officials and moved to dissolve all institutions associated with the 1999 Constitution.
Without the support of top army officials, the Carmona government collapsed. The next day, April 13,
found Carmona fleeing Miraflores, the pro-Chávez wings of the army in command, and chavista demon-
strators descending upon the palace. The stage was set for Chávez’s return. By early morning, April 14,
the army returned Chávez to Miraflores by helicopter for a hero’s welcome from the emotional crowd of
supporters gathered outside the palace.
A FATEFUL FACE-OFF: THE CHÁVEZ GOVERNMENT AND THE MEDIA
The role the media had played during the coup generated intense controversy. Chavistas accused private
television stations of slanting news coverage to favor the opposition. They claimed that the media present-
ed pro-government demonstrations in a highly negative light while portraying anti-government demonstra-
tions in a highly positive light.
Two examples were frequently cited. First, television stations broadcast footage appearing to implicate
chavistas in an ambush of peaceful marchers, a charge that was undermined by subsequent investigations.
Second, chavistas expressed outrage at television stations effectively for “blacking out” news unfavorable to
the Carmona government. For instance, on April 13, while Carmona’s government was falling apart and
chavistas were descending upon Miraflores to call for Chávez’s return, television stations replayed American
movies, soap operas, and sports clips. One television’s news ticker read: “Venezuela recovered its Liberty.
Chávez resigned.” By and large, print media followed suit. In both their news coverage and editorial line,
major newspapers strongly favored the opposition. To chavistas, private media’s news coverage during the
failed coup seemed part of a right-wing plot to topple the government.
THE OIL STRIKE AND REFERENDUM, 2002–2004
Despite a brief moment of reconciliation after Chávez returned to power, by the end of 2002, the opposi-
tion, then organized under the “Democratic Coordinator” umbrella group, mounted a renewed effort to
8 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | I I . THE VENEZUELAN CONTEXT

pressure Chávez to resign. It orchestrated a mass work stoppage in which both white- and blue-collar
petroleum workers walked off the job—the so-called paro petrolero. The work stoppage spread to other
sectors of the economy, including food and goods production. But the government was able to restart
oil production and develop alternative channels for goods production and distribution.
Next, the opposition turned to the Constitution’s Article 72, which allowed for a referendum to
recall a president once past the midway point of his or her term in office. After delays in verify-
ing signatures on the recall petition, the vote took place on August 15, 2004. By then, oil prices
had begun a steady climb, which enabled Chávez to implement a signature social assistance pro-
grams—Las misiones, block-level programs for basic health, education, and food services. The misio-
nes galvanized support for the president, and the initiative to recall him was defeated 59–41 percent.
That same year, the Chávez government reshaped the media landscape on two fronts. First, the
National Assembly passed a Law of Social Responsibility for Television and Radio (Resorte Law). It
included bans on vulgar language and images of “psychological” or physical violence during general
viewing hours (7 a.m.–11 p.m.). It also increased the penalties for slander or statements impugning
the honor of public officials. Second, the government began funding community radio and alternative
television outlets. By the end of 2004 there were approximately 300 of these new outlets. The govern-
ment also started a new state channel, Vive TV, and bankrolled Telesur, an international channel that
Chávez hoped would compete with CNN en español.
The failure of the 2004 referendum had split the “Democratic Coordinator” group into factions argu-
ing over whether the vote had been clean or fraudulent. Ultimately, the claims of vote fraud prevailed
within the opposition parties and led them, at the last minute, to abstain from the campaign for
the 2005 parliamentary elections. The opposition allowed pro-government parties to win complete
control of the 2005–2010 Congress. A year later, the opposition returned to electoral participation
for the 2006 presidential contest. Chávez easily defeated opposition candidate Manuel Rosales, 63–37
percent, setting a historical record with 7.3 million votes. His victory speech unveiled plans to build
“Twenty-First Century Socialism” in Venezuela.
The victory had coincided with a major geopolitical event—an increase in the price of oil. Armed
with windfall revenues, Chávez embarked on three ambitious initiatives: the nation-wide expansion
of the Consejos Comunales, an experiment in state-funded grassroots governance; the construction of a
Socialist political party, the Partido Socialista Unido de Venezuela (PSUV), a major change for pro-
Chávez forces that heretofore lacked an organizational structure; and the expanded jurisdiction of the
state over the economy, a process marked by nationalizations in the telecommunications, electricity,
and energy sectors.
Concurrently, Chávez promoted a constitutional reform to be submitted for public approval through
a referendum. Most notably, it proposed eliminating term limits for the presidency but maintaining
them for governors and mayors. Held in December 2007, just a year after the presidential election,
the referendum encountered a fatigued electorate. A significant segment of Chávez’s 2006 election
coalition failed to turn out and the “No” vote won by a slim margin, 1.5 percent, delivering Chávez
his first electoral defeat. The surprising result lifted spirits in the opposition. Indeed, the outcome
held out the promise of competitive elections. Yet the loss did not slow Chávez down. He repack-
aged some of the proposals in the constitutional reform as individual laws, which the pro-government

10 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | II. THE VENEZUELAN CONTEXT
National Assembly summarily passed.
RESHAPING THE MEDIA LANDSCAPE: THE CHÁVEZ GOVERNMENT AND RCTV
Television returned to the center of national political discussion in May 2007 when CONATEL
denied Radio Caracas Television’s (RCTV) request for license renewal, sparking public demonstra-
tions. RCTV, which had opened in 1953, enjoyed a strong following among two groups: average
Venezuelans who tuned in for the channel’s popular soap operas and anti-Chávez sectors that turned to
its news programs for hardline criticism of the government. Press freedom groups took issue with the
non-renewal decision that seemed to have only targeted RCTV. Previously, CONATEL license renew-
als had tended to be routine. Other private television stations that had broadcast highly unfavorable
coverage of the government during the 2002 coup, but changed their tone afterwards, saw their li-
censes renewed for five years. In addition, President Chávez had injected his personal views about how
CONATEL should rule, which seemed to have influenced the decision. Despite the protest, a new
state-funded, public service channel, Social Venezuelan Television (Televisora Venezolana Social, TVES),
replaced RCTV on the national dial. RCTV continued to broadcast in Venezuela, but via satellite and
cable and through its Miami-based international channel RCTV International.
A new chapter in this dispute began in January 2010, when CONATEL expanded the jurisdiction of
the 2004 Resorte Law. The Commission reviewed cable broadcasters’ content to determine whether 70
percent or more of their content and operations could be considered “national,” that is, pertaining di-
rectly to Venezuela. If that were the case, the channel would come under the Resorte Law’s regulations.
CONATEL ruled RCTV International’s content did offer mostly “national” coverage, and that it thus
had to air las cadenas, the national government-sponsored broadcasts.
EXPANDED EXECUTIVE POWERS AND THE OPPOSITION’S RESURGENCE, 2008–2010
Regional elections to choose mayors and governors marked 2008, and opposition candidates suc-
cessfully built on the momentum gained from defeating Chávez’s 2007 constitutional reform. Their
candidates won key posts in metropolitan Caracas, defeating prominent chavista officials campaigning
as members of the PSUV political party. The opposition had become a force to be reckoned with, but
it was still competing on an uneven field. Constraints on its ability to compete emerged on several
fronts. In particular, seven months before the regional elections, the Comptroller’s office published
a list of politicians accused of corruption. Existing legislation authorized that office to notify the
National Electoral Council (CNE) of citizens whose public activities involved irregularities, and who
therefore should be ineligible to hold public office, in effect banning them.
The list of ineligibles comprised mostly opposition politicians, including Leopoldo López, a rising
leader with a national profile. López appealed the administrative ruling through the courts but lost.
The case became emblematic of restrictions on the freedom to run for public office.
Meanwhile, Chávez expanded the central government’s jurisdiction over the economy and its
influence over elected sub-national authorities. In August 2008, Chávez nationalized the Banco de
Venezuela, a subsidiary of Spanish banking group Grupo Santander. He also used decree powers
to create so-called civilian militias, rename the armed forces as Bolivarian, and pass legislation that
allowed him to appoint regional vice-presidents who would oversee territories and sectors previ-
ously under the jurisdiction of elected governors. In 2009, Chávez returned to the issue of term
limits. Unlike the 2007 proposed constitutional reform that would only have eliminated presidential
term limits, Chávez now proposed an amendment that would ban term limits for all elected offices.

The Constitution’s rule banning direct reelection to a third term made this issue an urgent one for
chavismo. Under the 1999 Constitution, Chávez would be compelled to leave office in 2012. Thus,
the referendum on the amendment turned into something of a plebiscite on Chávez’s rule. With the
stakes raised, pro-Chávez mobilization reached high levels. The amendment passed 55 to 45 percent,
and it seemed that Chávez would remain president for the foreseeable future.
Parliamentary elections shaped the 2010 political agenda. Before the elections, the CNE reformed
the Electoral Law to put 60 percent of seats under a first-past-the-post system of selecting candidates
instead of the proportional, closed-list system. The winners of opposition primary elections ran
united under the electoral coalition Mesa de la Unidad Democrática (Coalition for Democratic Unity,
MUD). With the “unity” strategy, MUD candidates won 65 seats out of 165, enough to block the
government from a two-thirds majority. The shares of the popular vote, though, revealed a more
evenly divided electorate. Chávez’s PSUV party won 48 percent of the vote, the MUD 47 percent,
and small parties took the rest.
In June 2011, President Chávez startled the nation with the announcement that he had cancer. The
diagnosis called into question whether Chávez could stand for reelection the following year. But in
July 2012, he claimed to be cancer free. The Governor of Miranda state, Henrique Capriles, won the
opposition primaries and ran a strong campaign, sparking enthusiasm and attracting some new voters
to the MUD coalition. But, with the government undertaking a massive public spending campaign
and Chávez running for what seemed likely to be his final term, the president won reelection on
October 6 with 55 percent of the vote. The euphoria among chavistas quickly turned to gloom,
however, as Chávez announced a month later that he would undergo a new round of chemotherapy
treatment in Cuba, raising major doubts about the future of his movement.
CHÁVEZ’S DEATH AND MADURO’S ELECTION
Chávez’s renewed health crisis ushered in a period of extreme political uncertainty and constitutional
crisis. In December 2012, Chávez shuffled foreign minister Nicolás Maduro to the vice-presidency
and then named Maduro his successor, which addressed chavismo’s crisis of leadership succession but
posed other challenges. The timing of Chávez’s death, on March 5, 2013, brought a constitutional
crisis. According to the Constitution, Chávez should have been sworn in for his new term on January
10. But Vice-President Maduro announced that Chávez was still receiving medical treatment in
Havana and would not travel to Caracas by that date. On January 9, the Supreme Court, under the
principle of “administrative continuity,” ruled that the sitting government could continue exercising
its functions without the incumbent president going through a formal inauguration in the National
Assembly. The Court’s decision also affected the procedures for naming an interim president if Chávez
passed away, as seemed imminent. An interim president would hold office while the electoral authority
organized snap elections. The Constitution’s Article 233 provides for two scenarios: the President of
the National Assembly becomes interim president if the elected president becomes incapacitated before
the new term begins, while the vice-president, an unelected official, becomes interim president if the
president is incapacitated during the first four years of his term. With its ruling nullifying the need
for a formal inauguration, the Supreme Court effectively decided that Chávez began his new term on
January 10, even though he was hospitalized in Havana at the time. Throughout the crisis, government
officials maintained that Chávez was in control of his mental faculties and still exercising presidential
power.

12 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | II. THE VENEZUELAN CONTEXT
The Supreme Court’s decision put Article 233’s second scenario into motion. Vice-President
Maduro, not National Assembly President Diosdado Cabello, became interim president on
March 5, the day government officials declared Chávez’s “complete absence” following his death.
On March 8, Maduro was sworn in as interim president to hold office while the electoral author-
ity organized elections. On April 14, 2013, Maduro faced off against 2012 presidential candidate
Henrique Capriles. Amidst accumulating signs of economic slowdown, the election proved to be
a very tight contest. Upon tallying the votes, the electoral authority declared Maduro the winner
by a margin of 1.5 percent. Capriles rejected the results as fraudulent, and called for a recount.
Heated public demonstrations after the vote threatened to turn violent. But, before major
episodes of political violence broke out, Capriles called off a rally scheduled to end in downtown
Caracas, a chavista stronghold. The electoral authority reviewed the vote and Capriles filed a
complaint through the court system. Neither process convinced Capriles the election had not
been stolen.
Six months after his razor-thin victory, Maduro faced a new electoral challenge: the December
2013 municipal elections. Capriles raised the stakes by describing these local elections as a
national plebiscite on Maduro’s government. Meanwhile, the deteriorating state of the economy
and growing shortages in basic consumer goods emerged as the biggest election issues. A little
over a month before the election, Maduro forced consumer goods prices down to what the
government called “just prices,” using the National Guard to assist state regulators in enforcing
the mark. This coincided with the start of the Christmas season and boosted the government’s
popularity just when the opposition seemed to have gained the edge. On December 8, which
Maduro had declared the “Day of Loyalty and Love to the Supreme Commander Hugo Chávez
and the Fatherland,” pro-government candidates won 54 percent of the vote. The outcome was
an important victory for Maduro and a serious setback for Capriles, whose argument that the
government lacked popular legitimacy was undermined. His leadership of the opposition came
into doubt.
A NEW MEDIA ENVIRONMENT
In 2010 the government had amended the 2004 Resorte Law covering the media, banning content
that could “incite or promote hatred . . . foment citizens’ anxiety or alter public order . . . disre-
spect authorities . . . encourage assassination . . . or constitute war propaganda.” These changes
paved the way for a more active state role in regulating media. Alleging that “democratic security”
was at stake, CONATEL cited private television and print media for violations of the Resorte Law
and, in some instances, levied heavy fines. In August 2013, two private newspapers with national
circulation, El Nacional and Tal Cual, were fined one percent of their 2009 gross incomes for
publishing photographs in 2010 that sought to expose what they said were terrible conditions at
the Caracas morgue. In 2011, CONATEL fined television station Globovisión $2.1 million dol-
lars for the channel’s coverage of a prison riot. The Commission decided that Globovisión’s report
“promoted hatred and intolerance for political reasons.” In October 2013, the attorney general’s
office opened an investigation into the newspaper Diario 2001 for its reporting on gasoline short-
ages. Earlier that year, two important media outlets had been sold: the Globovisión news channel
and the Cadena Capriles newspaper conglomerate. Within months, a number of prominent edi-
tors and reporters at each outlet resigned their positions, alleging editorial pressure.

