california state university, northridge therefore
TRANSCRIPT
CALIFORNIA STATE UNIVERSITY, NORTHRIDGE
THEREFORE, SOCRATES IS A TALKING HEAD:
BELIEF BIAS AND SYLLOGISTIC REASONING ERRORS IN CABLE NEWS
A thesis submitted in partial fulfillment of the requirements
For the degree of Master of Arts
in Psychology, General Experimental
By
Alexander Benson Swan
May 2011
The thesis of Alexander Benson Swan is approved:
Dustin P. Calvillo, Ph.D. Date'
Mark P. Otten, Ph.D. Date
Abraham M. Rutchick, Ph.D., Chair Date
California State University, Northridge
ii
Signature Page
Dedication
Abstract
Introduction
Table of Contents
Rational Choice and The Hostile Media Bias
The Power of Persuasion
The Belief Bias in Syllogistic Reasoning
The Present Study
Method
Participants
Results
Design and Materials
Procedure
Preliminary Analyses
Primary Analyses
Secondary Analyses
Discussion
References
Appendix A
Appendix B
IV
11
111
v
1
2
4
6
12
14
14
14
16
17
17
19
20
21
25
33
37
ABSTRACT
THEREFORE, SOCRATES IS A TALKING HEAD:
BELIEF BIAS AND SYLLOGISTIC REASONING ERRORS IN CABLE NEWS
By
Alexander Benson Swan
Master of Arts
in Psychology, General Experimental
Cable news media has grown exponentially in audience over the last 30 years, and in that
time, argumentation and debate have become fractured and fallacious. Political partisans
refuse to watch pundits or commentators with viewpoints that differ from their own. Due
to this divide, a hostile perception of media outlets increases. Though many factors
contribute to media hostility, belief bias in logical reasoning was considered for this
study. Political ideology and congruent arguments may be believable but not necessarily
valid, perpetuating fallacious argumentation and improper debate. Participants (n = 98)
completed an argument validity evaluation task, determining if 16 arguments paraphrased
from cable news pundits and commentators were structurally valid. Half of the arguments
were liberal and the other half conservative. Additionally, half were valid and the other
half invalid. Thus, four argument types were used in the study. Overall, the beliefbias did
not influence liberals and conservatives on the argument evaluation task. The data do
v
suggest a trend for bias, with participants without formal logic training the most
susceptible to beliefs. Implications of these findings, including limitations of the current
experiment and paths for future research of this topic, are discussed.
VI
Introduction
In the United States, 24-hour cable news channels began in the 1980s. In the last
10 years, these channels began to offer news, political discussion, and commentary nearly
every day ofthe year. The audience of these channels has nearly doubled from the mid-
1990s to the mid-2000s (Coe et al., 2007), and it appears this trend is not about to stop.
They now focus their evening programs on political commentary (MSNBC, Lawrence
O'Donnell and Rachel Maddow; Fox News Channel, Bill O'Reilly and Sean Hannity).
Pmiisanship is no stranger to these programs, with MSNBC accused with a liberal bias
and Fox News accused with a conservative bias, and CNN trying to strike a middle
ground between the two (Coe et al., 2007). Due to these divisions, liberals and
conservatives watch their respective channels, making little effort to hear
counterattitudinal arguments that might rattle the certainty of their position on the issues
(Tormala & Petty, 2002).
Researchers have shown that liberals and conservatives are different in their
worldviews and morals (e.g., Graham, Haidt, & Nosek, 2009; Haidt & Graham, 2007;
Haidt & Joseph, 2004). Moreover, on a task examining open-ended responses on 12life
narrative scenes, in which respondents were to describe the scene in as much detail as
possible (e.g., a positive childhood scene), liberals were more likely to recall scenes in
which they developed empathy and learned to open up to new perspectives, while
conservatives recalled the scenes recognizing the importance of self-discipline and
personal responsibility (McAdams et al., 2008). Liberals are identified by their
agreeableness or compassion, whereas conservatives are characterized by politeness and
orderliness (Hirsh, DeYoung, Xu, & Peterson, 2010). Surprisingly, conservatives show a
1
greater well-being (measured by scales oflife satisfaction, sexual activity, marital status,
church attendance, employment status, education, and income) than do liberals (Napier &
Jost, 2008). The authors explain this difference by arguing that liberals are unable to
frame the negative effects associated with a lower socioeconomic status in a positive or
neutral manner (Napier & Jost, 2008). Overall, these differences shape how issues and
events are viewed. Additionally, these studies of partisan bias and motivated reasoning
have shown that if a partisan is confronted with information that is emotionally aversive
(i.e., it brings about feelings of anger or disgust), he or she will actively seek an
alternative conclusion (Westen et al., 2007). Ifliberals and conservatives approach issues
from different moral foundations, and have different governing philosophies, it follows
that worldviews could be exaggerated when presented with information that is
counterattitudinal. As more people look to cable news for their daily information, it is
important to consider whether the current polarization in news could be problematic. That
is, the ever-increasing divide reflected in cable news could make possible the
perpetuation of logically flawed yet highly believable information to the public.
Rational Choice and The Hostile Media Bias
The rational choice model, from a political science and economic perspective,
assumes that people act rationally when presented with a choice, weighing the costs and
the benefits and making the most utilitarian decision with the most benefits and least cost
(Lovett, 2006). From an early age, people are taught to act rationally or at least trained to
give the appearance of rationality in daily decision-making (Scott, 2006). However, from
a psychological perspective, the theory fits normal human behavior poorly. A slightly
2
different perspective in reinforcing future behavior is suggested: that of melioration, or
the process ofimproving one's choices overtime (Herrnstein, 1990).
Reinforcing the notion of improving decisions over time (melioration) is the claim
that consuming information from biased sources that are only compatible with a person's
worldview is performed in order to minimize errors in judgment and decision-making
(Calvert, 1985). If a person is constantly inundated with information that he or she does
not agree with, large errors might appear and create problems with discourse and
decision-making. However, it has also been suggested that viewing information that a
person opposes to a degree (without being hostile toward) is essential for a properly
functioning democracy (Mutz & Martin, 2001 ), and that consuming a greater number of
biased sources allows for improved information-gathering (Xiang & Sarvary, 2007).
However, because of this, cable news has become less homogenized and more slanted,
and framing of news stories to fit the slant has become the norm. It creates a power
struggle for being "fair," "balanced," or "most trusted." Proper argumentation and
discourse cannot survive in this type of environment. These activities are worsened when
perception becomes hostile and an accusation of bias becomes overt indignation with
overwhelmingly negative emotion (Hwang, Pan, & Sun, 2008). This can also be
described as a process of selective categorization, where liberals and conservatives apply
positive or negative emotions to events and facts in the same unbiased content differently
than their counterparts (Scmitt, Gunther, & Liebhart, 2004). In an environment full of
biased sources and hostility between opposing viewpoints arises the Hostile Media
Phenomenon (HMP).
