This paper was presented at the Association for Education in Journalism and
Mass Communication in San Antonio, Texas August 2005.
         If you have questions about this paper, please contact the author
directly. If you have questions about the archives, email
rakyat [ at ] eparker.org. For an explanation of the subject line, 
send email to
[log in to unmask] with just the four words, "get help info aejmc," in the
body (drop the "").

(Jan 2006)
Thank you.
Elliott Parker
====================================================================

What makes deliberation possible?
Structural Equation Models of online and face-to-face deliberation process

by

Yun Jung Choi
Doctoral Student

Hyo Jung Kim
Master's Student
S. I. Newhouse School of Public Communications
Syracuse University

Contact:
Yun Jung Choi
S. I. Newhouse School of Public Communications
Syracuse University
215 University Place
Syracuse, NY 13244-2100
[log in to unmask]




*Student Paper









Manuscript submitted to the Communication Theory & Methodology Division
of the 2005 AEJMC, San Antonio, TX


What makes deliberation possible?
A test of online and face-to-face deliberation process

Abstract
This paper identifies factors that influence the deliberation process 
by proposing a deliberation model. Six ideas, reciprocity, reasoned 
discourse, freedom of expression, open-mindedness, empathy and public 
interest were recognized as concepts that compose deliberation 
construct, and deliberation was measured based on the six concepts. 
The study found that those with high political interest are likely to 
engage in political discussion and seek diverse political views 
different from their own, and this lead them to exchange their 
political opinions with those who have different political views from 
them. A news concept, exposure to diverse political views was 
introduced as a strong mediating variable that influence the 
deliberation in the study. Internet media are found to have more 
impact on the deliberation process than the traditional media.
As increasing number of people are discussing politic on the 
internet, online deliberation model is also proposed. The online 
deliberation model is similar to the face-to-face deliberation model, 
but less variance in the online deliberation construct is explained 
by the model. This indicates that there are more factors influencing 
the online deliberation than the face-to-face deliberation.






Deliberation is a political process in which diverse ideas are 
exchanged by participants of diverse backgrounds, which can 
contribute to more informed and legitimate political decisions among 
the public (Fishkin, 1995; Cohen, 1989). Deliberation is considered 
as an integral process of democracy; however, whether the 
deliberation process is taking place in the real world is 
questionable because people like to talk with others who share 
similar views with themselves and avoid discussing with opponents 
(Jonas, Schulz-Hardt, Frey, & Thelen, 2001). This study provides a 
model of deliberation by identifying factors that lead to the 
deliberation process: how political interest, political knowledge, 
media use, and frequency of interpersonal discussion influence the 
public to engage in the deliberation process.
As an increasing number of people exchange their political viewpoints 
with unknown others on the internet, the potential of online 
deliberation is getting more attention. It is assumed by many that 
online media and discussion forums allow more individuals and 
minority groups to speak out, thus exposing the public to a more 
diverse spectrum of views. Online discussions and forums are believed 
to facilitate the online deliberation process. However, others are 
concerned that online media and discussion forums create a fragmented 
environment which encourages partisan cleavages and group 
polarization (Jones, 2002).
The deliberation concept is linked with and provides the basis for 
many political and social constructs, such as social capital, social 
trust, political tolerance, and democracy (Mutz, 2002, Putnam, 2000). 
However, the concept of deliberation is often used abstractly and has 
not been thoroughly explicated. This paper identifies six dimensions 
– reciprocity, reasoned discourse, freedom of expression, open 
mindedness, empathy, and public interest – as the components of deliberation.