2014: A YEAR OF DIFFICULTIES
The year 2014 started with the murder of a former beauty queen and soap opera star, and her ex-
husband, in a highway car-jacking incident. While the incident garnered international media atten-
tion because of the woman’s celebrity, the crime represented something to which ordinary Venezuelans
could relate. In response, Maduro called a meeting with all elected governors, including Capriles, to
discuss citizen security. At the meeting, Capriles shook Maduro’s hand, a gesture that seemed to con-
fer legitimacy on the government. Coordination between national and local level officials increased
somewhat, but soon fell apart when another round of heated demonstrations shook the country.
Two opposition leaders, Leopoldo López and María Corina Machado, promoted a campaign challeng-
ing the government. They called their effort “La Salida,” which can mean the Exit or the Solution.
They called for mass public demonstrations against the general state of affairs in the country, and the
drive gathered steam on the heels of news that government authorities had repressed student pro-
tests in the interior states Mérida and Táchira. The campaign’s largest demonstration took place on
February 12, National Youth Day. López led a rally that marched to downtown Caracas. As the rally
dispersed, shooting broke out and left three people dead.
Six days later, authorities detained López for trial on charges of inciting violence, an action that
spurred violent demonstrations throughout February and March. Students, opposition loyalists, and
politicians participated in rallies all over the country. Government security forces cracked down on the
protests, targeting student leaders who were sentenced to jail or placed under house arrest. The protests
eventually slowed and halted in mid-April, after the Union of South American Nations (UNASUR)
and the Vatican helped to broker talks between the government and MUD leadership. The talks
helped to calm tensions and to move political action from the street to institutional negotiations. But,
ultimately, the government and opposition failed to agree on mutual concessions. In mid-May, the
MUD’s Executive Secretary, Ramón Guillermo Aveledo, announced that the opposition was leaving
the talks, dissatisfied with the government’s unwillingness to negotiate. The talks remain frozen.
Economic issues moved to the forefront of national debate during the summer, when shortages of ba-
sic goods became dramatic. The shortages extended from food and hygiene goods to medical supplies
and automobile replacement parts. Signs of economic recession and the increasingly acute shortages
of basic goods caused a major drop in the government’s popularity. At 2014’s close Maduro’s popular-
ity was in the low 20s.
Also in 2014, the sale of the 105-year-old El Universal newspaper gave rise to renewed concern about
pluralism in the media landscape. On the heels of ownership changes at Globovisión and the Cadena
Capriles media group, the sale of El Universal generated fresh worries about the transparency of media
ownership and prospects for balanced news coverage under new ownership.

14 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | I I I . A P P R O AC H E S TO D E T E R M I N I N G M E D I A B I A S
III. APPROACHES TO DETERMINING
MEDIA BIAS
T
his study is organized around one central research question: in a politically polarized
Venezuela, did Globovisión offer balanced coverage of the most politically relevant
events during its first year under new ownership? We examine the question through
the lens of bias.
The terms balance and bias are empirically related but analytically distinct, with
subtle but important differences. A literal understanding of balance can refer to an equitable
amount of coverage of a political party or position. Alone, however, such an assessment does
not provide a thorough analysis of the quality of treatment given to competing groups or views.
Consider, for example, a two-party system in which the two sides receive equal amounts of news
coverage but the quality of the coverage varies dramatically, with one party consistently cast in
a negative light. In this sense, the issue is not balance, but how praise and criticism are allot-
ted. To capture the fact that balanced coverage is more a matter of the extent to which favorable
coverage is provided equally than the extent to which, for example, two parties receive equal
amounts of on-air coverage, this study turns to the issue of bias.
There are multiple ways to define bias and different types of bias. We chose to examine it in
terms of the degree of slant, a concept that calls attention to the extent to which coverage is
favorable or unfavorable. It is important to emphasize that bias may exist to different degrees.
For example, some studies see bias when “a systematic slant” exists (Waldman & Devitt 1998).
In our terms, this would be an example of a strong bias. Different amounts of favorable, unfa-
vorable, and neutral coverage may exist, which raises the possibility of strong, weak, or moder-
ate biases (Moriarty & Popovich 1991). Thus, bias, as used here, is not a negative pole of an
imagined standard of complete objectivity. Rather, it refers to a judgment about the character
of news coverage. Similarly, the extent of slant is our main dimension, and observations about
favorability serve as the main measures of whether coverage is slanted and to what degree.

I I I . A P P R O A C H E S TO D E T E R M I N I N G M E D I A B I A S | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 15
MANIFESTATIONS OF BIAS
To explore the nature of any bias that may exist in Globovisión’s coverage, it is important to ex-
plain how we empirically trace favorability. What are the links between coverage and favorability?
What aspects of coverage do we observe to determine favorability? What manifestations of bias
do we examine?
We consider three potential manifestations that provide a fairly comprehensive examination of
bias in terms of favorability. First is the visibility of actors—their appearance or on-air salience.
Second is the presentation and introduction of topics—how the news is framed. Third is the de-
piction of the actors in the story, which refers to the main content of the story. In the second and
third manifestations of potential bias, both visual and non-verbal aspects are assessed along with
verbal elements. Thus, in analyzing visibility, we ask how much coverage government representa-
tives receive compared to opposition representatives and whether there is greater salience for one
or the other. We look at how topics are framed by asking whether the visual and verbal presenta-
tions favor one side or another. Finally, we ask whether the visual and verbal depictions of actors
in the story treat them favorably or unfavorably.
A COMPARATIVE FRAMEWORK FOR RESEARCHING BIAS
This study’s comparative framework enables us to examine what factors may affect the appearance
of bias. We review news coverage over four temporally separate critical junctures to ask whether
coverage becomes more or less balanced over time and across the critical junctures. Since each
critical juncture introduced different issues, did any appearance of bias change in frequency or
in intensity? Additionally, the study examines three programs in depth, including weekly news
programming, week in review (Sunday) programming, and an interview program. The question
here is straightforward: if there is any bias, does it vary by program type?
RESEARCHING ON MEDIA BIAS AND POLITICAL COMMUNICATION
Media bias has long been a major area of research in political communication. Research on bias
stems from recognition that media play a significant role in shaping public opinion and political
outcomes. This role is inherent to the news medium: to inform is to make an impact. The no-
tion of media bias refers to issues specifically linked to the format and content of news messages.
Two fundamental questions are at the heart of inquiry into bias. First, do media outlets present
news content that lacks objectivity and balance? Second, under what circumstances is bias likely
to emerge? Answering these questions is not as straightforward as it may seem.
Many attempts to measure bias employ a single or comparative case framework. Such studies
have found evidence of ideological biases (Zeldes, Fico, Capenter, & Diddi 2008; Bozell & Baker
1990; Shoemaker & Reese 1991) where media outlets favored a political position. These and
similar case studies generally have two strengths. They focus on the local context of the media
landscape, and they can capture the specific ways in which media bias is likely to become mani-
fest in light of political circumstances.

Encompassing studies tell a different story. A major overview by D’Alessio and Allen (2000)
analyzed the findings of 59 studies of political bias in the U.S. media during presidential
elections.5
These studies examined bias in both print and television outlets. For all print
outlets surveyed, D’Alessio and Allen found no significant biases. For television news
they found a small degree of coverage and statement biases but judged them statistically
insignificant.6
Strictly speaking, their study does not disprove the thesis that media exhibit systematic bias.
Rather, it underscores a point commonly made in the literature on bias: because of the lack
of agreement on how to define bias, the fact that much of the research examines different
types of bias and that some research also takes into account the factor of audience reception,
it is very difficult to answer definitively whether ideological or partisan media bias exists.
In short, growth in the literature on media bias has resulted in a complex field of research
marked by multiple lines of inquiry but no underlying strong consensus.
TWO FRAMEWORKS: NEWS PRODUCTION AND NEWS RECEPTION
There are two dominant frameworks for examining the sources and mechanisms of bias.
Roughly, one considers news production and the other, news reception. The news produc-
tion approach examines the individual ideologies of news professionals, the institutional
structure of the outlet, and the journalistic principles that determine newsworthiness. The
news reception approach posits media bias to be dependent on viewers’ comprehension of
news. Therefore, bias is strongly influenced by the audiences’ prevailing political beliefs
(Ayton & Tumber 2001).
Within the news production approach, the question of agenda-setting is central. Agenda-
setting theory refers to both the salience and framing of news messages. In other words,
audiences acquire much of their views about the relative importance of political events and
the characteristics of political actors from the news content, as well as how the media pres-
ents the news (Graber 2005; McCombs, Shaw, & Weaver 1997; Soroka 2003). The theme
of salience focuses on the process of news selection. This means that audiences learn about
a particular issue and about how much importance it should be given based on whether
or not it is covered and on how much attention it receives (McCombs & Shaw 1972).
The theme of framing focuses on how stories are structured through verbal and nonverbal
information. This means that the media both informs audiences “what to think about” and
shapes “how to think about” a particular subject (McCombs 2005).
Media reception studies examine the role of perceptions, attitudes, and behavior of audience
members in their responses to news content. More specifically, reception research on media
bias considers the influences of pre-existing beliefs and cognitive processes on an audience’s
5
An encompassing study, also known as a meta-analysis study, is based on the review and aggregation of exist-
ing research on a given topic in order to utilize a larger sample size and make broader knowledge claims than
individual studies offer (D’Alessio & Allen 2000, 2012).
6
The coverage bias was detected in terms of the time allotted to Democratic candidates (52.7 percent) in rela-
tion to Republican candidates (47.3 percent). Statistically, this difference is small and most certainly would go
undetected by the audiences (D’Alessio & Allen 2010: 146–47).

perceptions of ideological bias and partisanship in the news (Arpan & Raney 2003; Vallone, Ross,
& Lepper 1985; Watts, Domke, Shah, & Fan 1999). This type of bias perception stemming from
audiences’ opinions, beliefs, and ideological dispositions has been termed the “hostile media” effect
(Gunther & Schimitt 2006; Vallone et al. 1985).
Studies of the hostile media effect suggest that it involves two distinct mechanisms—one based on
opinions and the other on cognition. With regard to the former, audiences evaluate media messages
in light of their own partisanship and what they consider a fair or accurate portrayal of the story in
relation to their position. With regard to cognition, pre-existing beliefs tend to affect how people
interpret media and what aspects of a particular story they retain and recall (Vallone et al. 1985). In
a widely cited experimental study, Vallone and colleagues (1985) found that, after viewing identical
samples of news coverage of a conflict situation, two partisan groups on opposite sides of the issue
evaluated the news segments as biased in favor of the other side. The responses suggest that the partic-
ipants watched “different” shows—in the sense that they recalled different information and disagreed
on the nature of the content viewed.
Whether due to partisanship or to varying levels of attention during exposure (Chaffee & Schleuder
1986), even “neutral” news stories on controversial issues may be perceived as biased. This is especially
true in polarized contexts marked by extreme degrees of partisanship and high tensions (Watts et al.
1999: 147; see also Gunther & Liebhart 2006). The main conclusion of reception research on media
bias bears emphasizing: Whether or not biases can be found in the media content, partisans will none-
theless perceive some degree of bias. As Gunther and Liebhart (2006: 464) contend, “partisans on all
sides of an issue will be systematically likely to criticize mass media for presenting unfavorable, slanted
or biased news coverage.”
THREE TYPES OF BIAS: GATEKEEPING, COVERAGE, AND STATEMENT
The focus of this study is on potential biases found in news content. Consequently, we are generally
working within the news production framework. Two concepts identified earlier—salience and fram-
ing—underlie much of the research on news production and point our research toward the important
mechanisms that may generate bias. D’Alessio and Allen (2000: 135–6) identify three types of media
bias: “gatekeeping bias,” “coverage bias,” and “statement bias.” In our terms, the three key processes
are the selection of news stories and topics, the use and visibility of sources, and the favorability of
verbal and visual presentation.
Gatekeeping or selectivity bias essentially concerns “information or knowledge control” (Donohue,
Tichenor, & Olien 1972: 42). Research addressing gatekeeping bias focuses on the choice of news
topics to cover. A narrowing of topics is inherent to the industry. The professional routine of journal-
ists and news producers requires them to make decisions not only about which specific news items
are relevant—according to a set of industry values and norms—but which ones are deserving of more
coverage. A number of other possible stories will necessarily be left out. The selection of topics and
construction of news stories may be influenced by the corporate structure of media outlets, internal
pressures from owners and executives, and market incentives (Entman 2007; Uscinski 2014; D’Alessio
& Allen 2000), or by the political beliefs of media professionals (Levite 1996, in D’Alessio & Allen
2000). Gatekeeping can filter the range of potential stories even before editorial decisions about
newsworthiness.

Donohue et al. (1972: 43) suggest that, beyond the news selection processes, gatekeeping biases can
also be found in the “shaping, display, timing, withholding or repetition of entire messages or message
components.” Decisions made in the process of producing, constructing, and displaying a news story
lie at the core of the gatekeeping process. For example, how many minutes should be devoted to a
particular story? In what order should it appear in the broadcast? Such questions are pivotal because
the context can determine for audiences which items are more newsworthy than others (Entman
2007). This fashioning of the news can also be linked with bias.
The second type of bias addressed by researchers is coverage bias. Generally, this is measured as the
amount of coverage granted to actors on different sides of a particular issue (Stempel & Windhauser
1989; Stovall 1985). Analyses of this kind also identify ideological imbalances in the selection of
sources, often relying on government officials (Croteau & Hoynes 1994; Reese, Grant, & Danielian
1994), and on experts and other spokespeople (Covert & Washburn 2007; Salwen 1995; Welsh,
Fenwick, & Roberts 1997). In highly polarized political contexts, for example, scholars consider that
“half the coverage should be accorded to one side and half to the other, and that deviations from this
pattern are consistent with a coverage bias of some kind” (D’Alessio & Allen 2000: 136). Such cover-
age bias can be tracked quantitatively, as has been done in this study of Globovisión.
Finally, the third type of bias, statement bias, refers to newsmakers’ explicit or implicit insertion of
their opinions into the news. Though it can vary in format, it is generally defined as a measure of
whether the coverage is “favorable” or “unfavorable” (Hopmann, Van Aelst, & Legnante 2011), or
“positive” or “negative” (Robinson & Sheehan 1983). Analysis of statement bias7
generally involves
identifying the use and frequency of positive, neutral, or negative statements (Niven 2003). In addi-
tion, facial expressions, images, and other visuals included in the story are understood to play a role in
determining the existence of statement biases. Nonverbal cues can be particularly relevant to com-
municate emotions (Banning & Coleman 2009; Ekman 1972), and research has shown that these
cues have as much influence as the verbal content, particularly on how political actors are perceived
(see also Graber 1990; Richmond, McCroskey, & Payne 1991).
LIVE COVERAGE
Live reporting, now ubiquitous in contemporary news coverage, can also be a source of news bias.
The growing availability of technology, the search for faster information delivery, and the attempt to
bring audiences “closer” to the scene have driven the increase in live reporting (Upshaw 1994; Tuggle
& Huffman 2001, 1999), sometimes at the expense of news values (Smith & Becker 1989). Scholars
are concerned that in choosing to report a particular story live, broadcasters signal that the issue de-
serves greater attention than others (Tuggle & Huffman 2001: 337; see also Morris & Nydahl 1983;
Weaver & Wilhoit 1996). Besides making the event more salient, broadcasting live from the scene
may affect the framing and accuracy of the story, as reporters have less time to gather and present
information. Scholars suggest that live coverage may also pressure leaders to respond instantly,
7
For studies that addressed media bias through examination of facial expressions of newscast, see Efron (1971); Ekman
(1972); Friedman, DiMatteo, & Mertz (1980); Friedman, Mertz, & DiMatteo (1980); Babad (1999); Babad & Peer
(2010); D’Alessio & Allen (2000).