3
The HMP was first identified during the mid-1980s in an experiment of people's
views on coverage of the Beirut Massacre (Vallone, Ross, & Lepper, 1985). The
experimenters found that people would claim that the news information was biased when
it contradicted their perspective on issues regarding the Middle East. The researchers
expanded it to account for any general tendency for partisans to claim bias of any source
that opposes their worldview. In other words, it is the pervasive belief that a media source
is biased if and only if the content contradicts a consumer's prior beliefs and attitudes.
However, at the time, the source of bias was identified as the consumer rather than the
media (Gunther, 1992). Additionally, the HMP has been identified in balanced news
stories, or stories free from slant or logical fallacy (Kim & Pasadeos, 2007). Now, the
political narratives chosen by these channels have influenced the commentators and
interviewers, creating a situation in which the interviewer has greater persuasive power
on the audience than a politician or expert (Babad, 2005). Many of these factors create a
highly partisan environment, where moderates and those not heavily involved in politics
cannot engage in public discourse and proper argumentation to discuss current and future
events (Eveland & Shah, 2003).
The Power of Persuasion
In the span of a day, people are faced with hundreds, if not thousands, of choices.
People have attitudes and beliefs on any number of topics and they are constantly
evolving and changing. Attitudes are defined as "general evaluations people hold in
regard to themselves, other people, objects, and issues. These general evaluations can be
based on a variety of behavioral, affective, and cognitive experiences, and are capable of
influencing behavioral, affective, and cognitive processes" (Petty & Cacioppo, 1986, p.
4
127). If attitudes are basic evaluations, it seems they can change on their own (evolve) or
change in the face of new information (persuasion).
Persuasion has been studied extensively (e.g., Cacioppo & Petty, 1979; Chaiken
& Eagly, 1976; Petty, Cacioppo, & Goldman, 1981). Early studies of attitude change
focused on the various media through which a persuasive message was presented. These
media consisted of written, audio, and audiovisual (video) messages, as the advent of the
Internet was years away. Audio and video were found to be more persuasive than written
messages, attributable to the non-message cues apparent in each medium, such as the
physical characteristics of the presenter, body language, or voice tone. Additionally,
messages that were easy to process (that is, fewer jargon words and shorter sentences)
were the most persuasive when given by audiovisual means (Chaiken & Eagly, 1976).
However, the persuasive power of the message should not be underestimated. If
the message's position on an issue that is proximal to a person's own position, then it is
said to be within the person's latitude of acceptance. As the distance between the
perceiver's beliefs and the position advocated in the message grows, then the message is
perceived as biased and eventually falls within the person's latitude of rejection. The
further the message's position from the perceiver's own, the less likely it is that attitude
change will occur (Hovland, Harvey, & Sherif, 1957).
The communicator of the persuasive message is an integral part for attitude
change. Likeable communicators are more persuasive through audio and video media,
suggesting that the effect is due to the nonverbal cues associated with audio and video
messages allowing faster and stronger attitude-shifting (Chaiken & Eagly, 1983). In
addition, the credibility of the message communicator affects people who are
5
unmotivated and unable to assess persuasive messages, with high source credibility more
persuasive to them than to people who are motivated and able to assess these messages
(Heesacker, Petty, & Cacioppo, 1983). These findings seem to suggest that the message
communicator (in this case, a news anchor or pundit) is as important in persuasion as the
message itself.
With a growing immersion in the bias-laden medium of cable news, people are
more likely to argue improperly and illogically reason through issues with others if
messages are highly persuasive and readily perpetuated. This is a worrisome trend. It is
easy to argue fallaciously while presenting truthful statements-the truth is not what is
important. A truthful argument without validity lacks strength, soundness, and cogency.
An important distinction between truth and validity would be to imagine the foundation
of a building as argument validity: the building and all that exists inside (truth) cannot
stand without a proper foundation. According to Hurley (2004), there are a number of
formal fallacies (which happen only to deductive arguments with an identifiable form)
and 22 informal fallacies that can occur in everyday argumentation (e.g., red herring, ad
hominem, straw man, etc.). Additionally, there are fallacies that are specific to syllogistic
reasoning (e.g., undistributed middle). Lurking behind many of these fallacies is the
belief bias, or the reasoning error in which a person judges an argument as logically valid
if the content is highly believable, even if the argument is logically invalid and/or false.
The Belief Bias in Syllogistic Reasoning
Syllogisms have been discussed since the time of Aristotle. They are the simplest
form of logical argumentation, and can take many different forms. For example, one
classic syllogism consists of the following: "All men are mortal. Socrates is a man.
6
Therefore, Socrates is mortal." Traditionally, a categorical syllogism contains two
premises (the first is termed the major premise; the second is termed the minor premise)
and a conclusion. These premises are linked by a middle term that appears in both
(Hurley, 2004). In deductive syllogisms, the two premises contain all the information
necessary and sufficient to draw a specific conclusion (Evans, Newstead, & Byrne,
1993). Due to their ease and structure, this thesis will focus on deductive categorical
(relating three categories together) syllogisms. In each of these, there is a major term, a
minor term, and a middle term. The conclusion contains the major and minor terms,
linked in the premises by the middle term (Evans, Barston, & Pollard, 1983). Evans et al.
describes four different forms (termed "figures") and combinations these terms can make
to form valid arguments. Note that the validity of an argument is based on its form, not its
content. For example, take the generic argument: "All A are B; All Bare C; Therefore,
all A are C". The items A and Care the major and minor terms, respectively, linked
together by the middle term, in this case, B. The argument is therefore valid; the actual
content of A, B, and Care not relevant to the argument's validity. If the Band Care
switched in the second premise, then the form is no longer valid, and the conclusion does
not necessarily follow from the premises, regardless of whether the premises are true.
Before a participant can judge validity, the argument's premises must be
evaluated for form and structure. There are two main theories of how syllogistic
arguments are represented in the brain. Mental Rules Theory ( cf. Braine, 1978) suggests
that when people evaluate arguments, they use the rules associated with that argument, in
this case, syllogisms using the words "all", "some are ... are not", and "no[ne]" to judge
validity (Goel, 2003; Gaiotti, Baron, & Sabini, 1986). The other is Mental Models
7
Theory, which suggests that people first formulate cognitive models of the premises and
then separately evaluate the conclusion multiple times to see if it fits these models
(Johnson-Laird, 1983; Johnson-Laird & Byrne, 1991; Johnson-Laird & Steedman, 1978).