Deliberation
The concept of deliberation derives from deliberative democracy. 
Waldman (2001) describes deliberative democracy as "a system in which 
decisions are made not by coercion or bargaining among interests but 
through a discursive process in which citizens collectively consider 
and debate alternatives." Deliberation is a sub-concept of 
deliberative democracy that is often refereed as an act of engaging 
in political discussion and the quality of discussion that 
contributes to deliberative democracy.
An early attempt to define deliberation has been made by Habermas 
(1981, 1984). He acknowledges that in ideal situations people should 
solve problems by inviting and encouraging arguments for all sides in 
pubic sphere. Waldman (2001) describes deliberation as "reasoned 
discussion among equals about public issues with the goal of 
ascertaining the nest course of action so as to optimize the common 
good." Rawls (1971/ 1999) defines deliberation as a process in which 
the reasoned public with equal voice and positions freely 
participates in deliberative discussion for a rationally motivated 
consensus. Gutmann and Thompson (1996) contend that deliberation is 
an ongoing process in which rational opinion is presented equally and 
reciprocally. According to Bohman, deliberation is reaching reasoned 
agreement among free and equal citizen (1996, p. 322).
Deliberation as a dependent variable
The attempt to operationalize and quantify the process of 
deliberation has been made by a handful of scholars. In an 
experimental study (Ryu, 2004), participants' deliberation about a 
school issue was measured as a dependent variable with nine items 
that construct four concepts of deliberation, willingness to express 
opinion, attention to their discussion, understanding of their 
discussion, and flaming. In a survey study (McLeod et al., 1999), 
three questions about whether to attend a local forum where citizens 
discuss local or community problems is called, the likeliness of 
speaking up at the meeting, the likeliness of expressing an opinion 
that is different from those of others at the meeting were asked to 
gauge deliberation.
More thorough measurement of deliberation has been developed in 
content analysis studies. In an analysis of discourse on abortion in 
two German newspapers, Gehards (1997) identified representativeness 
of the actors covered in the newspaper articles, the degree of 
respect expressed toward other participants in the debates, degree of 
justification of claims, and rationality of discourse quality as the 
four core dimensions of deliberation. Recently, Steenbergen and his 
colleagues (2003) developed a discourse quality index (DQI) that 
serves as a quantitative measure of discourse of deliberation. Their 
definition of DQI was based on the Habermas' discourse ethics. The 
discourse quality index measured nine concepts – openness of 
participation, justification of assertions and validity of claims 
exchanged by participants, the concerns for common good, the idea 
that participants should consider common good, mutual respect among 
discussion participants, the constructive politics, the idea that 
participants in a discourse should at least attempt to reach mutually 
acceptable comprises solutions, and authenticity, the absence of 
deception in expressing intentions – as indicators of deliberation.
Defining deliberation
Definitions of deliberation provided by political and communication 
scholars indicate that deliberation is composed of two dimensions, 
the actual act of engaging in political discussions to exchange 
different views and the quality of the political discussion itself. 
In this study both dimensions are considered as the components of 
deliberation.
Based on the literature, six concepts – reciprocity, reasoned 
discourse, freedom of expression, open mindedness, empathy, and 
public interest – are identified as the essential components of 
deliberation.  Reciprocity refers to the actual behavior of engaging 
in political discussion with opponents who have different views 
whereas the rest of concepts represent the discourse quality.  With 
those six concepts, deliberation can be defined as "a process of 
freely exchanging reasoned political views with those who have 
different political views with open mindedness that can result in 
understanding of different political views, which can contribute to 
common good."
             Reciprocity: Deliberation should accompany reciprocal 
exchange of two opposing ideas. It involves two-way communication 
processes where different political views are voiced and heard among 
participants (Park, 2002). Participants should be wiling to express 
their variety of political perspectives and views and further 
persuade opponents who have different political views (Waldman, 
2001).  Therefore, disagreement is an essential part of deliberation. 
Disagreement does not have to result in agreement in deliberation, 
but the process of exchanging different perspectives and arguing for 
their position should accompany the deliberation process.
Reasoned discourse: It is believed that much mass opinion is 
disorganized (Converse, 1964) and without basis in belief or 
knowledge (Herbst, 1993; Zaller, 1992). However, in deliberation, 
reasoned ideas supported by some evidence or justification should be 
exchanged (Waldman, 2001). Arguments that are not supported with 
reasons and justifications are not readily understood by opponents 
who have different political views, and thus no meaningful public 
discourse can be made. When participants are fully aware of 
justification of their political views as well as others, true 
deliberation occurs (Park, 2000).
Cappella et al. (2002) proposes "argument repertoire" as one measure 
of opinion quality.  It provides a way to measure how well reasoned 
each stated opinions are for one's own opinions. He thought that the 
understanding of reasons for their own opinions and reasons why 
others might disagree with my opinion provides one means of measuring 
opinion quality.
Freedom of expression: Deliberation occurs when there are no 
constraints on topics or point of views that can be discussed among 
participants. The public should feel free to introduce new ideas and 
perspectives to the discussion (Steenbergen et al., 2003). No one 
should be prevented from exercising freedom of expression due to 
internal or external coercions (Cohen, 1989).
Open mindedness: Participants in the deliberation discourse must have 
respect for others and their opinions. Steenbergen et al. (2003) 
identifies three dimensions of respect: respect toward the group, 
respect for demands under discussion, and respect toward 
counterarguments, arguments raised by opponents that contradict my 
opinion. Respect for different perspective can elicit more complete 
arguments with underlying reasons and justifications for such a view 
(Stewart & Logan, 1998).
Empathy: A sense of empathy should be shared among participants in 
deliberation. Deliberative participants develop a sense of empathy 
after careful reflection of opponents' point of view (Rogers, 1980). 
Although it is impossible to put oneself in another person's 
position, empathy helps participants to pay more careful attention to 
other's arguments and understand opposing views even though they do 
not necessarily agree with their opponents' opinion.
Public interest: The discussion should involve public issues and the 
consideration for the common good (Steenbergen et al., 2003). If the 
discussion reflects self interest of a small group or the arguments 
presented in the discussion harm the common good, the discussion does 
not become deliberation.