“circumvent[ing] established avenues of political communication” and making “the process of gov-
erning more reactive than deliberative” (Tuggle & Huffman 2001: 336; see also Beschloss & Talbott
1993; Friedland 1992).
While the Internet has considerably affected news delivery, the study of live broadcasting remains a
significant aspect of political communication research. This is particularly true for Latin America,
where television is still a major source of information (IBOPE 2013).8
Despite increasing levels of
Internet penetration, broadcast television remains a central medium for news reporting and political
communication.
8
For a comparison of rates of penetration across mediums in Latin American countries, see also Latin Link. Media Penetra-
tion in Latin America. Available at http://latinlink.usmediaconsulting.com/2013/08/media-penetration-in-latin-america/
(Accessed on 27 March 2015).

20 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | I V. CO N T E N T A N A LY S I S
IV. CONTENT ANALYSIS OF GLOBOVISION’S
NEWS AND OPINION COVERAGE
T
his section of the report begins by outlining the content analysis methods used in the
study. First, it operationalizes different indicators of political bias. Second, it discusses
the selection of cases for quantitative analysis. This includes a discussion of the “critical
junctures” chosen as the focus of study, the news and opinion shows of interest, and
the types of stories and interviews within these broadcasts that are deemed relevant
for being coded for analysis. Third, it outlines the method of data collection, which entailed the
following five steps: developing the codebook and coding form, training coders, conducting pilot
tests of validity and reliability, repeating intercoder reliability tests, and assigning coders to review
specific broadcasts.
Measuring Political Bias
A. ACCESS TO MEDIA
The first dimension of political bias that is measured through the content analysis is “visibility.”
Much research on media balance has focused on the visibility of actors affiliated or associated with
different parties or political factions (Hopmann et al. 2011). This form of bias is also referred
to as “coverage bias,” and is defined by “the physical amount of coverage each side of some issue
receives” (Stempel 1969; Stempel & Windhauser 1989; Stovall 1985). Scholars generally contend
that “in a two-party electoral system, it is reasonable to assume that half the coverage should be ac-
corded to one side and half to the other, and that deviations from this pattern are consistent with a
coverage bias” (D’Allessio & Allen 2000: 136). This study proceeds from this premise to determine
whether and to what extent greater visibility is awarded to social and political actors identified with
the government of Nicolás Maduro in Venezuela or with the opposition.
Determining visibility requires four steps. First, we identify the first four actors that appear in each
story. Second, we identify up to two affiliations of each actor9
and categorize the type of actor.
The types of actors include: 1) non-elected official or government bureaucrat; 2) current or former
9
These affiliations could include the political parties or government agencies the individuals are representing, the insti-
tutions where they work, the cities where they hold office, and so forth.

I V. CO N T E N T A N A LY S I S | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 21
elected representative; 3) candidate for office or surrogate; 4) civil society group or representative; 5)
issue expert, pundit, or journalist; 6) member of the citizenry; 7) private sector representative; 8) politi-
cal party militant; 9) other. Third, each actor’s level of “visibility” is recorded. Those who are simply
mentioned are given a value of “1”; those who are quoted are counted as “2”; those who are pictured
and mentioned or quoted but who do not speak on air (e.g., image-bites) are coded as “3”; and those
who are shown while speaking or being interviewed are coded as “4.” Finally, the amount of time that
each actor is visible is recorded and used as a second indicator of visibility.
Using this information, we can compare the level and length of visibility of different types of actors and
of actors with different affiliations. In addition to comparing the visibility of explicitly political actors
associated with the opposition and the government, we compare the visibility of nominally objective—
or disinterested—actors (such as pundits, experts, or members of the public) who represent positions
favorable or unfavorable to the government. This relates to the second key goal of the inquiry into the
issue of bias: assessing the favorability of coverage.
B. FAVORABILITY OF ACTOR PORTRAYAL
To determine the second dimension of political bias, the favorability of actors presented in news cover-
age or interviewed with respect to the government or opposition, we measured several indicators. To
classify the verbal favorability of actors, coders examined the presentation or portrayal of the actor in
terms of what was said and then identified whether this was favorable to the government or not. The
researchers coded this in terms of how each actor would perceive her or his portrayal. Verbal favorabil-
ity was classified with reference to the way the hosts discussed the actor and his/her actions or, in the
case of interviews, with how the actor presented him/herself.
Non-verbal actor favorability was coded based on terms of what was shown on the screen. This could
include the gestures and facial expressions of the hosts in discussing the actor or the situation in which
the actor was involved, the choice of images and video clips, and how the actor was visually presented
during an interview.
Both the verbal and non-verbal favorability of the actors were coded as 1) favorable to the government’s
position; 2) unfavorable to the government’s position; 3) neutral with respect to the government’s posi-
tion (i.e., supports NEITHER the government nor the opposition narrative); or 4) ambivalent, denot-
ing a mix of both pro- and anti-government positions.
C. FAVORABILITY OF TOPIC PRESENTATION AND FRAMING
Another way we measure the favorability of news coverage focuses on how the topics covered in news
stories and interviews are framed and presented by the hosts and the overall broadcast. Measuring this
variable required two principal steps. First, coders identified at least one and up to four “subtopics”10
per news story or interview segment. As a rule, a topic was only included if it was introduced by the
host(s) or presented in a non-verbal way by the production of the broadcast itself. The exception to
this general rule came when a topic was introduced by the guest, but was subsequently engaged and
10
See Appendix 2.

discussed by the host(s).
Second, once at least one subtopic was identified, coders assessed the verbal and non-verbal
favorability given to each topic by the hosts and narrators. As with the coding of actor favorabil-
ity, we classified topic presentation as favorable, unfavorable, neutral, or ambivalent vis a vis the
government’s position. Failure to interject with a pro-government counterargument in response
to points made by an opposition representative was coded as neutral, rather than unfavorable.
Interjection with counterarguments that balanced the overall presentation was viewed as “am-
bivalent” (i.e. playing devil’s advocate). When presentations and interventions supported the
government’s position, this was coded as “favorable,” and when they opposed it, they were coded
as “unfavorable” (this was the case independently of whether the actors are pro-government,
pro-opposition, or disinterested parties). For interviews, the verbal presentation of the topic was
defined by the nature of the questions, statements, and arguments that introduced the discus-
sion and moved it forward. For news stories, verbal presentation was defined by the way in
which the host(s) introduced the topic, narrated throughout, interviewed people “on location,”
and so forth.
For the non-verbal presentation of each topic, topics introduced by hosts who had an indeter-
minate facial expression or did not make gestures that indicated support or rejection of claims
being made, as well as those presented with innocuous images, were coded as neutral. Failure to
show images that would undercut the government (or opposition) position was deemed neutral,
rather than favorable/unfavorable. For interviews, the non-verbal presentation of the topic was
determined based on the facial expressions and body language of the host when asking ques-
tions, making statements, and presenting arguments throughout the segment. For news stories,
non-verbal presentation included the facial expressions, hand movements, and body language
of hosts and narrators as well as the still and video images used to introduce the topic. This also
included the way a story or interview was edited, including if an interview was cut off in a way
that was favorable/unfavorable to the government position.11
D. GATEKEEPING BIAS
A third type of bias identified in the research literature is referred to as selectivity or gatekeeping
bias. This includes decisions by television stations and programs on which news stories, inter-
view guests, and topics to cover, when to cover them, and what questions to ask. Stories that re-
ceive coverage are selected from a much wider potential body of stories, guests are recruited from
a universe of candidates, and topics that receive coverage or generate questions are chosen over
others. By extension, the choice of what to select for coverage effectively deselects other topics
(D’Alessio & Allen 2000). Gatekeeping functions as a means of “information or knowledge
control” (Donohue et al. 1972: 42). In addition to decisions about selection, we understand
gatekeeping bias to also include decisions about display, timing, and withholding of information
(Donohue et al. 1972: 43).
To identify and quantify gatekeeping bias, our analysis estimated the share of total news and
opinion broadcast hours among Globovisión’s most popular shows (see below for discussion of
11
Such cuts do not impact the verbal presentation code, which was strictly based on what the hosts said.

the selection of shows) that was devoted to topics and sub-topics related to the four critical junc-
ture periods.12
Within each of the broadcasts analyzed, we selected for coding individual news
stories that included topics relevant to each critical juncture. For each critical juncture period,
we identified the number of stories, their length, and their placement in each broadcast sam-
pled. In addition to identifying which stories and interview segments covered topics relevant to
the critical junctures, our identification of sub-topics within each juncture allowed for a more
refined estimate of the topics that were covered and those that were not.
Our ability to strictly identify the presence of gatekeeping bias was limited insofar as we lacked
an objective baseline of what topics and stories were relevant for coverage and, crucially, what
precise share of broadcast time and what precise placement of relevant stories and interviews
was appropriate. That said, the estimation of a statistically valid share of broadcast time devoted
to coverage of these topics provided a crucial first step toward identifying the degree to which
Globovisión has provided sufficient coverage to the crucial issues of the day.
Case Selection
The selection of cases for analysis in this study occurs at several levels. First, we identified four
“critical juncture” periods, ranging in length from four weeks to two months, that particu-
larly exemplify the political polarization of the country. Second, we chose Globovisión’s most
watched news program and two most watched information and opinion programs as the focus
of our analysis of coverage during these periods. Third, within each critical juncture, we ran-
domly selected 25 broadcasts of these three programs for content analysis. Fourth, we reviewed
each of the sampled broadcasts, divided them into “stories” (for news programs) and “inter-
view segments” (for interview/opinion programs), and determined if the topic(s) covered fell
within the desired universe of analysis for the given critical juncture. What follows is a more
detailed discussion of this process of case selection, sample selection, and data collection unit
identification.
A. SELECTION OF THE CRITICAL JUNCTION PERIODS
This study assesses the extent to which Globovisión meets its goal of providing balanced report-
ing and analysis of Venezuelan politics. This is particularly challenging, given the deep divisions
in the country between chavismo and the opposition not only about the best direction for the
country, but also about the underlying facts. It is in this context that we aim to understand the
relative balance of Globovisión’s coverage of Venezuelan politics in general since late 2013.
Rather than randomly select sample broadcasts of Globovisión’s news and analysis over the
course of a whole year, we chose instead to sample from key, critical juncture periods. While
this might make our analysis less useful for understanding the nature of Globovisión’s coverage
in times of relative continuity and calm, these are arguably few and far between in contempo-
rary Venezuela.
In addition to representing critically important moments during the network’s first year under
new ownership, each of the four periods chosen represents a different set of topics and issues
12
The junctures are discussed above and elaborated in Appendices 2 and 3.

across which Venezuelan politics is particularly polarized. We chose to focus on these junctures
because they are moments of inflection characterized by high degrees of polarization and as such rep-
resent the kinds of moments when the significance of the media’s voice is amplified, and thus when
the provision of balanced coverage is especially important. We also focus our analysis on coverage
of these events and topics because they are microcosms of the main political tensions and issues of
contention that characterize political polarization in Venezuela.
Thus, the selection of these four junctures is important not only because they represent what we
assessed to be the most important moments in Venezuelan politics from late 2013 through mid-
2014, but also because this allowed us to assess coverage of a range of topics that were most salient
during each period. While a simple random sample of broadcasts would have given this study a
higher level of external validity—that is, provide a more pure representation of Globovisión’s cover-
age of Venezuelan politics in general—this method would not have ensured a sufficient sample of
broadcasts during these most critical periods of Venezuelan politics, nor of the most salient issues
to generate statistically valid estimates. The methodology we chose makes possible statistically valid
estimates of different types of bias during each period, while making possible comparisons of the sta-
tion’s coverage across periods and topics. We now turn to a brief description of the four periods that
are the focus of this study.
1) The Municipal Elections Period (Nov. 9–Dec. 15, 2013)
Venezuela held municipal elections on December 8, 2013. The vote took place against a backdrop
of high-intensity political contestation. The April 14, 2013 presidential elections saw opposition
candidate Henrique Capriles reject the official outcome—a victory for President Maduro by 1.5
percent. After the opposition’s appeal of the results failed to bear fruit, Capriles and the opposition
turned their attention to the municipal elections.
Until November the opposition was ahead in the polls. With one month left before the elections
and the all-important Christmas holiday looming afterwards, the government undertook policy
measures that lowered prices at consumer goods stores. Government candidates won more mayoral
offices than did the opposition, and nationally pro-government candidates won 6 percent more of
the vote. In defeat, Capriles spoke on behalf of the opposition and complained of facing highly
“unequal campaign conditions.”
2) Street Demonstrations (Feb. 12–Mar. 21, 2014)
Venezuela experienced a wave of mostly student-led demonstrations from approximately January
23–March 21. The demonstrations originated in Táchira and Mérida (interior states on the western
border with Colombia) and spread to Caracas (the capital) on February 12. They snowballed after
opposition leader Leopoldo López was taken prisoner on February 18. The demonstrations took a
variety of forms, from peaceful protest to more violent barricades and road blocks, and generated
counter-demonstrations by government supporters. State security institutions responded with exces-
sive force in some cases and in step with accepted standards of policing in others. The demonstra-
tions dissipated toward the middle of March, with fatigue and state repression undercutting their
momentum.