The main difference between these two theories of reasoning is in the type of reasoning
that is being evaluated. Mental Rules Theory suggests that people tend to follow more
formal paths of reasoning, such as conditional reasoning ("if, then" statements; Braine,
1978), whereas Mental Models Theory is more attuned to looser reasoning, such as
categorical reasoning (e.g., deductively linking three items together; Johnson-Laird &
Steedman, 1978). According to Mental Rules Theory, errors result from not following the
rules oflogic (i.e., the structure or the words). According to Mental Models Theory,
errors result from arguments not being tested exhaustively, through each iteration of
possible combinations of premises and conclusions. Most of the time, deduction goes
from general to specific (Hurley, 2004), so Mental Models Theory posits that proper
argumentation is about "spatial manipulation and search"-that is, organizing the
iterations of an argument visually and spatially (Goel, 2003, p. 3). The two theories posit
different mechanisms in the brain to allow for argument evaluation and production:
Mental Rules should engage the brain's language system in the brain (usually during
abstract reasoning), while Mental Models should engage the visuo-spatial system (usually
during concrete reasoning). According to brain imaging done by Goel and colleagues
(Goel, 2003; Goel, 2007; Goel & Dolan, 2004; Vartanian, Goel, Tierney, Huey, &
Grafman, 2009), both systems work together in what they described as a dual mechanism
theory, whereby the content of the argument dictates the path of reasoning in the brain.
8
The belief bias, in various forms, has been the most closely examined (e.g., Evans
et al., 1983; Feather, 1964; Morgan & Morton, 1944) and has survived the most scrutiny
(Evans, Handley, & Harper, 2001; Newstead, Pollard, Evans, & Allen, 1992) of the errors
that can occur in syllogistic reasoning tasks. Two other explanations for errors in logical
reasoning have also drawn considerable interest: atmosphere effects and illogical
conversion. The atmosphere effect is described as the general mood (positive or negative,
universal or particular) of the premises dictating the valid conclusion (Begg & Denny,
1969; Sells, 1936; Sells & Koob, 1936; Woodworth & Sells, 1935). Illogical conversion
was an alternate explanation for errors made during reasoning tasks not attributable to the
atmosphere effect. This process was described as participants taking the premise "All A
are B" and interpreting/encoding it as "All Bare A" (Chapman & Chapman, 1959,
Revlis, 1975a, 1975b; Revlin, Ammerman, Petersen, & Leirer, 1978; Revlin & Leirer,
1978). However, both of these errors can be easily trained against and removed
(Dickstein, 1975; Simpson & Johnson, 1966). Beliefbias was the only factor for errors
that remained in content-laden arguments; thus, it is the perspective that likely governs
the arguments from cable news that will be considered in the current study.
Studies on the content of syllogisms affecting error making began in the mid-20th
century. Researchers argued that people were adept at logical reasoning (Henle, 1962) or
at least were competent if given simple training (Morgan & Morgan, 1953), and errors
developed due to prior beliefs and attitudes (Janis & Frick, 1943). Whether it was
studying agreeable or disagreeable premises (Janis & Frick, 1943) or premises meant to
elicit personal convictions (Morgan & Morton, 1944), the fact remained that "when an
issue is injected which relates to the personal opinions, wishes, fears, or convictions of an
9
individual, the distortion shifts ... to the meaning of the terms of the syllogism" (Morgan
& Morton, 1944, p. 59). Various aspects of the same problem have appeared in many
places. Agreeable positions to arguments are sought out, while disagreeable positions are
ignored (Dawson, Gilovich, & Regan, 2002).
Additionally, syllogisms were used to study personal prejudices and attitude
effects on reasoning, wherein highly emotional content based upon existing racial and
prejudicial beliefs will elicit falsely judged arguments (Thistlethwaite, 1950). Along the
same vein, ethnocentrism is positively correlated with rigidity in beliefs and less latitude
of movement in those beliefs. This, in tum, was positively correlated with poor abstract
reasoning ability (O'Connor, 1952). It makes sense: If someone believes very strongly
about something, he or she will make every effort to judge arguments containing those
beliefs as valid. Moreover, arguments to the contrary will be met with judgments of
invalidity, even if such arguments are equally valid. An additional early case of attitude
and belief biases apparent in reasoning tasks was an experiment in which high and low
dogmatic individuals were tested to evaluate religious syllogisms (Feather, 1964). Highly
dogmatic (or those participants who were rigid and unwavering in their thinking and
attitudes) and pro-religious participants tended to judge pro-religious arguments valid and
antireligious arguments as invalid. Not surprisingly, these participants also had low
critical reasoning ability. However, one difficulty remained in interpreting the results:
because the antireligious participants held only a slightly unfavorable view of religion as
opposed to the pro-religious participants, they did not implement the same evaluation
processes for the arguments. Essentially, the pro-religious were more in favor of religion
than the anti-religious were against religion. The findings suggests the more dogmatic
10
and rigid a person is in their attitudes, the less he or she will judge an argument as valid
the further it is from his or her unalterable worldview.
In recent research, the belief bias has been refined and elaborated. It has been
shown that arguments that are both invalid and believable are more likely to be judged
valid than arguments that are both valid and unbelievable (71% to 56%; Evans et al.,
1983; Evans et al., 2001). In addition, beliefbias has been differentiated into positive and
negative belief biases (Evans et al., 2001). Positive belief bias occurs when a normally
avoided fallacious argument is enoneously judged valid; negative belief bias occurs
when the fallacies are plausible (or even ignored), thus reducing validity rates for
unbelievable arguments. Further refmement has included production tasks of valid
conclusions, whereby the task requires the participant to input a logical conclusion to a
set of given premises (rather than evaluation of included conclusions). In these tasks, the
belief bias was found in the conclusions produced by the participants, with the incidence
of bias significantly greater in produced conclusions than evaluated conclusions
(Markovits & Nantel, 1989; Oakhill & Johnson-Laird, 1985).
In addition, dual processing research has been extended to belief bias. In dual
processing theory, System 1 processing is quick and prone to errors, whereas System 2
processing is slower and more effortful. In belief bias research, studies have found that
System 1 is typically controlled by the beliefs of a person, whereas System 2 is typically
controlled by logical rules (Evans, 2003, 2008). For example, Evans and Curtis-Holmes
(2005) performed an experiment with a syllogistic reasoning task that required either a
rapid response or a normal, free-time response. They found that the rapid responses
(System 1) had more errors indicative ofbeliefbias than the untimed condition (System 2
11
could be engaged during this condition). Evidence suggests that reasoners with a high
level of skill can engage System 2 more successfully than reasoners with a low skill level,
but not more often (Evans, 2008). Furthermore, people with high working memory show
less bias overall, but can still be susceptible to the bias if under heavy cognitive load (De
Neys, 2006). This suggests an overall trend toward System 1 processing while reasoning,
even if System 2 will lead one to the correct evaluation.
For a more realistic perspective, research on the belief bias in syllogistic
argumentation has moved toward using arguments that are less abstract and more
concrete in content (Morley et al., 2004). The implications in other situations may allow
for proper education of proper argumentation, or fundamentally change the way people
argue and debate in everyday situations.