Online deliberation
Deliberation is not limited to face-to-face situations. Deliberation 
can occur on the internet through diverse channels such as online 
forums, chat rooms, usernet groups, and listserv emails. People can 
log onto the internet and join political discussion forums with 
people who can not be encountered otherwise due to physical and 
social constraints.  This study will look at how online deliberation 
differs from the traditional deliberation process that occurs in 
face-to-face situations.
The potential of the internet as the deliberative public sphere has 
been evaluated. Engaging in chat room dialogue exposes participants 
to a variety of opinions, attitudes, and sources of information that 
they may have never encountered otherwise (Weger & Askhus, 2003). The 
internet provides a perfect place to find different views of a very 
diverse group of people who are at the same time open to such 
difference and disagreement needed for deliberation.
The internet also allows citizens to "break our public silence" and 
become involved in public communication" (Fernback, 1997). The 
internet provides a more liberated environment for those who are 
afraid to speak up their voices because anonymity is relatively 
guaranteed on the internet, as long as they do not have to log on 
with their real names. The internet may liberate people from 
psychological constraints and help them express opinions that are 
different and sometimes the opposite of their discussion partners' opinions.
However, the contribution of the internet to deliberation is 
questioned by some. The internet may provide a virtual public sphere 
where diverse groups of people can join and exchange opposing ideas, 
but the question is whether all these different people actually do 
find each other on the internet, or whether they seek the 
like-minded. Sunstein (2001) expresses worries about the polarizing 
effects of the internet might hold. He fears that people will seek 
the like-minded to talk on the internet. This is possible because the 
internet provides very segmented online forums where only those 
people with similar opinions can meet and discuss, unlike the real 
world where people are sometimes forced to mingle with people with 
different opinions.
The online deliberation process is very different from the 
deliberation process that occurs with face-to-face interactions. The 
diversity of the networked, liberated environment for free speech, 
equality in participation provided by the internet may facilitate 
deliberation, while segmented online forums may hinder deliberate 
discussions on the internet.