3) International Dimensions of Crisis (Mar. 25–May 14, 2014)
From March 25–May 14, Venezuela’s political crisis moved to the international level, with
dynamics playing out in two theaters. Within the region, UNASUR (the Union of South
American Nations) and the Vatican served as the formal conveners of political talks between the
government and the opposition. To facilitate these talks aimed at defusing the political crisis,
UNASUR sent a delegation of foreign ministers from Brazil, Colombia, and Ecuador and the
Vatican operated through its Ambassador (El Nuncio) in Caracas. In the United States, mean-
while, the Senate Foreign Relations Committee took up the issue of human rights violations in
Venezuela. The Committee held a high-profile committee meeting with the Assistant Secretary
of State for Western Hemispheric Affairs to discuss proposed legislation for imposing targeted
sanctions against Maduro government officials. The talks produced multiple rounds of discussion
but yielded few concrete gains for alleviating the crisis. When the talks failed to produce conces-
sions from the government on the issue of releasing high-profile political prisoners from jail, this
increased the likelihood for actions by the Senate and the State department.
4) Shortages in Basic Goods (Aug. 7–29, 2014)
From August 7–August 29, Venezuela’s intermittent problems with scarcity of foodstuffs and
goods reached new heights. Beyond decreased availability of food, the crisis involved decreased
availability of both over-the-counter and prescription medications and of medical supplies.
Consequently, the situation at hospitals and medical clinics moved to the front and center of
media discussion of the shortages. The government announced a major new policy measure to
combat hoarding during this period. The acute shortages dissipated but underlying supply prob-
lems continued.
B. IDENTIFICATION OF NEWS AND OPINION PROGRAMS FOR STUDY
To further focus our analysis, we identified thirteen news or opinion programs that regularly
covered contemporary political topics of interest to this study. Rather than randomly select
broadcasts from these thirteen programs, however, we chose to focus on those with the highest
viewership, while seeking representation of both traditional news programming and information
and opinion (interview-based) programming. The programs chosen for this study were Noticias
Globovisión Estelar, the network’s flagship daily news show; Aló Venezuela, a weekly interview and
news program; and Vladimir a la Una, a long-form interview program that airs every weekday.
We ranked the popularity of all thirteen shows with relevant political news content using the
audience levels—number of individuals watching—for selected week-long periods within
each juncture.13
The thirteen shows were ranked based on the average number of viewers per
broadcast-hour over the course of a selected week. We selected week-long periods since some
shows aired every day and others only once a week. The “average” number of viewers for shows
that only aired once a week—such as Aló Venezuela—was based on the per-hour average of that
broadcast. The ranking of shows that aired every day was based on the average number of viewers
per-broadcast-hour for all seven days.
13
No ratings data was available for the “Municipal Elections” period.

To determine the most popular shows during these periods, we focused on one-week periods from
February 12th through 18th, April 1st through 7th, May 16th through 22th, and August 7th through
14th. There were no ratings data available during the November–December 2013 juncture. After
ranking the thirteen shows within each of the four one-week periods, we calculated the average of
the four ranks. The most popular news show, and the most popular of all shows by a considerable
margin, was Noticias Globovisión Estelar (avg. rank = 1.25), which airs from 9:00 to 10:30 p.m. every
night. The two most popular information and opinion programs were Aló Venezuela (avg. rank
= 4.25), which airs every Sunday, followed by Vladimir a la Una (avg. rank = 5), which airs every
weekday.
C. SAMPLING OF BROADCASTS
The sampling frame was thus determined based on 1) the four week-long periods and 2) the three
programs chosen for study based on viewership. From this, randomly selected twenty-five broadcasts
within each period. This random selection was done using a three-step process in the OpenOffice
Calc application. First, researchers listed all broadcasts of the three selected programs in chronologi-
cal order over the entire period. Second, researchers used the random number generator function,
“RAND(),” which generates a random number between 0 and 1 for each broadcast. Third, research-
ers sorted all broadcasts in ascending order based on these random numbers and selected the first
twenty-five broadcasts within each critical juncture. The sample included eight broadcasts of Aló
Venezuela, sixty broadcasts of Noticias Globovisión Estelar, and thirty-two broadcasts of Vladimir a la
Una.
D. DATA COLLECTION
Once the sample of broadcasts was selected as outlined above, each broadcast was divided into news
stories or interview segments. These were the data collection units, as distinct from sampling units
(Neuendorf 2002: 72). In addition to identifying the start and end time of each news story or inter-
view segment, we identified whether each unit included discussion of a topic or subtopic relevant to
the period under study (i.e. municipal elections, protests, international dimensions of the crisis, and
shortages). Those stories and segments not covering a relevant topic were not coded for indicators of
bias.
A detailed codebook was developed to generate the relevant data for the quantitative analysis on the
framing of key topics by the hosts and the station, visibility of different types of actors, and the rela-
tive favorability of these actors’ presentation and portrayal vis a vis the government’s position. The
codebook was developed, tested, and revised over several coder training sessions and group and indi-
vidual pilot coding tests. The research team conducted a total of five independent coding pilot tests
of a selection of news and opinion content from all of the Globovisión programs of interest within the
periods of the study, but outside of the selected sample of broadcasts. After each of the first five pilot
tests, the four coders met together with one or both of the lead researchers to review points of confu-
sion, disagreement, and agreement, and to further refine the instrument and clarify the operational-
ization of the variables. Based on feedback from these meetings, the lead researchers made revisions
to variables, their operationalization, and added clarifications with respect to how certain pieces of
content should be coded. At this point the codebook and accompanying coding form were finalized
(see Appendix 1).

After these independent, off-sample pilot tests were completed, the lead researchers con-
ducted two paired-independent coding exercises using broadcasts that were sampled for
the study. As acceptable levels of reliability had not yet been reached, pairs of coders met
with each other to identify and reconcile any points of disagreement, and provide the lead
researchers with finalized data for those broadcasts.14
It was determined that reliability levels
were sufficient to proceed to full, independent coding at this point, and the remaining 79
of 100 sampled broadcasts were assigned randomly to the four coders, based on the number
of work hours each was free to dedicate each week. To calculate final reliability statistics, 15
broadcasts were coded by two coders. This reliability sample of 15 broadcasts consisted of
296 data collection units (stories or interview segments), 23 percent of the remaining data
collection units in the sample.15
Any disagreements between coders in the final reliability
sample were reconciled by randomly selecting the codes of one of the coders for each variable
(Lombard, Snyder-Duch, & Bracken 2002: 601).
14
Importantly, the bulk of these disagreements are trivial in the sense that they are related to where in the
coding spreadsheet a given coder lists a particular actor rather than disagreement about the type, visibility, or
favorability of the actor to the government’s position.
15
Final reliability for the key variables was calculated using “ReCal: reliability calculation for the masses,” cre-
ated by Deen Freelon of American University, and are reported in Table A-1 in Appendix 5, Tables and Figures
(see Freelon 2010).

28 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | V. E M P I R I C A L A N A LY S I S A N D F I N D I N G S
W
e now present the analysis and findings regarding different types of
bias in the news, information, and opinion broadcasts of Globovisión.
The quantitative content analysis seeks to identify and measure bias
reflected in visibility of pro-government and opposition voices and
the presence of any favorable or unfavorable slant in the treatment,
presentation, and framing of competing perspectives. In addition to this favorability
dimension of bias, we examine a third type of bias, known as gatekeeping. We do this
by using the data gathered to estimate the share of broadcast coverage that focused on
key topics of national importance during each of the periods under study. The programs
included in the sample were the most popular news program Noticias Globovisión Estelar
and the two most popular opinion programs Aló Venezuela and Vladimir a la Una. These
estimates provide evidence of whether Globovisión exercised bias in selecting topics to
cover.
This section is divided into three parts. The first presents descriptive statistics on different
measures of each dimension of bias (visibility, favorability, and gatekeeping). In addition,
this part identifies variations in bias across the four discrete periods under study. The
second part presents a comprehensive index of bias, which combines estimates of visibility
and favorability bias, and identifies key factors associated with relatively more (and less)
bias in favor of the government. The third and final subsection discusses implications of
the results.
Indicators of Bias: Descriptive Statistics and Comparisons
A. ACCESS TO MEDIA: VISIBILITY
The first set of estimates in this study identifies bias with respect to the access afforded to
individuals associated with the opposition compared to those associated with the govern-
ment and those who are not partisan. We used two distinct measures of “visibility.” The
first consisted of a scale of actor visibility, ranging from 1 to 4, with 4 being the highest
V. EMPIRICAL ANALYSIS AND FINDINGS

V. E M P I R I C A L A N A LY S I S A N D F I N D I N G S | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 29
level of visibility and 1 being the lowest.16
The second indicator of visibility measured the amount
of time (in minutes) that each individual actor is “visible” (i.e., is mentioned or discussed, quoted,
shown in images, and mentioned or shown while speaking).
To identify the presence of any partisan bias, we classified the individuals appearing in stories or being
interviewed as pro-government partisans, pro-opposition partisans, or nonpartisan.17
We defined par-
tisanship narrowly, such that actors were only considered partisan insofar as they directly participated
in electoral politics in Venezuela. Thus, individuals such as U.S. Vice-President Joe Biden or Bolivian
President Evo Morales were not classified as partisan, nor were actors from such organizations as the
National Association of Journalists, the National Electoral Council (CNE), or unaffiliated political
analysts.
Overall, nonpartisan actors were most commonly featured in coverage (45.3 percent) and pro-
government and pro-opposition partisans were equally likely to receive attention (28.4 and 26.3
percent, respectively). Regardless of partisanship, individuals appeared most often while speaking or
being interviewed.18
As illustrated in Figure 1, of all individuals who were interviewed or shown while
speaking, 45.7 percent were nonpartisan, 27.4 were pro-government, and 26.9 were pro-opposition.
Nonpartisans were the largest group regardless of how visibility was measured. At the lowest level of
the visibility (“Mentioned”) and the third level of visibility (“Shown while Mentioned or Quoted”),
pro-government figures were somewhat more represented than their opposition counterparts, but
opposition voices were more likely to be quoted. Globovisión’s coverage achieved noteworthy balance
with respect to the average amount of time allotted to pro-government (1.6 minutes) and and pro-
opposition voices (1.97 minutes).
16
Specifically, actors who were mentioned or discussed by the hosts were categorized as “1”; those who were quoted were
counted as “2”; those who were shown in the picture and mentioned or quoted but not speaking (e.g., image-bites) were
coded as “3”; and those who were shown while speaking or interviewed were coded as “4.”
17
In order to ensure the independence of our classifications of visibility and favorability, we classified actor partisanship
only after completing the content analysis. These classifications were made based on the affiliations of each actor as re-
ported through coverage (e.g. the political party they represented), based on coders’ own expertise, and through secondary
internet research. Finally, in some cases where these classifications were not clear, researchers consulted specific video foot-
age to identify the partisanship of actors.
18
There was no difference in the average qualitative visibility of the three actor types.
42.9
32.9
24.3
54.1
18.9
27.0
43.4
35.4
21.2
45.7
27.4
26.9
020406080100
PercentofActors
Nonpartisan Pro-Government
Pro-Opposition
Mentioned Quoted Shown Shown Speaking
Figure 1. Visibility Scale by Actor Partisanship

Figure 2. Minutes Visible by Actor Partisanship
As shown in Figure 2, partisans tended to be visible for considerably longer than nonpartisans (0.86
minutes). Another way to examine the relative access to the media of different types of individuals
is by calculating the number of minutes in which they were visible. Table 1 shows the total number
and percentage of broadcast minutes that nonpartisan, pro-government, and opposition actors were
visible. The results show balance, with those from the opposition receiving a slightly larger percentage
of coverage (37.5 percent) than those associated with the government (33.6 percent) or nonpartisan
actors (28.9 percent).

Table 1. Total Number of Minutes Visible
Thus, although considerably more individuals who appeared on these programs were classified as
nonpartisan, these actors were visible for less time. Among partisans, pro-government actors were no
more likely to appear, and did not appear for longer periods of time, when compared to those who
were pro-opposition. Furthermore, pro-government voices did not have a higher level of visibility
compared to others. However, these trends vary across the periods that we measured. Figure 3 shows
the average number of minutes each perspective was made visible within each critical period.19
During
coverage of “Municipal Elections” and “Shortages of Basic Goods,” the average opposition actor was
visible for a full minute more than the average pro-government counterpart. The average visibility of
19
The qualitative indicator (range 1–4) did not vary significantly across the four periods, so it is not plotted on this graph.
0.86
1.60
1.97
0.511.52
Average
Nonpartisan Pro−Government Pro−Opposition

Nonpartisan Government Opposition
N % N % N %
Municipal Elections 58 16.4 87 24.7 208 58.9
Street Demonstrations 169 25.7 250 38.1 238 36.3
International Dimensions 72 66.4 25 22.9 12 10.8
Shortages of Basic Goods 98 38.4 100 39.1 57 22.5

All 396 28.9 461 33.6 515 37.5

nonpartisan actors was higher than that of partisans only during coverage of international engagement
with Venezuela, and much lower during the other three junctures we analyzed. The magnitude of
these differences can be seen in Table 1.
0.60
1.43
2.59
0123
Average
Municipal Elections
0.79
2.49
0123
Average
Shortages of Basic Goods
1.56
1.67
1.23
0.78
0123
Average
International Dimensions of the Crisis
Figure 3. Nonpartisan, Pro-Government, and Pro-Opposition Actors
0.85
1.75 1.66
0123
Average
Street Demonstrations

B. FAVORABILITY OF ACTORS PORTRAYED
The second set of estimates of political bias relate to favorability, measured in terms of whether voices
featured in news, information, and opinion programs express views that favor the government’s position.
Both verbal and non-verbal favorability are assessed, and we measure whether the first four actors that
appeared in a story or segment are 1) favorable, 2) unfavorable, 3) neutral, or 4) ambivalent toward the
government.
We analyzed favorability 1) at the actor level, 2) at the story/segment level,20
and 3) at the broadcast
level. Reporting favorability of entire stories and broadcasts, in addition to the discrete ratings of indi-
viduals, is important because it provides measures of bias in the messages that viewers perceive. That is,
a viewer’s perception of the messages occurs at the level of a whole news story, and sometimes a whole
broadcast.21
For example, a story could be unbiased on net if it presented equal numbers of perspectives
favorable and unfavorable to the government, or if all perspectives were neutral.22
1. Actor-Level Analysis

Critical Juncture
1 2 3 4 All
N % N % N % N % N %
Favorable 64 26.1 170 33.9 24 29.6 68 31.5 326 31.2
Unfavorable 103 42.0 234 46.6 25 30.9 87 40.3 449 43.0
Neutral 78 31.8 83 16.6 27 33.3 55 25.5 243 23.3
Ambivalent 0 0.0 15 3.0 5 6.2 6 2.8 26 2.5