The Present Study
The purpose of this thesis is to examine whether the belief bias effect is present in
paraphrased arguments made by actual cable news commentators, such as those pundits
and anchors from MSNBC and Fox News, which may account as a driving force in HMP.
If identification of argumentation errors can be found, then an effort can be made to
address these errors. This could conceivably result in a change in cable news from
entertainment-oriented to harder, fact-driven news, which in turn could decrease
partisanship and increase civil public discourse.
In the present experiment, participants encountered arguments in syllogistic form
paraphrased from commentators and pundits (e.g., Glenn Beck, Keith Olbermann, Bill
O'Reilly, and Rachel Maddow). They evaluated if arguments' conclusions followed
necessarily and logically from the premises. Half of the conclusions were valid and the
12
other half invalid. Additionally, half were liberal arguments and the rest were
conservative arguments. Arguments were either compatible or incompatible with the
participants' own political ideologies. It is hypothesized that liberal participants will rate
liberal arguments as more valid than conservative arguments, while conservative
participants will do the opposite; that is, compatible arguments will be more likely to be
viewed as valid, indicating a stronger overall bias.
13
Method
Participants
Ninety-eight (61.2% female) undergraduate psychology students participated in
this study. Each received course credit for their participation. Ages ranged from 18 to 51
years, with a mean of20.76 years. More than half (62.2%) of the participants stated that
they had not received any formal logic training (e.g., college courses). This trend is likely
due to many of the participants being one or two semesters out of high school, where
minimal emphasis is placed on logical reasoning.
Design and Materials
Arguments. The arguments used were paraphrased from cable news sources.
These sources include Fox News Channel, CNN and MSNBC. The commentators
included Glenn Beck, Keith Olbermann, Bill O'Reilly, Sean Hannity, Roland Martin and
Rachel Maddow, among others. Sixteen arguments were used, halfliberal and the other
half conservative. The context of each argument was established by the overall political
ideology or stated affiliation ofthe pundit or commentator. Half of each set were valid
and the other invalid. The arguments varied in mood, incorporating universal affirmatives
and negatives as well as particular affirmatives and negatives. The purpose of the mixture
of syllogism type was to prevent atmosphere effects. Thus, there were four types of
argument: liberal-valid (LV), liberal-invalid (LI), conservative-valid (CV), and
conservative-invalid (CI). Appendix B contains a complete list of all arguments used in
this study.
Argument Pretest. A pretest of the believability of the conclusions of the
arguments was performed on a large social psychology class (n = 115). To achieve the
14
best selection, 24 total arguments were created, with eight intended to be discarded after
the pretest. On a scale of 0- 10 (1 0 representing maximum believability), participants
were shown only the conclusions of the arguments (given to them as "statements") and
asked to record how plausible the statement was to them and to others. The 16 best
conclusions were chosen by correlating political identity (also asked of the pretest
participants) with believability scores. The strongest correlations of each conclusion type
were chosen (four of each). A mean believability score was computed for each participant
for liberal and conservative conclusions. A two-way Analysis ofVariance (ANOVA) of
the composite scores and political identity revealed a significant interaction, F(l, 83) =
11.17, p = .00 1. Those indicating a liberal identity tended to believe liberal conclusions
more and conservative conclusions less than the participants indicating a conservative
identity. The opposite was true for conservatives and conservative vs. liberal conclusions.
Thus, the 16 chosen conclusions fit the believability manipulation and were used in the
main task.
Design. Political ideology of the participant, rated on a scale from 1 to 7, 1
representing extreme liberalism and 7 representing to extreme conservatism, was
prescreened prior to involvement in the study. To ensure maximum partisanship
evaluation on each argument, subject pool students rating themselves as moderates (based
on the sample distribution, a response of 3 ), were excluded from participation. A
response of 4 was included for comparison to the liberal and conservative poles.
Recruitment was achieved by conducting three separate sub-studies, with a specific
response requirement for each. The design of this experiment was a three-group
comparison on the four argument types; one independent variable was the political
15
identity of the participant (a pre-existing quasi-experimental variable) and the other
independent variable was whether the argument was valid or invalid. The dependent
measure was calculated as the number of conclusions accepted as valid.
All materials were presented using Media Lab 2008 on Windows PCs. Personal
data were not collected by the program.
Procedure
Two participants entered the laboratory space and were instructed to sit at a
computer station. Each participant worked independently while evaluating each
argument. All of the instructions necessary to complete the task were displayed prior to
starting the evaluation task. The instructions were as follows:
Welcome to our study. This is an experiment to test people's reasoning ability. You will be given sixteen problems in total. For each problem you will be shown two statements and you are asked if a certain conclusion (given below the statements) may be logically deduced from them. You should answer this question on the assumption that the two statements are, in fact, true. If, and only if, you judge the conclusion necessarily follows from the statements, you should select "valid" using the mouse, otherwise select "invalid." Please take your time and be sure you have the right answer before giving your response.
The arguments displayed on the computer screen with a choice of either "valid"
or "invalid." The order of arguments was randomized. On each screen, a single argument
appeared. After the participant completed all 16 arguments, the final screen asked basic
demographic information such as age and gender, including if he or she had any previous
formal logic training. A posttest question using the same political ideology scale was also
used. If asked for help, the experimenter was instructed to state they could not give any
assistance. All participants were debriefed upon completion and thanked for their
participation.
16
Results
Preliminary Analyses
Prior to testing group differences, preliminary analyses were performed. Political
identity was prescreened prior to participation in the task (the pretest score). A check of
political identity was assessed following the argument validity task. The two scores were
highly correlated (r = .70) and the means were not significantly different from each other
(t(97) = -.36,p = .72); however, some participants did change their answers between the
time of the pretest and the time of the task. An average of the scores was utilized for all
subsequent analyses. The average political identity score was then grouped into nominal
political identity, with scores of 1.00 to 2.99 classified as "liberal," scores of 3.00 to 4.49
classified as "moderate," and scores of 4.50 to 7.00 classified as "conservative." The
coding was performed this way to match the distribution of this sample and the overall
characteristics of the university.
The 16 individual arguments were quite varied in their acceptance rates. As
shown in Table 1, a larger rate for the valid arguments indicated greater accuracy,
whereas a larger rate for the invalid arguments indicated a greater tendency for error. One
argument, a conservative-invalid, had a mere 7% acceptance rate, while two liberal-valid
arguments had an 84% acceptance rate, indicating a greater accuracy overall. The
conservative-invalid argument had a moderate correlation during pretest as a conservative
conclusion (r = .24). However, the liberal-valid conclusions suffered from very weak
correlations as liberal conclusions (rs = -.09). This might explain why some arguments
were more easily accepted as valid than others. Most arguments, however, fell in the 50%
to 70% acceptance range.
17
For the main analysis, the 16 arguments were condensed into sub-scores based on
argument type. Thus, there were four sub-scores: CV, LV, CI, and LI. A participant could
have a score ranging from zero to four on each sub-score, indicating how many times a
conclusion was accepted as valid. Three indices were then calculated based on these sub
scores (per Evans & Curtis-Holmes, 2005). The first index measured a participant's
logical response to the arguments. It combines the acceptance scores of valid conclusions
minus the acceptance scores of the invalid conclusions (CV +LV- LI- CI). The higher
this score, the more logical the reasoning attempts by the participant to accept only valid
conclusions. The second index was the belief index. Due to the overwhelming political
ideology of the sample, a liberal perspective was chosen for this index. It assumes that all
liberal arguments are believable while all conservative arguments are unbelievable (LV +
LI- CV- CI). The higher the number, the stronger the liberal bias is, and the more belief
bias might be accounting for this liberal bias (a conservative belief index was calculated
and it was the direct inverse of the liberal belief index). The third index was also
calculated from a liberal perspective, and it determines the interaction of belief and
political ideology, specifically the desire to choose invalid over valid conclusions (CV +
LI- LV- CI). The larger the number is on this index means that, from a liberal
perspective, the participant showed stronger belief bias on invalid than valid conclusions.
One-sample t-tests were conducted on each of these indices to determine if they
were significantly different from zero (a value that has no bias). Both the logic index and
the liberal belief index were significantly greater than zero (ps < .001), suggesting that
most participants were successful on the reasoning task and there was a tendency for
liberal bias, respectively. The interaction index was also significantly different from zero
18
(p = .026), suggesting participants overall tend to show belief bias on valid conclusions
more than invalid conclusions from a liberal perspective.
Separating the data by whether or not the participant had any formal logic
training, some significant differences between liberals and conservatives did arise. For
those participants stating they had never received formal logic training (n = 61),
conservatives (n = 21) tended to accept conservative-valid arguments (t(43) = -2.72,p <
.01, Cohen's d= .82), as well as conservative-invalid arguments (t(43) = -2.14,p = .04,
Cohen's d = .64), more than liberals (n = 24). Interestingly, they also tended to accept
more liberal-invalid arguments (t(43) = -2.75,p < .01, Cohen's d = .81). For those
participants stating they had some form of formal logic training (n = 37; conservative, n =
15; liberal, n = 13), each ofthese effects disappear. None of the indices showed any
significant belief bias effects.
Primary Analyses
A series of Independent-Samples t-tests were conducted to determine if liberals (n
= 37) and conservatives (n = 36) differed on the three belief bias indices. Liberals and
conservatives did not differ significantly on any of the indices (illustrated by Figure 1),
but the liberal belief index was in the predicted direction, with liberals showing a stronger
liberal bias (M= 1.78, SD = 1.67) than conservatives (M= 1.25, SD = 1.30; t(71) = 1.52,
p = .13, Cohen's d = .35). Liberals tended to have a stronger logic index score (M = 1.73,
SD = 1.37) than conservatives (M = 1.42, SD = 1.50). Liberals and conservatives showed
no overall difference on the interaction index, suggesting a lack of belief bias tendency
on invalid conclusions over valid conclusions; in fact, the opposite was trending: liberals,
M= -0.43 (SD = 1.72) vs. conservatives, M= -0.41 (SD = 1.83).
19
To determine overall group differences on the sub-scores, another set of
Independent-Samples t-tests was conducted. Liberals and conservatives did not differ
significantly on any of the sub-scores, but acceptance rates were in the predicted
directions for conservative sub-scores (CV and CI). As shown in Figure 2, conservatives
tended to have higher acceptance rates (M= 2.25, SD = .91) than liberals (M= 2.00, SD =
.91) on conservative-valid arguments (t(71) = -1.17,p = .24, Cohen's d = .27). This is
also shown on conservative-invalid arguments, which is marginally significant (t(71) =-
1.68,p = .09, Cohen's d = .39). However, these trends are not shown on the liberal sub
scores, with liberals (M= 3.11, SD = .94) slightly higher on liberal-valid arguments than
conservatives (M= 3.08, SD = .87), while the reverse trend on liberal-invalid arguments
shows conservatives (M = 2.17, SD = .91) higher than liberals (M = 2.03, SD = .93).
Secondary Analyses
Pmiicipants identifying themselves as "moderates" behaved similarly to those
who identified themselves as "liberals", not significantly changing the comparisons of
liberals and conservatives. Additionally, post-hoc analyses were conducted to determine
differences in participant scores using demographic variables.
Dividing the data by gender revealed that liberal and conservative men and liberal
and conservative women do not differ in their reasoning abilities, nor do they differ when
all males and females are compared.
Reaction time was recorded for each argument and then averaged across all 16
arguments. Conservatives, on average, took approximately 5-6 seconds longer to make a
decision (t(71) = -2.57,p = .01, Cohen's d= .60). However, reaction time was not
correlated with overall performance on the task.
20
Discussion
In the investigation to determine if political liberals and conservatives succumb to
the belief bias effect in syllogisms containing content from cable news pundits and
commentators, the hypothesis was partially supported. More specifically, liberals and
conservatives did not differ significantly in their acceptance rates of the four. Overall,
conservatives accepted valid arguments of a conservative slant 56% of the time and
invalid arguments 44% of the time, while liberals accepted these arguments 50% and
34%, respectively. Liberal arguments seemed to affect conservative participants more
than liberals in this instance. Liberals did not perform as predicted. However, from a
liberal perspective on the index of belief bias, liberals had a stronger bias than that of
conservatives; the same can be reported for conservatives on the inverse index of belief
bias.
The effects can be attributed to those who stated they had not received any formal
logic training, or those who would be considered novice reasoners. These reasoners lack
the skills to determine accurately an argument's validity. As evidenced by the exclusion
ofthose with training, the overall marginal effects on the conservative arguments were
stronger. One interpretation for this trend is that beliefs and attitudes overwhelmingly
control evaluation of arguments for those with no training.
Although not as marked, taken together, these results suggest a trend toward the
results seen in Evans et al. (1983) and Evans and Curtis-Holmes (2005), and an important
first step in determining the overall impact of cable news in the United States and on
political partisans in particular. Liberal participants tended to accept liberal arguments as
valid while conservative participants tended to accept conservative arguments as valid.
21
With myriad arguments on cable news every hour of every day, it is difficult to pinpoint
when and how often beliefbias or persuasion is overriding the rational aspect of political
debate. Liberals and conservatives are shown to be fundamentally different (Graham et
al., 2009; Hirsh et al., 2010; McAdams et al., 2008; Napier & Jost, 2008), and it is
possible each approach argumentation differently and are more willing to accept
congruent arguments over incongruent arguments. While affect was not tested directly,
this willingness can alter perceptions of media sources and instigate HMP (Hwang et al.,
2008).
However, there were limitations to the study that had significant effects on the
data produced. Perhaps the main issue was the argument collection and production. These
materials, paraphrased and converted into syllogisms from full prose arguments,
prevented uniformity in political slant of the arguments. Some arguments were easier to
accept as valid, while others proved more difficult. Due to this high variance, the
difference between liberals and conservatives had little variance on argument subtypes.