Total Actors 245 100 502 100 81 100 216 100 1,044 100
Breaking the data down by actors gives as much weight to a voice that appears in a news story for 5 or
10 seconds as to someone who is interviewed for an entire 10- or 15-minute segment, or to an individu-
al who appears at the beginning or the end of coverage. Table 2 shows the data on the verbal favorability
of individuals who appeared in each of the four periods we studied. The most important finding here
is that the portrayal of actors was considerably more likely to be unfavorable to the government posi-
tion. The “unfavorable” category was most common in all instances except for those focusing on the
“International Dimensions of the Crisis.” On that topic, “neutral” was the most common characteristic
of coverage. Overall, the portrayal of actors was more likely to be unfavorable to the government (43.0
percent) than favorable (31.2 percent), with about 1 in 4 being categorized as either neutral (23.3 per-
cent) or ambivalent (2.5 percent).
During the period of the “Municipal Elections,” more than two out of every five actors (42 percent)
were portrayed in a way that was unfavorable to the government, more than a quarter (26.1 percent)
20
As was described in the Case Selection section, the unit of data collection was the story or segment, and data were collected
on up to four topics, and up to four actors within each data collection unit.
21
The exception of an individual interviewee who is the only person in a particular segment proves the rule. Technically
speaking, the favorability of this person toward the government at the actor level would be equivalent to that at the story
level.
22
A similar logic is at play when comparing the net favorability of an entire broadcast with that of its component stories and
individuals represented.
Table 2. Favorability of Individuals toward the Government (Verbal)

were presented favorably, and 31.8 percent were classified as neutral.23
There was a similar favorability
bias away from the government position during the period of “Street Demonstrations” (February 12–
March 21, 2014). During this period, 46.6 percent of actor portrayals were classified as unfavorable,
compared to 33.9 percent favorable and only 19.6 percent either neutral or ambivalent.
As noted, the exception to the rule that actor portrayals lean against the government was found in our
analysis of the “International Dimensions of the Crisis” (March 25–May 14, 2014), where there were
no significant differences. During this period, one in three (33.3 percent) was neutral, similar to the
share that was favorable (29.6 percent) or unfavorable (30.9 percent) to the government. The trend
of bias against the government continued during the fourth period of study on “Shortages of Basic
Goods” (August 7–August 29, 2014). Specifically, 40.3 percent were coded as unfavorable, compared
to 31.5 percent coded as favorable and 28.2 percent as either neutral (25.5 percent) or ambivalent (2.8
percent).

Critical Juncture
1 2 3 4 All
N % N % N % N % N %
Favorable 31 12.6 88 17.7 13 16.1 31 14.4 163 15.7
Unfavorable 34 13.8 137 27.6 5 6.2 19 8.8 195 18.8
Neutral 181 73.6 270 54.1 61 75.3 165 76.4 677 65.0
Ambivalent 0 0.0 4 0.8 2 2.5 1 0.5 7 0.7

Total Actors 246 100 499 100 81 100 216 100 1042 100
Table 3 shows statistics for the non-verbal favorability of actors. In notable contrast with the findings
for the verbal portrayal of actors, the non-verbal portrayal was most likely to be classified as neutral.
Across all four junctures we analyzed, about two-thirds of actors were portrayed in a neutral fashion
(65.0 percent), compared to 18.8 percent as unfavorable and 15.7 percent favorable to the govern-
ment. These patterns roughly held across time and theme, with neutrality being the most common
throughout. Individuals featured in coverage of street demonstrations were considerably less likely to
be non-verbally neutral; 54.1 percent compared to roughly three-fourths for the other three junctures
we analyzed.
2. Story- and Segment-Level Analysis
To provide a slightly different way of analyzing how actors are portrayed, we also created measures of
verbal and non-verbal actor favorability at the level of the news story/interview segment. For stories or
segments with multiple actors, the measures of favorability combine the favorability of all actors (1–4)
who appear. To capture this segment-level actor favorability, we create an indicator of “net actor favor-
ability,” calculated by taking the difference between the number of actors portrayed in such a way that
was favorable to the government and the number portrayed in a manner unfavorable to the govern-
ment (range –4 to 4).
23
During this juncture, no actor was identified as having been ambivalently portrayed.
Table 3. Favorability of Individuals toward the Government (Non-Verbal)

Table 4. Net-Favorability of Actors in Stories

Mean by Critical Juncture Period

Verbal Non-Verbal N
Municipal Elections -0.22* 0.02 144
Street Demonstrations -0.28* -0.20* 270
Int’l Dimensions -0.04 0.13 55
Basic Goods Shortages -0.11 0.10 133
All -0.20* -0.05 602

* Statistically different from 0 at the 0.05 level, two-tailed.
Note: 0 means no bias. Positive numbers indicate Pro-Government bias and negative numbers
indicate Pro-Opposition bias.
The results presented in Table 4 are consistent with what was found in the actor-level analysis.
Across all the periods we studied, the net verbal favorability at the level of the news story or
interview segment was shaded against the government (–0.20). Our estimates of non-verbal
net-favorability did not find bias overall, but there are some important differences across peri-
ods that need to be highlighted.
The coverage of street demonstrations was unfavorable to the government on both the verbal
and non-verbal dimensions, as was coverage of the municipal elections on the verbal dimen-
sion. Figure 4 depicts the distribution of the verbal and non-verbal favorability of actors, and
shows these for the street demonstrations and all other periods separately. One thing these fig-
ures reflect is a large number of stories/segments that are neutral (having net-favorability close
to 0), and this tendency is even more pronounced when considering the non-verbal scores.

0100200300400
Frequency
−4 −2 0 2 4
Pro−Opposition Pro−Government
All Stories: Non−Verbal
0100200300400
Frequency
−4 −2 0 2 4
All Stories: Verbal
050100150
Frequency
−4 −2 0 2 4
Street Demonstrations: Verbal
050100150
Frequency
−4 −2 0 2 4
Street Demonstrations: Non−Verbal
050100150200250
Frequency
−4 −2 0 2 4
All Other Periods: Verbal
050100150200250
Frecuencia
−4 −2 0 2 4
All Other Periods: Non−Verbal
Figure 4. Net-Favorability of Actors at the Story/Segment Level of Analysis
Street Demonstrations, Others, and All Periods

3. Broadcast-Level Analysis
To offer a slightly different way to examine the favorability of how actors are portrayed, we present
summary measures of favorability by hour-long broadcast.24
As shown in Table 5, no bias was found in
the favorable or unfavorable direction at the broadcast level. A noteworthy difference appeared in the
non-verbal favorability in coverage of the “International Dimensions of the Crisis.” Here, non-verbal
coverage was on average favorable to the government.
C. TOPICS OF COVERAGE AND GATEKEEPING
This subsection presents data on the range of subtopics covered within each critical juncture and on the
share of broadcasts that did not cover the target topic during each period of study. For each broad-
cast, stories and segments within them were identified as covering the target topic of each juncture:
1) Municipal Elections, 2) Street Demonstrations, 3) International Dimensions of the Crisis, and 4)
Shortages of Basic Goods. Within each of these target topics, we identified several sub-topics, and each
story was coded as covering at least one and up to four of these topics.
24
These variables were calculated simply by taking the average of net-favorability scores at the segment level across the full
hour of each broadcast we coded.

Mean by Critical Juncture Period

Verbal Non-Verbal N
Municipal Elections -0.27* 0.00 21
Street Demonstrations -0.18* -0.11 25
International Dimensions 0.10 0.18* 21
Basic Goods Shortages -0.05 0.06 21
All -0.10 0.00 88
* Statistically different from 0 at the 0.1 level, two-tailed.
Note: 0 means no bias. Positive numbers indicate Pro-Government bias and negative numbers
indicate Pro-Opposition bias.
Table 5. Net-Favorability of Actors in Broadcasts

Figure 5. Percent of Broadcasts Covering Target Topic by Critical Juncture Period
Across all of the broadcasts in the three programs analyzed, 602 stories or segments touched on the rel-
evant topic and 1,030 did not. The variation in frequency of “on-topic” segments between the four peri-
ods is noteworthy; ranging from only 12 percent of stories focusing on “International Dimensions of the
Crisis” during the mid-late spring of 2014 to more than three-fifths focusing on “Street Demonstrations”
earlier that year (See Figure 5).
Statistical findings for each of the critical junctures we analyzed are presented in Figure 6 (see also Tables
A-2 and A-3 in Appendix 5). During the period of the municipal elections, the most common subtopic
covered was the “Economic Situation” (30.2 percent). Some 45 percent of topics covered focused on
different aspects of the elections themselves: the campaign, 26.0 percent; party leader reactions to the
election results, 14.9 percent; electoral institutions, 10.7 percent; and turnout and results, 5.3 percent.
During the period of “Street Demonstrations,” it was most common for stories to focus on “Violent
Anti-Government Demonstrations” (20.3 percent), followed by “Peaceful Anti-Government
Demonstrations” (16.9 percent), “General” coverage of demonstrations (15.4 percent), and “Police and
Security Forces” (10.7 percent). “Pro-Government Counter-Demonstrations” only made up 4.4 percent
of topics covered.
As noted above, the broadcasts sampled during the period of the study devoted to the “International
Dimensions of the Crisis” were considerably less likely to contain stories or interview segments that
N=310
46.5%
53.5%
Municipal Elections
N=465
11.8%
88.2%
N=436
64.4%
35.6%
N=421
31.6%
68.4%
Target Topic Not Target Topic

touched on our target topic. Within this smaller number of stories, the plurality focused generally on in-
ternational dimensions of the crisis, rather than on any of the specific subtopics identified (26.6 percent).
Of the subtopics of interest, the most commonly covered were the Union of South American Nations
(UNASUR) and the Vatican, both of which were key players in external attempts to launch talks to resolve
Venezuela’s domestic political conflicts.
Finally, our analysis identified, unsurprisingly, that the most common subtopics covered during the period
where we focused on “Shortages of Basic Goods,” during August 2014, were the “Economic Situation”
(29.7 percent) and the “Economic Policy Context” (21.9 percent). There was also a considerable amount
of coverage of the “Fingerprint System,” or Captahuellas (12.9 percent), and “Hospitals and Health
Clinics” (9.7 percent).
D. FAVORABILITY OF TOPIC PRESENTATION AND FRAMING
In addition to classifying the favorability of the portrayal of actors who appear in stories, we similarly clas-
sified whether the topics covered in news stories and interviews were framed and presented by the hosts in
a way that was favorable to the government. As with actor favorability, we classified the verbal and non-
verbal framing and presentation of the topics covered.
0.02
0.02
0.02
0.03
0.03
0.05
0.05
0.08
0.13
0.14
0.19
0.27
N=64
0 .1 .2 .3
Share of Int’l Dimensions Subtopics
General UNASUR
Vatican Other
U.S.−Venezuelan U.S. Sanctions
OAS Venezuela−South America
Venezuela−Brazil Venezuela−Colombia
United States Cuba
0.04
0.05
0.09
0.11
0.15
0.26
0.30
N=169
0 .1 .2 .3
Share of Election Subtopics
Municipal Elections
Economy Campaign
Reactions to Elections Electoral Institutions
Results General
Other
0.04
0.06
0.08
0.09
0.09
0.11
0.15
0.17
0.20
N=385
0 .1 .2 .3
Share of Demonstration Subtopics
Violent Protests Non−Violent Protests
General Police/Sec. Forces
Judicial Process Other
Int’l Actors Human Rights
Counter Protests
0.05
0.05
0.08
0.09
0.10
0.13
0.22
0.30
N=155
0 .1 .2 .3
Economic Situation Economic Policy
Captahuellas Hospitals and Health
Other General
Cesta Básica Int’l Actors
Figure 6. Subtopics Covered by Critical Juncture Period

What is most noteworthy here is that neutrality is the norm, with more than three-fifths of topics cov-
ered being categorized as verbally neutral (61.8 percent) and 80.3 percent categorized as non-verbally
neutral. Figures 7 and 8 show the verbal and non-verbal favorability of topic presentation for each of
the four periods we analyzed. Most importantly, topics were equally likely to be framed for or against
the government, both verbally and non-verbally and in each period. The percentage of topics framed
either favorably or unfavorably was very low for non-verbal favorability. The only exception to this
trend was in coverage of street demonstrations, where 14 percent of topics were portrayed in a way that
was favorable to the government and 14 percent were unfavorable. The overarching conclusion here is
that unbiased topic presentation was the norm, with the bulk of topics being neutrally presented, and
in those that were not, pro- and anti-government presentations balanced each other out.
71
11
14 14
60
17
14 14
63
13 13 13
58
16
13 13
0102030405060708090
PercentofSubtopics
CJ−1 CJ−2 CJ−3 CJ−4
Neutral Ambivalent
Favorable Unfavorable
Figure 7. Verbal Favorability of Topic Presentation
85
8
4 4
74
6
14 14
89
3 2 2
89
1
3 3
0102030405060708090
PercentofSubtopics
CJ−1 CJ−2 CJ−3 CJ−4
Neutral Ambivalent
Favorable Unfavorable
Figure 8. Non-Verbal Favorability of Topic Presentation

Measuring Overall Bias and What Explains It
COMPREHENSIVE INDEX OF BIAS
In this section we present a total bias score for each news story. This is based on a comprehensive
index which combines the different indicators of visibility and favorability of the actors portrayed in
news coverage as well as the favorability of topic presentation and framing. This provides a total bias
score that simultaneously reflects the different dimensions of media bias in a single indicator. We cre-
ated this comprehensive indicator by combining the following variables at the level of the news story:
1) Visibility Type, 2) Minutes Visible, 3) Actor Verbal Favorability, 4) Actor Non-Verbal Favorability,
5) Verbal Topic Presentation, and 6) Non-Verbal Topic Presentation.25
This analysis creates a single
measurement of bias that incorporates all of the different dimensions of bias analyzed so far. Next,
we conducted an analysis to identify the factors that are associated with different levels of bias as
measured by this comprehensive index.

Figure 9 shows the plots the comprehensive index variable. The height of each bar represents the
number of stories at each value of the index and stories with higher values on the index were more
biased in favor of the government relative to other stories.26
In addition to estimating the degree
to which bias varied across the periods analyzed in this study, we examined variation across time.
Additional variables of interest include the sampled programs (Aló Venezuela, Vladimir a la Una, and
Noticias Globovisión Estelar), however, these programs had similar levels of total bias. In addition, we
measured the order in which each story appeared within each broadcast. Analyzing how the prioriti-
zation of the story was associated with the direction of bias allowed us to show that pro-government
stories tended to be prioritized at the beginning or relegated to the end of broadcasts.27
25
The precise method used to calculate this index variable is described in the Methodological Appendix.
26
Note that the factor analysis technique sets the mean at 0. Figures A-1–A-6 in the appendix show the distribution of the
six indicators of bias that are the component parts of the comprehensive index.
27
Descriptive statistics for all of these variables, as well as the component parts of the bias index variable, are included in
Table A-4 in Appendix 5.
050100150
Frequency
−4 −2 0 2 4
Index Value
Frequency Distribution
Figure 9. Comprehensive Index of Bias

We identified through statistical analysis a couple of noteworthy findings that build on the descriptive
analysis of the different types of bias. First, stories and interviews that occurred during the “Municipal
Elections” period and the weeks of “Street Demonstrations” tended to be less favorable to the govern-
ment. Second, in contrast to the general findings of an absence of bias in favor of the government,
and the occasional bias in favor of the opposition, we find evidence here that the placement of stories
favored the government.28
Figure A-7 in the Appendix 5 presents four scatter plots (one for each period we studied), which
show the correlation between bias and story prioritization.29
The negative correlation between them
indicates that stories that are relatively more favorable to the government tend to be presented earlier
in broadcasts. The mild downward slope depicted in Figure A-7 is visible in all four critical juncture
periods. It should be emphasized, however, that the substantive significance in this difference is rela-
tively minor.30
Thus, if anything, this result should serve as a very minor exception to a general rule
that systematic bias was not present in the coverage of Globovisión’s most popular news and opinion
programming during the critically important periods of political polarization examined in this study.
28
See Ordinary Least Squares (OLS) results in Table A-5 in Appendix 5.
29
This variable, which we call “Story Number,” simply indicates where each story appeared in the sequence of an hour-long
broadcast.
30
Specifically, moving from the latest story (Number 29) to the first story (Number 1) would imply an expected movement
down the Index of Bias of less than 1, about 1/8th of the total range of the index.

42 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | V I . CO N C LU S I O N
T
his study set out to determine the degree of balance in news coverage provided
by Venezuelan television station Globovisión. To answer this question we con-
ducted a rigorous content analysis of 88 hour-long broadcasts, including some
600 news stories or interview segments featuring more than 1,000 individual
appearances and discussion of 36 different topics. We focused exclusively
on whether there was evidence of bias in coverage during four critical junctures, span-
ning from November 2013 to August 2014, that were deemed best suited to determining
whether Globovisión programming provides fair coverage of major political events and
controversies.
NEUTRALITY AND BALANCE
The study found that Globovisión’s framing of the issues tended to be neutral, and that
there was no significant bias in favor of the government or the opposition. While overall
Globovisión achieved its goal of providing unbiased coverage, we found modest variation
across the periods and issues covered as well as across particular measurements of bias.
Specifically, opposition voices received more total coverage than pro-government voices in
the periods surrounding the municipal elections and the street demonstrations, and this
coverage tended to be unfavorable to the government. However, pro-government perspec-
tives received slightly more coverage in broadcasts focused on international dimensions
of political crisis or shortages of basic goods. The comprehensive analysis of the different
dimensions of bias also found that—although opposition partisans received more on-air
time overall—news stories that appeared earlier in a broadcast tended to be relatively more
favorable to the government, whereas those appearing toward the end of a broadcasts were
more favorable to the opposition.
The placement of news stories is important since it speaks to the priority an outlet assigns
to the topic. Stories reported earlier in a program are what the channel “leads with” in
prioritizing its presentation of the news. This study found a pattern in sequencing that is
VI. CONCLUSION

V I , CO N C LU S I O N | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 43
likely to hold significance. News stories appearing earlier in a broadcast tended to be more favor-
able to the government than those appearing toward the end. Because we were unable to to secure
access to data to determing whether audience levels fluctuated over the course of each program, this
prevents us from assessing the full significance of these tendencies in Globovisión presentation of the
news. The absence of data on viewership also prevents us from reaching conclusions about the signif-
icance of the finding that coverage tilted modestly in favor of the opposition during the elections and
street demonstrations, and modestly in favor of the government when the focus was international
engagement with Venezuela’s political crisis or with shortages of consumer goods.
In the absence of information about viewership, this study’s data consisted exclusively of specific
footage supplied by the station at the request of the research team. Analysis of that information
enables us to establish that Globovisión did not strongly favor the government position by providing
pro-government voices significantly more visibility or by presenting them in a more positive light
than their opponents. This runs counter to widely reported opposition sentiment during this period
regarding the supposed capture of Venezuelan media by the government. The “hostile media” effect,
a phenomenon identified in the “Approaches to Determining Bias” section of this report, may be
relevant here. The hostile media effect refers to how partisan perceptions of what constitutes a fair
portrayal of events can predispose audiences to believe that a media outlet favors a contrary political
position. This effect powerfully shapes how audiences interpret news and helps to explain the preva-
lence among government opponents of a conception that Globovisión is a surrogate for official posi-
tions. Moreover, the hostile media effect is an especially relevant phenomenon in polarized political
contexts such as that found in Venezuela today. With Venezuela experiencing high levels of polariza-
tion since at least 2002, and Globovisión’s previous editorial line giving the channel a pronounced
anti-government slant, much of Globovisión’s audience inevitably holds views about the station’s
editorial line under the ownership group that bought the channel in June 2013. Nonetheless, this
study finds that expectations of a shift from support for the opposition to strongly favoring the gov-
ernment is not evidenced by rigorous empirical analysis.
While we did not assess the quality of news coverage during this period in terms of the depth of
investigative reports, the quality of sources, or the checking of facts, the findings also suggest that
Globovisión programming often achieved a degree of neutrality with its news coverage while also
providing a space for government, opposition, and neutral voices during important moments. This
conclusion is especially important for assessments of balance.
Moreover, the correlation between visibility and favorability—for example, in coverage of street
demonstrations in early 2014, in which there were more appearances by opposition figures and more
news segments unfavorable for the government—provides quantitative confirmation for an observa-
tion derived from the qualitative review of Globovisión news programming. In reviewing footage,
researchers noted the tendency of Globovisión’s reporters to provide both street- and studio-based
interview guests significant time and latitude to answer questions. The non-confrontational, open-
ended approach of Globovisión journalists afforded partisan guests ample opportunity to express
themselves clearly and without obstacle. The manner in which Globovisión frames and presents the
issues of the day suggests that, both quantitatively and qualitatively, the hosts and production teams
at the stations have largely succeeded in producing unbiased programming in the areas that are argu-
ably most directly under their control.

Two characteristics of the local context in Caracas are useful for thinking through some prac-
tical reasons why Globovisión might have provided greater space for opposition voices during
particularly critical moments. First, generally speaking, pro-government politicians use pub-
lic or state-owned media outlets to increase their exposure, promote their images, and articu-
late their positions. It is less frequent for pro-government politicians or government officials
to utilize private media outlets for these purposes, and their preference for using state-owned
media may affect the size of the pool of potential recruits for private media outlets such as
Globovisión. At the same time, opposition politicians know that their base audience is much
more likely to watch a privately-owned channel such as Globovisión. As a result, it is entirely
possible that the incentive structure for appearing on Globovisión varies significantly between
pro-government and anti-government actors, i.e. individuals associated with the opposition
are simply more interested in appearing on Globovisión.
Second, the geographic location of the channel, which is headquartered in Caracas, pro-
vides some context for considering the effects of Globovisión’s social environment and the
networks it accesses. The metropolitan area of Caracas is split more or less evenly between
government and opposition supporters. However, five of the city’s six municipalities—
Baruta, Chacao, Sucre, El Hatillo, and Metropolitan Caracas—have tended to vote for
opposition mayoral candidates. With Globovisión often reporting on events in the Caracas
area, and a large portion of its traditional audience residing in these markets, the channel
may have found it most convenient, and in its market self-interest, to invite interviewees
from these mayoral administrations or from Caracas-based universities or research centers.
The inevitable effect would be to augment the relative proportion of guests who would hold
views highly critical of the government.
A broader political factor is also relevant for interpreting this study’s general finding of an
absence of bias in favor of the government. After the death of President Chávez in March
2013, Venezuela entered a political crisis. This crisis heightened when the opposition presi-
dential candidate, Governor Capriles, rejected the results of a tight election held in April
2013. Then, by the end of 2014, Venezuela had fallen into economic crisis. This combina-
tion of political and economic travails created a situation in which an overwhelming major-
ity of Venezuelans judged the country’s situation as negative.
Hence, the Maduro government faced daunting conditions in 2013 and 2014, and the chal-
lenges it faced constituted the country’s most important issues. As a result, media outlets
covered matters that, by their very nature, placed the government in a mode of having to
defend its record amid objective signs of poor performance. The prevalence of news top-
ics that were unpleasant for the government meant that independent media would report
on issues that the government would rather avoid, and this may explain a tendency for
Globovisión to invite guests associated with the opposition, as they would be more inclined
under the circumstances to discuss these controversial topics.
44 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | V I . CO N C LU S I O N

V I . CO N C LU S I O N | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 45
A CLOSING NOTE ON BREADTH AND DEPTH: COVERAGE OF ISSUES AND ACTORS
The study revealed Globovisión to have given coverage to a fairly wide range of issues and perspec-
tives. A total of 36 topics received coverage, with many of these directly related to the main themes
and problems prevalent during the junctures. This suggests that Globovisión covered the most im-
portant issues and problems facing the country. It is difficult to draw a similarly positive conclusion
about the extent to which Globovisión strengthened political pluralism with fresh perspectives on
controversial topics or discussing new elements of hot topics. For example, the case of jailed opposi-
tion leader Leopoldo López received less attention from Globovisión than from international outlets
such as CNN in Spanish or NTN24. Lillian Tintori, López’s wife and his spokeswoman while he
has been in jail awaiting trial, appeared on Globovisión a handful of times. But, she was not fea-
tured as an in-studio interviewee.
Though the study’s quantitative analysis of empirical data did not find evidence of bias in favor of
the government, there is a noteworthy qualitative finding about the visibility of key government and
opposition leaders. Specifically, our study revealed that former presidential candidate and Gover-
nor of Miranda state Henrique Capriles appeared 11 times out of approximately 1,000 instances of
appearances by individuals. Under Globovisión’s previous ownership group, Governor Capriles fre-
quently appeared on the channel for invited interviews and as a partisan actor in news reports. In the
channel’s first year of operation under new ownership, Governor Capriles, who won 49 percent of the
vote in the 2013 elections and led the opposition’s campaign in advance of the 2013 municipal ballot-
ing, did not appear as an interview guest on any of the Globovisión programs in the research sample.
Meanwhile, President Maduro appeared as a partisan actor in 42 segments. Governor Capriles’s
degree of visibility on Globovisión does not seem commensurate with his political prominence.

46 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S I V. CO N C LU S I O N
References
Arpan, L., & Raney, A. (2003). An Experimental Investigation of News Source and the Hostile
Media Effect. Journalism & Mass Communication Quarterly, 80(2), 265–281.
Ayton, P., & Tumber, H. (2001). The Rise and Fall of Perceived Bias at the BBC. We Should
Look to Psychology not Content Analysis to Understand Allegations of Political Partiality.
InterMedia, 29(4), 12.
Babad, E. (1999). Preferential Treatment in Television Interviewing: Evidence from Nonverbal
Behavior. Political Communication, 16(3), 337–358.
Babad, E., & Peer, E. (2010). Media Bias in Interviewers’ Nonverbal Behavior: Potential
Remedies, Attitude Similarity and Meta-Analysis. Journal of Nonverbal Behavior, 34(1),
57–78.
Banning, S. A., & Coleman, R. (2009). Louder Than Words: A Content Analysis of Presidential
Candidates’ Televised Nonverbal Communication. Visual Communication Quarterly, 16(1),
4–17.
Beschloss, M., & Talbott, S. (1993). At the Highest Levels. Boston: Little, Brown.
Bozell, L. B., & Baker B. H. (1990). And That’s the Way It Isn’t: A Reference Guide to Media
Bias. Alexandria, VA: Media Research Center.
Chaffee, S. H., & Schleuder, J. (1986). Measurement and Effects of Attention to Media News.
Human Communication Research, 13(1), 76–107.
Covert, T., & Wasburn, P. (2007). Measuring Media Bias: A Content Analysis of Time and
Newsweek Coverage of Domestic Social Issues, 1975-2000. Social Science Quarterly, 88(3), 690–
706.
Croteau, D., & Hoynes, W. (1994). By Invitation Only: How the Media Limit Political Debate.
Monroe, ME: Common Courage Press.
D’Alessio, D. (2012). Media Bias in Presidential Election Coverage, 1948–2008: Evaluation via
Formal Measurement. Lanham, MD: Lexington Books.
D’Alessio, D., & Allen, M. (2000). Media Bias in Presidential Elections: A Meta-Analysis.
Journal of Communication, 50(4), 133–156.
Donohue, G. A., Tichenor, P., & Olien, C. N. (1972). Gatekeeping: Mass Media Systems and
Information Control. In F. G. Kline & P. J. Tichenor (Eds.), Current Perspectives in Mass
Communication Research (pp. 41–69). Beverly Hills: Sage Publications.
Efron, E. (1971). The News Twisters. Los Angeles: Nash.
Ekman, P. (1972). Universals and Cultural Differences in Facial Expressions of Emotion. In J.
Cole (Ed.), Nebraska Symposium on Motivation (Vol. 19). Lincoln: University of Nebraska
Press.
Entman, R. M. (2007). Framing Bias: Media in the Distribution of Power. Journal of
Communication, 57(1), 63–173.
Freelon, D. (2010). ReCal: Intercoder Reliability Calculation as a Web Service. International
Journal of Internet Science, 5(1), 20–33.
Friedland, L. A. (1992). Covering the World. New York: Twentieth Century Fund.
Friedman, H. S., DiMatteo, M. R., & Mertz, T. J. (1980). Nonverbal Communication on Televsion
News: The Facial Expressions of Broadcasters during Coverage of a Presidential Election
Campaign. Personality and Social Psychology Bulletin, 6(3), 427–435.