Additionally, many of the arguments were created almost a year prior to their
presentation, so it is possible that the timeliness of the "current events" were no longer in
the minds of the participants. For example, some arguments were taken from the 2008
election campaign season, the debate over torture, and the 2009 health care debate. Many
of these issues have been discussed thoroughly or are no longer relevant. Although the
arguments passed the believability manipulation in the pretest, their overall correlations
with the intended political slant were small. This might have led to dynamic political
attitudes (of the participants) on specific worldview issues; a liberal argument might have
been plausible enough to appeal to a conservative participant and vice versa.
22
Furthermore, the sample used was from a large and diverse California university,
skewing the political identity distribution toward liberalism. The "conservatives" in the
sample, while the minority, are consistently exposed to liberal ideology, especially on
this particular college campus, known for its activism and promotion of diversity. In
addition, the participants were college students, with often changing political attitudes or
attitudes that are an eclectic mix based on individual issues, stemming from parental
attitudes or the fledgling independent attitudes of a new adult. It is not far-fetched to
suggest this sample is not as well defined as the divide in the US House of
Representatives or the US Senate. The skewness toward liberalism is strong and
particularly difficult to separate partisans on a continuum.
Future research can address both of these limitations, as well as others.
Arguments should begin as written prose taken verbatim from the cable news pundits or
commentators, substituted carefully into both valid and invalid syllogisms (or only
invalid), and then the conclusions thoroughly tested for their believability and whether
the correlations are strong enough with political ideology and the intended slant of the
argument. To achieve a more representative sample of political partisans, placing the task
in an online setting is recommended. Both greater sample size and partisanship is
intended by this method. A smaller issue to modify would be to change the question
asked of participants during the task, from "valid" and "invalid" to "yes" and "no", which
might have a subtle effect in the thought process of the participants.
Other paths of research in this topic could include associating the argument with
the speaker, or manipulating argument with speaker, to test ifbeliefbias is overridden by
the suggestions of the pundit and the associations that might be linked. For example,
23
Glenn Beck could be associated with unearthing left-wing propaganda arguments, which
could either strengthen or weaken the belief bias effect. Moreover, an effort to verify
political identity on the traditional 7 -point scale could be instituted with a collection of
scales on political ideology (e.g., Weber & Federico, 2007).
This investigation has shown the belief bias effect on reasoning from arguments
presented on cable news channels, particularly with individuals without logic training. If
politically active Americans continue to watch and place importance on these networks
and those who speak on the networks, it stands to reason that a political partisan must be
cognizant of the powerful forces of belief congruency and persuasion on these channels.
Blindly accepting an argument as valid due to its believability adds to fallacious
argumentation, perpetuating HMP and increasing the political divide. While it appears
the majority of audience members might not have the aptitude to reason through each
argument heard or read, seeking incongruent arguments or counterexamples has its
benefits.
24
References
Babad, E. (2005). The psychological price of media bias. Journal of Experimental
Psychology: Applied, 11(4), 245-255.
Begg, I. & Denny, J.P. (1969). Empirical reconciliation of atmosphere and conversion
interpretations of syllogistic reasoning errors. Journal of Experimental
Psychology, 81(2), 351-354.
Braine, M.D. S. (1978). On the relation between the natural logic of reasoning and
standard logic. Psychological Review, 85(1 ), 1-21.
Cacioppo, J. T. & Petty, R. E. (1979). Effects of message repetition and position on
cognitive response, recall, and persuasion. Journal of Personality and Social
Psychology, 37(1), 97-109.
Calvert, R. L. (1985). The value of biased information: A rational choice model of
political advice. The Journal of Politics, 47(2), 530-555.
Chaiken, S. & Eagly, A. H. (1976). Communication modality as a determinant of
message persuasiveness and message comprehensibility. Journal of Personality
and Social Psychology, 34(4), 605-614.
Chaiken, S. & Eagly, A. H. (1983). Communication modality as a determinant of
persuasion: The role of communicator salience. Journal of Personality and Social
Psychology, 45(2), 241-256.
Chapman, L. J. & Chapman, J.P. (1959). Atmosphere effect re-examined. Journal of
Experimental Psychology, 58(3), 220-226.
25
Coe, K., Tewksbury, D., Bond, B. J., Drogos, K. L., Porter, R. W., Yahn, A., & Zhang,
Y. (2008). Hostile news: Partisan use and perceptions of cable news
programming. Journal ofCommunication, 58, 201-219.
Dawson, E., Gilovich, T., & Regan, D. T. (2002). Motivated reasoning and performance
on the Wason selection task. Personality and Social Psychology Bulletin, 28(10),
1379-1387.
De Neys, W. (2006). Dual processing in reasoning: Two systems but one reasoner.
Psychological Science, 17(5), 428-433.
Dickstein, L. S. (1975). Effects of instructions and premise order on errors in syllogistic
reasoning. Journal of Experimental Psychology: Human Learning and Memory,
104(4), 376-384.
Evans, J. St. B. T. (2003). In two minds: Dual-process accounts ofreasoning. TRENDS in
Cognitives Sciences, 7(1 0), 454-459.
Evans, J. St. B. T. (2008). Dual-processing accounts of reasoning judgment and social
cognition. Annual Review of Psychology, 59, 255-278.
Evans, J. St. B. T., Barston, J. L., & Pollard, P. (1983). On the conflict between logic and
belief in syllogistic reasoning. Memory & Cognition, 11 (3), 295-306.
Evans, J. St. B. T. & Curtis-Holmes, J. (2005). Rapid responding increases belief bias:
Evidence for dual-process theory of reasoning. Thinking and Reasoning, 11(4),
382-389.
Evans, J. St. B. T., Handley, S. J., & Harper, C. N.J. (2001). Necessity, possibility and
belief: A study of syllogistic reasoning. The Quarterly Journal of Experimental
Psychology, 54A(3), 935-958.
26
Evans, J. St. B. T., Newstead, S. E., & Byrne, R. M. J. (1993). Human reasoning:
Psychology of deduction. East Sussex, UK: Lawrence Erlbaum Associates, Ltd.
Eveland, W. P., Jr., & Shah, D. V. (2003). The impact of individual and interpersonal
factors on perceived news media bias. Political Psychology, 24(1), 101-117.
Feather, N. T. (1964). Acceptance and rejection of arguments in relation to attitude
strength, critical ability, and intolerance of inconsistency. Journal of Abnormal
and Social Psychology, 69(2), 127-136.
Gaiotti, K. M., Baron, J., & Sabini, J.P. (1986). Individual differences in syllogistic
reasoning: Deduction rules or mental models? Journal of Experimental
Psychology: General, 115(1), 16-25.
Goel, V. (2003). Evidence for dual neural pathways for syllogistic reasoning.