V I . CO N C LU S I O N I AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 47
Friedman, H. S., Mertz, T. J., & DiMatteo, M. R. (1980). Perceived Bias in the Facial
Expressions of Television News Broadcasters. Journal of Communication, 30(4), 103–111.
Graber, D. A., & Smith, J. M. (2005). Political Communication Faces the 21st Century.
Journal of Communication, 55(3), 479.
Graber, D. (1990). Seeing Is Remembering: How Visuals Contribute to Learning from
Television News. Journal of Communication, 40(3), 134–155.
Gunther, A. C., & Liebhart, J. L. (2006). Broad Reach or Biased Source? Decomposing the
Hostile Media Effect. Journal of Communication, 56(3), 449–466.
Gunther, A. C., & Schmitt, K. (2004). Mapping Boundaries of the Hostile Media Effect.
Journal of Communication, 54(1), 55–70.
Hopmann, D. N., Van Aelst, P., & Legnante, G. (2011). Political Balance in the News: A
Review of Concepts, Operationalizations and Key Findings. Journalism, 13(2), 1–18.
IBOPE (Instituto Brasileiro de Opinião Pública e Estatística). (2013). Media Book 2013.
Retrieved from http://www.mediabook.ibope.com/pais/america-latina/2013/abetura-2/
highlights.
Lombard, M., Snyder-Duch, J., & Bracken, C. C. (2002). Content Analysis in Mass
Communication: Assessment and Reporting of Intercoder Reliability. Human Communication
Research, 28(4), 587–604.
McCombs, M., & Shaw, D. (1972). The Agenda-Setting Function of Mass Media. Public
Opinion Quarterly 36(2), 176–187.
McCombs, M. (2005). A Look at Agenda-Setting: Past, Present and Future. Journalism
Studies, (4), 543–557.
McCombs, M. E., Shaw, D. L., & Weaver, D. H. (1997). Communication and Democracy:
Exploring the Intellectual Frontiers in Agenda-Setting Theory. Mahwah, NJ: Erlbaum.
Moriarty, S. E., & Popovich, M. N. (1991). Newsmagazine Visuals and the 1988 Presidenial Election.
Journalism & Mass Communication Quarterly, 68(3), 371–380.
Morris, B., & Nydahl, J. (1983). Toward Analyses of Live Television Broadcasts. Central States Speech
Journal, 34(3), 195–202.
Neuendorf, K. A. (2002). The Content Analysis Guidebook. Thousand Oaks, CA: Sage Publications.
Niven, D. (2003). Objective Evidence on Media Bias: Newspaper Coverage of Congressinal
Party Switchers. Journalism and Mass Communication Quarterly, 80(2), 311–326.
Reese, S. D., Grant, A., & Danielian, L. H. (1994). The Structure of News Sources on
Television: A Network Analysis of CBS News, Nightline, MacNeil/Lehrer, and This Week
with David Brinkley. Journal of Communication, 44(2), 84–107.
Richmond, V. P., McCroskey, J. C., & Payne, S. K. (1991). Nonverbal Behavior in Interpersonal
Relations, 2nd edition. Englewood Cliffs, NJ: Prentice-Hall.
Robinson, M., & Sheehan, M. (1983). Over the Wire and On TV: CBS and UPI in
Campaign ‘80. New York: Russell Sage Foundation.
Salwen, M. B. (1995). News of Hurricane Andrew: The Agenda of Sources and the Sources’
Agendas. Journalism & Mass Communication Quarterly, 72(4), 826–840.
Shoemaker, P. J., & Reese, S. D. (1991). Mediating the Message: Theories of Influences on
Mass Media Content. New York: Longman.

48 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S I R E F E R E N C E S
Smith, C., & Becker, L. (1989). Comparison of Journalistic Values of Television Reporters and
Producers. Journalism Quarterly, 66(4), 793–800.
Soroka, S. N. (2003). Media, Public Opinion, and Foreign Policy. The International Journal of
Presidential Politics, 8(1), 27–48.
Stempel, G. H. (1969). The Prestige Press Meets the Third-Party Challenge. Journalism Quarterly,
46(4), 699–706.
Stempel, G. H., & Windhauser, J. W. (1989). Coverage by the Prestige Press of the 1988 Presidential
Campaign. Journalism & Mass Communication Quarterly, 66(4), 894–919.
Stovall, J. (1985). The Third-Party Challenge of 1980: News Media Coverage of the
Presidential Candidates. Journalism Quarterly, 62, 266–71.
Tuggle, C. A. (1991). Media Relations during Crisis Coverage: The Gainesville Student
Murders. Public Relations Quarterly, 36(2), 23–28.
Tuggle, C. A., & Huffman, S. (2001). Live Reporting in Television News: Breaking News or Black
Holes? Journal of Broadcasting & Electronic Media, 45(2), 335–344.
Tuggle, C. A., & Huffman, S. (1999). Live News Reporting: Professional Judgment or
Technological Pressure? A National Survey of Television News Directors and Senior Reporters.
Journal of Broadcasting & Electronic Media, 43(4), 492–505.
Upshaw, J. (1994). Visual Newsworkers’ Attitudes toward Local Television News. Paper
presented at the national convention of the Broadcast Education Association, Las Vegas, NV.
Uscinski, J. (2014). The People’s News: Media, Politics, and the Demands of Capitalism.
New York: NYU Press.
Vallone, R. P., Ross, L., & Lepper, M. (1985). The Hostile Media Phenomenon: Biased
Perception and Perceptions of Media Bias in Coverage of the Beirut Massacre. Journal of
Personality and Social Psychology, 49(3), 577.
Waldman, P., & Devitt, J. (1998). Newspaper Photographs and the 1996 Presidential Election:
The Question of Bias. Journalism & Mass Communication Quarterly, 75(2), 302–311.
Watts, M. D., Domke, D., Shah, D. V., & Fan, D. P. (1999). Elite Cues and Media Bias in
Presidential Campaigns Explaining Public Perceptions of a Liberal Press. Communication
Research, 26(2), 144–175.
Weaver, D., & Wilhoit, G. (1996). The American Journalist in the 1990s: U.S. News People at the
End of an Era. Mahwah, NJ: Lawrence Erlbaum Associates.
Welsh, M., Fenwick, M., & Roberts, M. (1997). Primary Definitions of Crimes and Moral Panic:
A Content Analysis of Experts’ Quotes in Feature Newspaper Articles on Crime. Journal
of Research in Crime and Delinquency, 34(4), 465–74.
Zeldes, G., Fico, F., Capenter, S., & Diddi, A. (2008). Partisan Balance and Bias in Nework Cover
age of the 2000 and 2004 Presidential Elections. Journal of Broadcasting & Electronic Media,
52(4), 563–580.

V I I . A P P E N D I C E S I AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 49
Appendix 1. General Codebook
Sampling Unit: Show broadcast. First hour of the broadcast should be coded if longer than an
hour (e.g., Aló Venezuela).
Unit of data collection: For news shows, the unit of data collection will be the news story. For
in-depth interview and opinion shows, the unit of data collection is the segment. A segment
is defined by periods between commercial breaks. All of the segments in the selected broad-
cast should be considered for coding. If a news story or interview continues after a segment
break, treat the second part as a new coding unit (i.e., a column in the spreadsheet).
Stories/segments that have content on the target topics will be identified separately and coders
will be provided with a list of stories/segments (by start time and number) that they should
code.
1. Coder Initials
2. Broadcast Code: Unique code for each broadcast (provided)
3. Juncture Code: Unique code for the critical juncture period (provided)
1 = 11/9–12/15/13; 2 = 2/12–3/21/14; 3 = 3/25–5/14/14; 4 = 8/7–8/29/14
4. Show Name: Write name of show in this field (provided)
5. Show ID: Fill in ID number for each show (provided)
6. Date: Enter date of broadcast (mm/dd/yy) (provided)
e.g., November 20, 2013 is “11/20/13”
7. BroadcastStartTime:Enterstarttimeofselectedbroadcast(hh:mm)(provided)
e.g., a show airing at 2:00 p.m. should be coded as “14:00,” at 2:00 a.m. as “02:00”
NOTE: This should be the same value for ALL stories/segments in a given broadcast
VII. APPENDICES

8. BroadcastEndTime:Enterendtimeofselectedbroadcast(hh:mm)(provided)
e.g., a show ending at 2:00 p.m. should be coded as “14:00,” at 2:00 a.m. as “02:00”
9. File ID: Digital file name for video (provided)
10. Story/Segment Number: Indicate where the story or segment falls within the broadcast
(provided).
1. first story/segment
2. second story/segment
3. third story/segment
4. fourth story/segment
5. fifth story/segment
6. sixth story/segment
7. seventh story/segment
etc...
11. Story/Segment Start: Enter timer value when story/segment starts (provided)
Format: hh:mm
12. Story/Segment End: Enter timer value when story/segment ends (provided)
Format: hh:mm
13. Video Start: Time the video starts (provided)
Format: mm:ss
14. Video End: Time the video ends (provided)
Format: mm:ss
Topics and Presentation
15. Target Topic: Indicate whether the segment includes discussion of a topic or subtopic
relevant to the period under study. To make this determination, refer to supplementary
document for the relevant period of study (provided)
Yes=1 (e.g., discussion of municipal elections, discussion of protests, discussion of international
talks, discussion of shortages, etc.)
No=0 (e.g., no significant discussion of the target topic(s) [do not code, move on to next story
or segment]).
16. Topic Notes: Notes entered by pre-coders on the guests, focus, topic, etc. of each story/
segment
17. Topic 1: Enter the detailed code for the first relevant topic discussed in the story/segment.
Make your best judgment and only enter one code in each field. If multiple relevant topics
are discussed, code for the second topic (even if introduced simultaneously) under Topic 2, etc.
NOTE: Only code for topics introduced by an interview guest when the host(s) directly engage(s) with the
topic. 99=missing n/a (more likely for Topics 2–4)
18. Other Topic: If “other” relevant topic juncture, briefly identify topic discussed.
19. Topic 1 Verbal Presentation: Identify if the verbal presentation of Topic 1 by the hosts,
narrators, etc., is favorable, unfavorable, neutral, or ambivalent vis a vis the government’s
50 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | V I I . A P P E N D I C E S

V I I . A P P E N D I C E S | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S 51
position). Failure to interject with a pro-government counterargument to points made by an
opposition representative is neutral, rather than unfavorable. Interjection with counterarguments
that balance the overall presentation is “ambivalent” (i.e. playing devil’s advocate). When
presentations and interventions support the government’s position, this is favorable and when they
oppose it, this is unfavorable (this is the case independently of whether the actors are pro-government,
pro-opposition, or disinterested parties).
NOTE: If this is an interview, the verbal presentation of the topic is defined by the nature of the questions, state-
ments, and arguments that introduce a new topic for discussion AND move the discussion forward. As stated above,
topics should only be considered if engaged directly by the host (e.g., don’t code for every topic mentioned by a guest,
unless there is back and forth with the host or unless the host asks the question to begin with). For news stories, ver-
bal presentation is defined by the way in which the host(s) introduce(s) the topic, narrate(s) throughout, interview(s)
people “on location,” etc.
1. favorable to the government’s position
2. unfavorable to the government’s position
3. neutral with respect to the government’s position
4. ambivalent: presents both government and opposition positions
88. not verbally introduced
20. Topic 1 Non-Verbal Presentation: Identify if the non-verbal presentation of Topic 1 is favorable,
unfavorable, neutral, or ambivalent vis a vis the government’s position. Neutrality is the baseline (e.g.
a lack of facial expression, gestures that do not indicate support or rejection of claims being made,
innocuous images, etc.). Failure to show images that would undercut the government (or opposition)
position is neutral, rather than favorable/unfavorable. If this is an interview, the non-verbal
presentation of the topic is defined by the facial expressions and body language of the host when
asking questions, making statements, and presenting arguments throughout the segment. For news
stories, non-verbal presentation includes the facial expressions, hand movements, body language,
etc. of hosts and narrators as well as the still and video images used to introduce the topic. This also
includes the way a story or interview is edited, including if an interview is cut off in a way that is
favorable/unfavorable to the government position (such cuts should not impact the verbal
presentation code, which should be based solely on what is said by the hosts).
4. ambivalent: presents both government and opposition positions
88. Topic not introduced by hosts or with visuals
21.–32. Topic 2–Topic 4 coded in the same way as above
Actor Affiliation, Visibility, and Favorability toward Government
For this section, the FIRST FOUR (4) actors that appear and/or are discussed in a news story/ segment should
be coded for the variables below. “Actors” can be individuals (e.g., politicians, experts, etc.) or groups (e.g.,
“students,” “police,” etc.) that have done something, said something, etc. (e.g, in a story that discusses ac-
tions of supporters of Nicolás Maduro, the “supporters” are the actor, not Maduro). For interview or debate
segments, the actors are the individuals interviewed or who sit on a panel/round table. For news stories or
reports, the actors can be people interviewed, people discussed by the hosts or others, or people shown on still
photographs or video clips while being discussed.

33. Actor 1 Paternal Last Name: Enter text of paternal last name (no accents)
99=not given
34. Actor 1 Maternal Last Name: Enter text of maternal last name (no accents)
99=not given
35. Actor 1 First Name: Enter first name (no accents)
99=not given
36. Actor 1 Affiliation 1: Enter first affiliation mentioned (e.g., Minister of X)
99=none given
37. Actor 1 Affiliation 2: Enter second affiliation mentioned (e.g., Political Party Y)
99=none given
38. Actor 1 Type: Enter primary actor type. If more than one type, enter the one that is most
relevant to the given topic/story.
1. Government bureaucrat, security forces, police, government agent, etc.
(non-elected, current or former)
2. Elected representative (e.g., president, mayor, assembly member, governor,
Member of Congress, former elected representatives, etc.)
3. Candidate for office or surrogate (e.g., secretary general of a political party, party plat
form, party issues, MUD, political party)
4. Civil society group or representative (e.g., chamber of commerce, labor union,
human rights organization, NGO, international organization, etc.)
5. Issue expert, pundit, journalist (e.g., journalist, intellectual, professor, scientist, etc.)
6. Member of the public (e.g., someone interviewed “on the street” to represent the
opinion of the “ordinary” citizens)
7. Private sector representative (e.g., business owner, CEO of corporation, etc.)
8. Political party militant (if not a candidate or surrogate)
9. Other
39. Actor 1 Other Type: Enter text of other actor type (only if “9,” “other” for previous
question).
40. Actor 1 Visibility: How visible is Actor 1 in the segment?
NOTE: code as the highest relevant (i.e., if mentioned AND quoted, code “2”).
1. Mentioned
2. Quoted
3. Shown in picture and mentioned or quoted but not speaking (e.g., image-bites)
4. Shown while speaking or interviewed
41. Actor 1 Visibility 2: Enter amount of time Actor is visible, discussed, or heard in the
segment. Format: mm:ss
42. Actor 1 Favorability (verbal): Enter score for how favorable to the government’s position
the presentation or portrayal of the actor is in terms of what is said as would be perceived
by the individual or group appearing in the story. This includes the way the host(s)
discuss(es) the actor and his/her actions, as well as how the actor presents him/herself.
Partisans who are guests on interview shows will usually be coded as “favorable” (if govern
52 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S V I I . A P P E N D I C E S

ment partisans/allies) or unfavorable (if opposition partisans).
3. neutral with respect to the government’s position (i.e., supports NEITHER the
government nor opposition narrative)
4. ambivalent: mixes both pro- and anti-government positions
43. Actor 1 Favorability (visual): Enter score for how favorable to the government’s position
the presentation or portrayal of the actor is in terms of what is shown as would be perceived
by the individual or group appearing in the story. This might include the gestures and facial
expressions of the host in discussing the actor or the situation, as well as the choice of images
and video clips used, as well as how the actor is visually presented during an interview.
4. ambivalent: represents both pro- and anti-government positions
44–76. Actors 2–4 coded in the same way
Appendix 2. Supplementary File 1: Topics and Subtopics
100 Municipal Elections
110 Economic Situation
120 The Election Campaign
130 Election Day Turnout and Results
140 Party and Leader Reactions to Election Results
150 Electoral Institutions—The Consejo Nacional Electoral (El CNE)
160 Other
200 Street Demonstrations
210 Anti-Government Demonstrations—Peaceful Protests
220 Anti-Government Demonstrations—Violent Protests
230 Counter Demonstrations
240 Judicial Process
250 Human Rights
260 Police and Security Forces
270 International Actors
280 Other
300 International Dimensions of Crisis
310 The Vatican
320 UNASUR
321 Venezuela-South American Relations
322 Venezuela-Brazil Relations
323 Venezuela-Colombia Relations
324 Venezuela-Ecuador Relations
330 The United States
331 U.S.-Venezuela Relations
332 Sanctions Proposals
340 Cuba
341 Venezuela-Cuba Relations
350 Alianza Bolivariana para los Pueblos de América (ALBA)
360 Organization of American States (OAS)
370 Other

400 Shortages of Basic Goods
410 Economic Situation
420 International Actors
430 Economic Policy Context
440 Hospitals and Health Clinic
450 Basic Food Basket (Cesta básica)
460 Fingerprint system (Captahuellas)
470 Other
Key Words and Phrases for Topics
100 Municipal Elections
We’re the Majority, Economic War, Economic Counter-Offensive, Day of Loyalty, Illegitimate,
Plebiscite
200 Street Demonstrations
The Street, Scarcity, Right to Protest, Separation of Powers, Fascism, Las guarimbas, Inflation,
Colectivos, Political Prisoners, Insecurity, Miami
300 International Dimensions of the Crisis
Hombre de paz, Amnesty, Presos políticos, Colectivos, Truth Commission, Miami, Economic War,
Mutual Respect, Golpe continuo
400 Shortages in Basic Goods
Acaparamiento, Falta de producción, Guerra económica, Contrabando, Corrupción, Crisis de salud
pública, Emergencia en el sector médico, Las colas (The lines)
Appendix 3. Codebook Supplementary File 2: Summary of Critical Juncture Periods
I. Municipal Elections (Nov. 9–Dec. 15, 2013)
Venezuela held municipal elections on December 8, 2013. The vote took place against a backdrop of
high-intensity political contestation. The April 14, 2013 presidential elections saw opposition candi-
date Henrique Capriles reject the official outcome—a victory for President Maduro by 1.5 percent.
After the opposition’s appeal of the results failed to bear fruit, Capriles and the opposition turned
their attention to the municipal elections.
Until November, the opposition was ahead in the polls. With one month left before the elections and
the all-important Christmas holiday looming afterwards, the government undertook policy measures
that lowered prices at consumer goods stores.
Government candidates won more mayoral offices and nationally pro-government candidates won 6
percent more of the vote. In defeat, Capriles spoke on behalf of the opposition and complained of
facing highly “unequal campaign conditions.”
II. Street Demonstrations (Feb. 12–Mar. 21, 2014)
Venezuela experienced a wave of mostly student-led demonstrations from approximately January
23–March 21. The demonstrations originated in Táchira and Mérida (interior states on the west-
ern border with Colombia) and spread to Caracas (the capital) in mass form on February 12.
They snowballed after opposition leader Leopoldo López was taken prisoner on February 18. The
54 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S I V I I . A P P E N D I C E S

demonstrations took a variety of forms, from peaceful protest to more violent barricades and road
blocks, and generated counter-demonstrations. Likewise, state security institutions responded with
the use of excessive force in some cases and in step with accepted standards of policing in others. The
demonstrations dissipated toward the middle of March, with fatigue and state repression undercutting
momentum.
III. International Dimensions of Crisis (Mar. 25–May 14, 2014)
From March 25 to May 14 Venezuela’s political crisis moved to the international level, with dynamics
playing out in two theaters. Within the region, UNASUR (the Union of South American Nations)
and the Vatican served as the formal conveners of political talks between the government and the op-
position. To help convene these talks over political issues for defusing the political crisis, UNASUR
sent a delegation of foreign ministers from Brazil, Colombia, and Ecuador and the Vatican operated
through its Ambassador (El Nuncio) in Caracas. In the United States, meanwhile, the Senate Foreign
Relations Committee took up the issue of human rights violations. The Committee held a high-
profile committee meeting with the Assistant Secretary of State for Western Hemispheric Affairs to
discuss proposed legislation for imposing targeted sanctions against Maduro government officials.
The talks produced multiple rounds of discussion but yielded few concrete gains for alleviating the
crisis. When the talks failed to produce concessions from the government on the issue of releasing
high profile political prisoners from jail, this increased the likelihood for actions by the Senate and the
State Department.
IV. Shortages in Basic Goods (Aug. 7–29, 2014)
From August 7–August 29, 2014 Venezuela’s intermittent problems with availability of foodstuffs
and goods reached new heights. Beyond decreased availability of food, the crisis was marked by
problems with decreased availability of medical supplies, both over-the-counter and prescription.
Consequently, the situation at hospitals and medical clinics moved to the front and center of media
discussion of the shortages. The government announced a major new policy measure to combat
hoarding during this period. The acute problems with shortages dissipated but underlying issues with
availability continued.
Appendix 4. Index of Bias
The index of bias presented in the second part of “Empirical Analysis and Findings” is calculated us-
ing factor analysis of six variables at the level of the news story/interview segment. The variables are:
1) Visibility Type, 2) Minutes Visible, 3) Actor Verbal Favorability, 4) Actor Non-Verbal Favorability,
5) Verbal Topic Presentation, and 6) Non-Verbal Topic Presentation. As each unit is potentially
composed of up to four actors and up to four topics, these six variables are themselves composed of
multiple elements, the operationalization of which we will now discuss.
Visibility Type is measured on a scale of 1–4 for each actor that appears in a story or is interviewed.
To generate a story-level variable of actor visibility, we take the difference of the sums of the “visibility
type” variable for all pro-government actors and all pro-opposition actors. As such, a story with four
pro-government actors who spoke while being shown (visibility type=4) would have a net story-level
government visibility value of 16, and a story with four pro-opposition actors who spoke while being
shown would have a story-level net government visibility score of –16.
The Minutes Visible variable would be similarly aggregated at the story level by subtracting the

56 V I I . A P P E N D I C E S | AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S
total number of minutes pro-opposition actors were visible from the total number of minutes pro-
government actors were visible. To measure actor favorability at the story-level, as already presented
in the Findings section, we operationalize variables for the net-verbal and non-verbal favorability of
actors. To do this, we take the sum of actors portrayed in a way that was favorable to the government
and subtract the sum of actors portrayed in a way that was unfavorable to the government. The net
favorability of topic presentation (verbal and non-verbal) was measured in a similar way. Specifically,
we added up the total of topics in each story that were framed in such a way that was favorable to the
government and subtracted from them the total of topics that were unfavorable to the government.
The results of a factor analysis of these six variables are presented in Table A-6.
Appendix 5. Supplementary Tables and Figures
Table A-1. Final Content Analysis Reliability Indicators
% Agree Scott’s Pi Cohen’s Kappa K’s Alpha
Topic 79.6 0.396 0.401 0.400
Verbal Presentation 79.2 0.319 0.324 0.324
Topic Non-Verbal Presentation 86.6 0.358 0.361 0.362
Actor Visibility 88.4 0.773 0.774 0.775
Actor Favorability (Verbal) 80.6 0.646 0.648 0.648
Actor Favorability (Non-Verbal) 88.0 0.745 0.745 0.747
Calculated using “ReCal: reliability calculation for the masses.” Created by Deen Freelon, American
University (http://dfreelon.org/utils/recalfront/)

Table A-2: Topics and Subtopics Covered by Critical Juncture Period

Topic N %
Municipal Elections-General 9 5.3
Economic Situation 51 30.2
The Election Campaign 44 26.0
Election Day Turnout and Results 15 8.9
Party and Leader Reactions to Election 25 14.8
Electoral Institutions 18 10.7
Other 7 4.1
Street Demonstrations-General 57 14.8
Peaceful Anti-Gov. Demonstrations 65 16.9
Violent Anti-Gov. Demonstrations 78 20.3
Counter Demonstrations 17 4.4
Judicial Process 36 9.4
Human Rights 23 6.0
Police and Security Forces 41 10.7
International Actors 31 8.1
Other 37 9.6
International Dimensions of Crisis 17 26.6
The Vatican 9 14.1
UNASUR 12 18.8
Venezuela-South American Relations 2 3.1
Venezuela-Brazil Relations 2 3.1
Venezuela-Colombia Relations 1 1.6
The United States 1 1.6
U.S.-Venezuela Relations 5 7.8
Sanctions Proposals 3 4.7
Cuba 1 1.6
Organization of American States (OAS) 3 4.7
Other 8 12.5
Shortages of Basic Goods 12 7.7
Economic Situation 46 29.7
International Actors 7 4.5
Economic Policy Context 34 21.9
Hospitals and Health Clinics 15 9.7
Basic Food Basket (Cesta Básica) 7 4.5
Fingerprint system (Captahuellas) 20 12.9
Other 14 9.0
Table A-3. Coverage of Target Topic

N %
Municipal Elections 144 46.5
Street Demonstrations 270 61.9
International Dimensions 55 11.8
Basic Goods Shortages 133 31.6
All 602 36.9

58 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | V I I . A P P E N D I C E S
Table A-5. Explaining Variations in the Index of Bias OLS Model of Bias Index
β s.e. P-value
Municipal Elections -0.25 0.143 0.075
Street Demonstrations -0.24 0.133 0.067
Shortages of Goods -0.01 0.144 0.936
Noticias Globovisión Estelar 0.11 0.133 0.400
Vladimir a la Una -0.04 0.163 0.785
Story Number -0.03 0.006 <0.001
Intercept 0.38 0.183 0.040

Observations = 597
Adj R-squared = 0.0447

Note: The“International Dimensions of the Crisis” Juncture and “Aló Venezuela”
were the reference categories, and thus omitted from the models.
Table A-6. Factor Analysis of Measures of Bias
Variable Factor Loadings
Visibility Type 0.699
Minutes Visible 0.375
Actor Verbal Favorability 0.792
Actor Non-Verbal Favorability 0.757
Topic Verbal Favorability 0.601
Topic Non-Verbal Favorability 0.641
Observations 597
Eigenvalue = 2.6
Table A-4. Dependent and Independent Variables Used in OLS Regressions
N Mean S.D. Min Max
Dependent Variables
Bias Index 597 0.00 0.92 -4.09 3.04
Net Visibility Value 598 0.10 3.85 -16 13
Net Minutes Visible 598 -0.09 3.57 -29 17.83
Net-Actor Favorability (Verbal) 598 -0.20 1.26 -4 4
Net-Actor Favorability
(Non-Verbal) 598 -0.05 0.90 -4 4
Topic Favorability (Verbal) 598 -0.05 0.58 -2 3
Topic Favorability (Non-Verbal) 598 -0.04 0.46 -2 2
Independent Variables
Municipal Elections 598 0.24 0.43 0 1
Street Demonstrations 598 0.44 0.50 0 1
International Dimensions 598 0.09 0.29 0 1
Shortages of Goods 598 0.22 0.42 0 1
Noticias Globovisión Estelar 598 0.77 0.42 0 1
Vladimir a la Una 598 0.14 0.35 0 1
Aló Venezuela 598 0.09 0.28 0 1
Story Number 598 9.14 6.21 1 29

Figure A-1. Net Government Visibility Quality
Figure A-2. Net Government Minutes Visible
Figure A-3. Net Government Actor Favorability
050100150200
Frequency
−16 −12 −8 −4 0 4 8 12 16
0100200300400
Frequency
−30 −20 −10 0 10 20
050100150200
Frequency
−4 −3 −2 −1 0 1 2 3 4

Figure A-4. Net Government Actor Non-Verbal Favorability
60 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S I V I I . A P P E N D I C E S
Figure A-5. Net Government Topic Favorability
Figure A-6. Net Government Topic Non-Verbal Favorability
0100200300400
Frequency
−4 −3 −2 −1 0 1 2 3 4
0100200300400
Frequency
−4 −3 −2 −1 0 1 2 3 4
0100200300400500
Frequency
−4 −3 −2 −1 0 1 2 3 4

Figure A-7. Pro-Government Bias and Story Prioritization: Scatter Plots by Critical Juncture
−4−2024−4−2024
0 10 20 30 0 10 20 30
Municipal Elections Street Demonstrations
International Dimensions Shortages
Index of Bias 95% CI
Fitted values
Story Number
Graphs by Juncture Number

4 AU C E N T E R F O R L AT I N A M E R I C A N & L AT I N O S T U D I E S | C H A P T E R T I T L E H E R E
MAILING ADDRESS
Center for Latin American & Latino Studies
American University
4400 Massachusetts Ave., NW
Washington, DC 20016-8137 USA
OFFICE LOCATION
4545 42nd St., NW, Suite 308
Washington, DC 20016 USA
T: 1-202-885-6178
F: 1-202-885-6430
clals@american.edu
www.american.edu/clals
CenterForLatinAmerican
&LatinoStudies
@AU_CLALS

Sphere layout final-051415

More Related Content

Similar to Sphere layout final-051415 (20)

More from hinterlaces (20)

Recently uploaded (20)

Sphere layout final-051415