Psychologia, 32, 301-309.
Goel, V. (2007). Anatomy of deductive reasoning. Trends in Cognitive Sciences, 11 (1 0),
435-441.
Goel, V. & Dolan, R. (2004). Differential involvement of left prefrontal cortex in
inductive and deductive reasoning. Cognition, 93, B109-B121.
Graham, J., Haidt, J., & Nosek, B. A. (2009). Liberals and conservatives rely on different
sets of moral foundations. Journal of Personality and Social Psychology, 96(5),
1029-1046.
Gunther, A. C. (1992). Biased press or biased public? Attitudes toward media coverage
of social groups. Public Opinion Quarterly, 56, 14 7-167.
Haidt, J. & Graham, J. (2007). When morality opposes justice: Conservatives have moral
intuitions that liberals may not recognize. Social Justice Research, 20(1 ), 98-116.
27
Haidt, J. & Joseph, C. (2004). Intuitive ethics: How innately prepared intuitions generate
culturally variable virtues. Daedalus, 133(4), 55-66.
Heesacker, M., Petty, R. E., & Cacioppo, J. T. (1983). Field dependence and attitude
change: Source credibility can alter persuasion by affecting message-relevant
thinking. Journal of Personality, 51(4), 653-666.
Henle, M. (1962). On the relation between logic and thinking. Psychological Review,
69(4), 366-378.
Hernnstein, R. J. (1990). Rational choice theory: Necessary but not sufficient. American
Psychologist, 45(3), 356-367.
Hirsh, J. B., DeYoung, C. G., Xu, X., & Peterson, J. B. (2010). Compassionate liberals
and polite conservatives: Associations of agreeableness with political ideology
and moral values. Personality and Social Psychology Bulletin, 36(5), 655-664.
Hovland, C. I., Harvey, 0. J., & Sherif, M. (1957). Assimilation and contrast effects in
reactions to communication and attitude change. Journal of Abnormal and Social
Psychology, 55(2), 244-252.
Hurley, P. J. (2004). A concise introduction to logic: Selected chapters. Belmont, CA:
Wadsworth Thomson Learning.
Hwang, H., Pan, Z., & Sun, Y. (2008). Influence of hostile media perception in
willingness to engage in discursive activities: An examination of mediating role
of media indignation. Media Psychology, 11, 76-97.
Janis, I. R. & Frick, F. (1943). The relationship between attitudes toward conclusions and
errors in judging logical validity of syllogisms. Journal of Experimental
Psychology, 33(1), 73-77.
28
Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language,
inference, and consciousness. Cambridge, Mass.: Harvard University Press.
Johnson-Laird, P. N. & Byrne, R. M. J. (1991). Deduction. Hillsdale, NJ: Lawrence
Erlbaum Associates, Ltd.
Johnson-Laird, P. N. & Steedman, M. (1978). The psychology of syllogisms. Cognitive
Psychology, 10, 64-99.
Kim, K. S. & Pasadeos, Y. (2007). Study of partisan news readers reveals hostile media
perceptions of balanced stories. Newspaper Research Journal, 28(2), 99-106.
Lovett, F. (2006). Rational choice theory and explanation. Rationality and Society, 18(2),
237-272.
Markovits, H. & Nantel, G. (1989). The belief-bias effect in the production and
evaluation of logical conclusions. Memory & Cognition, 17(1 ), 11-17.
McAdams, D.P., Albaugh, M., Farber, E., Daniels, J., Logan, R. L., & Olson, B. (2008).
Family metaphors and moral intuitions: How conservatives and liberals narrate
their lives. Journal of Personality and Social Psychology, 95( 4), 978-990.
Morely, N.J., Evans, J. St. B. T., & Handley, S. J. (2004). Belief bias and figural bias in
syllogistic reasoning. The Quarterly Journal of Experimental Psychology, 57 A( 4),
666-692.
Morgan, J. J. B. & Morton, J. T. (1944). The distortion of syllogistic reasoning produced
by personal convictions. The Journal of Social Psychology, 20, 39-59.
Morgan, W. J. & Morgan, A. B. (1953). Logical reasoning: With and without training.
Journal of Applied Psychology, 37(5), 399-401.
29
Mutz, D. C. & Martin, P. S. (2001). Facilitating communication across lines of political
difference: The role of mass media. American Political Science Review, 95(1), 97-
114.
Napier, J. L. & Jost, J. T. (2008). Why are conservatives happier than liberals?
Psychological Science, 19(6), 565-572.
Newstead, S. E., Pollard, P., Evans, J. St. B. T., & Allen, J. L. (1992). The source of
belief bias effects in syllogistic reasoning. Cognition, 45, 257-284.
O'Connor, P. (1952). Ethnocentrism, "intolerance of ambiguity", and abstract reasoning
ability. Journal of Abnormal and Social Psychology, 47(2 Suppl.), 526-530.
Oakhill, J. V. & Johnson-Laird, P. N. (1985). The effects of belief on the spontaneous
production of syllogistic conclusions. The Quarterly Journal of Experimental
Psychology, 37A, 553-569.
Petty, R. E. & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of persuasion.
In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp.
123-205). New York: Academic Press.
Petty, R. E., Cacioppo, J. T., & Goldman, R. (1981). Personal involvement as
determinant of argument-based persuasion. Journal of Personality and Social
Psychology, 41(5), 847-855.
Revlin, R. & Leirer, V. (1978). The effects of personal biases on syllogistic reasoning:
Rational decisions from personalized representations. In R. Revlin & R. E. Mayer
(Eds.), Human reasoning. Washington, D.C: Winston-Wiley.
Revlin, R., Ammerman, K., Petersen, K., & Leirer, V. (1978). Category relations and
syllogistic reasoning. Journal of Educational Psychology, 70(4), 613-625.
30
Revlis, R. (1975a). Syllogistic reasoning: Logical decisions from a complex data base. In
R. Falmagne (Ed.), Reasoning: Representation and process. Hillsdale, N.J.:
Erlbaum.
Revlis, R. (1975b). Two models of syllogistic reasoning: Feature selection and
conversion. Journal of Verbal Learning and Verbal Behavior, 14, 180-195.
Scmitt, K. M., Gunther, A. C., & Liebhart, J. L. (2004). Why partisans see media as
biased. Communication Research, 31,623-641.
Scott, J. (2000). Rational choice theory. In G. Browning, A. Halcli, & F. Webster (Eds.),
Understanding contemporary society: Theories of the present (pp.126-138).
London: Sage Publications, Ltd.
Sells, S. B. (1936). The atmosphere effect: An experimental study of reasoning. Archives
of Psychology, 29, 3-72.
Sells, S. B. & Koob, H. F. (1937). A classroom demonstration of"atmosphere effect" in
reasoning. Journal of Educational Psychology, 28(7), 514-518.
Simpson, M. E. & Johnson, D. M. (1966). Atmosphere and conversion errors in
syllogistic reasoning. Journal of Experimental Psychology, 72, 197-200.
Thistlethwaite, D. (1950). Attitude and structure as factors in the distortion of reasoning.
Journal of Abnormal and Social Psychology, 45(3), 442-458.
Tormala, Z. L. & Petty, R. E. (2002). What doesn't kill me makes me stronger: The
effects of resisting persuasion on attitude certainty. Journal of Personality and
Social Psychology, 83(6), 1298-1313.
31
Vallone, R. P., Ross, L., & Lepper, M. R. (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), 577-585.
Vartanian, 0., Goel, V., Tierney, M., Huey, E. D., Grafman, J. (2009). Frontotemporal
dementia selectively impairs transitive reasoning about familiar spatial
environments. Neuropsychology, 23(5), 619-626.
Weber, C. & Federico, C. M. (2007). Interpersonal attachment and patterns of ideological
belief. Political Psychology, 28(4), 389-416.
Westen, D., Blagov, P. S., Harenski, K., Kilts, C., & Hamann, S. (2006). Neural bases of
motivated reasoning: An fMRI study of emotional constraints on partisan political
judgment in the 2004 U.S. presidential election. Journal of Cognitive
Neuroscience, 18(11), 1947-1958.
Woodworth, R. S. & Sells, S. B. (1935). An atmosphere effect in formal syllogistic
reasoning. Journal of Experimental Psychology, 18(4), 451-460.
Xiang, Y. & Sarvary, M. (2007). News consumption and media bias. Marketing Science,
26(5), 611-628.
32
Appendix A: Tables and Figures
Table 1
Acceptance Rates of Conclusions with Intended Political Slant and Validity
Arguments Acceptance as
Valid Conclusion
Argument Conclusion Political Slant Validity Rate SD
Therefore, some torture is vital Conservative Valid 0.6 0.49
for America's safety.
Therefore, no TV reporting is on Conservative Valid 0.53 0.5
expansion of government.
Therefore, Barack Obama is anti-Conservative Valid 0.13 0.34
American. Therefore, some Obama followers do not understand Obama' s Conservative Valid 0.82 0.39 thought process. Therefore, some Congresspeople
Liberal Valid 0.84 0.37 are not for Obama's policies.
Therefore, some endorsements of protests against the government Liberal Valid 0.69 0.46 are by Fox News hosts.
Therefore, all guns are harmful. Liberal Valid 0.73 0.44
Therefore, some things that are beneficial for corporations would
Liberal Valid 0.84 0.37 not be supported by liberal politicians.
Therefore, all illegal immigrants Conservative Invalid 0.56 0.5
should be deported.
Therefore, global warming has Conservative Invalid 0.27 0.44 decreased.
Therefore, Barack Obama is not Conservative Invalid 0.07 0.26
for free speech. Therefore, all abortion is killing
Conservative Invalid 0.62 0.49 life. Therefore, JP Morgan Chase is not readily disclosing details Liberal Invalid 0.24 0.43 about bailouts. Therefore, health care reform is
Liberal Invalid 0.59 0.49 protecting the people. Therefore, some Obama presidency deniers are not
Liberal Invalid 0.65 0.48 demanding Obama' s birth certificate.
33
Therefore, some bankers were damaged by corporate greed.
Liberal Invalid 0.67 0.47
Note. Rates represent the average rate participants accepted each conclusion as valid.
Rates either show a trend of correct or incorrect responses by using the Validity column
and comparing it to the rate.
34
• Liberal ~ Conservative
2.5
2
1.5
1
0.5
0 Liberal Belief Index Logic Index
-0.5
-1
Figure 1. Mean differences between liberals and conservatives on the three belief bias
indices. The belief index is taken from a liberal perspective. No significant differences
were found between liberals and conservatives on the indices. Standard errors are
represented in the figure by the error bars attached to each column.
35
• Liberal 11 Conservative
3.5
3
0.5
0 Conservative-Valid Liberal-Valid Conservative-Invalid Liberal-Invalid
Argument Type
Figure 2. Mean acceptance scores (out of a possible score of 4) for each argument type.
Scores were in the predicted direction for the conservative arguments, but the liberal
arguments were unclear in trending scores. Standard errors are represented in the figure
by the error bars attached to each column.
36
Appendix B: Experimental Arguments
Legend: C = Conservative argument, L = Liberal argument; V = Valid, I = Invalid; (Mood)
All coerced interrogation techniques are vital for America's safety. Some coerced interrogation techniques are torture. Therefore, some torture is vital for America's safety. CVl (All)
All expansion of government is a left-wing conspiracy. No TV reporting is on left-wing conspiracies. Therefore, no TV reporting is on expansion of government. CV2 (AEE)
Barack Obama attended Jeremiah Wright's church for 20 years. Jeremiah Wright is anti-American. Therefore, Barack Obama is anti-American. CV3 (AAA)
All illegal immigrants are breaking the law. Deportation is the only punishment for breaking this law. Therefore, all illegal immigrants should be deported. en (AAA)
Global warming causes extreme heat deaths. Deaths from extreme heat have decreased in recent years. Therefore, global warming has decreased. CI2 (AAA)
Some Democrats are not for free speech. Barack Obama is a Democrat. Therefore, Barack Obama is not for free speech. CI3 (IAA)
All abortion is murder. Killing life is murder. Therefore, all abortion is killing life. CI4 (AAA)
Some banks are not readily disclosing details about bailouts. JPMorgan Chase is a bank. Therefore, JP Morgan Chase is not readily disclosing details about bailouts. Lll (IAA)
37
Health care reform is good for this country. Protecting the people is good for this country. Therefore, health care reform is protecting the people. LI2 (AAA)
Some Tea Party members are not demanding Obama's birth certificate. All Obama presidency deniers are Tea Party members. Therefore, some Obama presidency deniers are not demanding Obama's birth certificate. LI3 (OAO)
No Republicans are for Obama's policies. Some Congress people are Republicans. Therefore, some Congresspeople are not for Obama's policies. LVl (EIO)
Some Fox News hosts are promoting the Tea Party movement. All promotion of the Tea Party movement is an endorsement of protests against the government. Therefore, some endorsements of protests against the government are by Fox News hosts. LV2 (IAI)
All members of the middle class were damaged by corporate greed. Some bankers are not members of the middle class. Therefore, some bankers were damaged by corporate greed. LI4 (AOI)
Some working Americans do not understand Obama's thought process. Some Obama followers are working Americans. Therefore, some Obama followers do not understand Obama's thought process. CV4 (010)
All guns are associated with violence and force. All violence and force is harmful. Therefore, all guns are harmful. LV3 (AAA)
No liberal politician supports exporting jobs. Exporting some jobs is beneficial for corporations. Therefore, some things that are beneficial for corporations would not be supported by liberal politicians. LV4 (EIO)
38