Factors that influence deliberation
Factors that affect deliberation also have been identified.
Political Interest
Political interest is defined as the "degree to which politics 
arouses a citizen's curiosity" (Van Deth 1990, p. 278), and political 
information seeking and use of newspaper as a consequence of 
generalized interest in political life (McLeod et al, 1996, 1999). 
Political interest has been shown to be linked to news exposure, 
political knowledge, and some dimensions of political participation. 
Schuefele and Shah (2000) argued that political interest spurs 
increased attention to news media and encourage other forms of 
political information consumption. In their survey, respondents with 
a higher level of political interest were more likely to expose 
themselves to political contents of the news media.
Political interest is also known to be correlated with political 
knowledge (Tilley et al.,2004). Many authors have tended to posit 
political interest and media exposure as its primary exogenous causes 
for political knowledge (Bartle, 2000). Tolbert et al. (2003) also 
found that people with higher political interest and efficacy, media 
consumption, and education have greater political knowledge from 
their survey studies.
The relationship between political interest and exposure to diverse 
views has been rarely studied.  Rather, political interest has been 
known to be positively related to political discussions. Schuefele & 
Shah (2000) found that political interest is positively related to 
civic activities and that it fosters engagement in collective 
actions. We are particularly interested in political discussion among 
various forms of civic activities.  It can be assumed that political 
interest has a positive relationship with the frequency of political 
discussion.
In the sense that political interest spurs increased attention to 
news media as well as and encourage participating in political 
discussions, we can assume that political interest may lead to 
exposure to diverse views.
Media Use
Media use, how much people read, watch and listen to news stories 
through mass media, can increase political knowledge. Moy (2004) 
found that attention to local news on TV and news in newspapers 
enhanced perceptions of knowledge. Political knowledge is gained as 
the consequence of the news consumption.
The role of media use has been given little attention in the 
political participation literature (McLeod et al.,1999). Recently, 
however, some researchers found that public affairs content in 
newspaper and television is positively related to traditional forms 
of participation (Chaffee & Kanihan, 1997; McLeod et al., 1996; Moy, 
Scheufele, & Holbert, 1999). Particularly, McLeod et al. (1996) found 
that consumption of both local newspaper and television public 
affairs content to positively influence a citizen's willingness to 
participate in a public forum. Based on these early studies, we 
assume the positive relation between media use and the frequency of 
political discussion.
In addition, Media use is also assumed to influence the exposure to 
diverse views. Mutz and Martin (2001) assert that mere exposure to 
mass media messages increase exposure to diverse views because media 
reports tend to include a diversity of opinion to make stories look 
more balanced. McLeod et al. (1999) also found the positive 
correlation between media use and heterogeneity of network. Mutz and 
Martin (2001) suggest that this role of media is rooted in the 
relative difficulty of selectivity exposing oneself to those sources 
of information. On the other hand, as for internet use other than 
newspaper and television, such researchers as Chaffee (2001) concern 
that the user-controlled nature of the internet might return campaign 
communication to the selective-exposure pattern that obtained in the 
era of highly partisan newspapers.
Exposure to diverse views
One's willingness to expose oneself to diverse point of views by 
engaging in conversation with someone with different political views 
or by consuming mass media whose opinion differ from mine is an 
important gateway to deliberation. Deliberative discussion does not 
happen unless one is exposed to different opinions. By the way, it 
should be noted that this concept of exposure to diverse views is 
distinguished from reciprocity, one of the six concepts which define 
deliberation. Reciprocity refers to actual exchanging of ideas that 
includes disagreement and persuasion among discussants, whereas 
exposure to diverse views refers to paying attention to different 
opinions in mass media and political discussion without expressing 
one's own opinions.
The concept of exposure to diverse views is sometimes referred as 
network heterogeneity (McLeod et al., 1999), which is defined as a 
network of heterogeneous people in their social network in terms of 
their demographics and political opinions. Individuals who compose a 
network whose members come from diverse socio-political backgrounds 
are more likely to encounter a diversity of viewpoints and opinions 
in their political discussion (McLeod, Sotirovic, & Holbert, 1998). 
This exposure to diverse viewpoints leads to deliberation. McKuen 
(1990) addresses that through heterogeneous networks, comprised of 
diverse opinion, public dialogue and deliberative democracy come alive.
In the classical two-step flow study, (Berelson et al., 1954, 
Lazarsfeld et al., 1944) the term  "cross-pressure" is used to 
represent people's exposure to diverse opinions. "Cross-pressures" 
are created by disagreement among participants, and these 
cross-pressures lead to political withdrawal as a consequence of 
psychic stresses produced by political uncertainty and mixed 
political messages.
Mutz (2002) also presents a similar concept called "cross-cutting 
exposure." She found that cross-cutting network increases awareness 
of rationales for one's own viewpoints and also awareness of 
rationales for oppositional viewpoints and these awareness increase 
political tolerance. However, she also found that cross-cutting 
exposure discourages political participation (Mutz, 2002). Exposure 
to those with political views different from one's own creates 
greater ambivalence about political options, and this makes it more 
difficult to take part in political actions.
Two opposing views have been presented concerning the exposure to 
diverse views. Encountering diverse viewpoints may lead to more 
active political participation while this may also lead to confusion 
among the public and thus withdrawal from the politics. Although two 
opposing views are presented, it is difficult to deny that exposure 
to diverse views provide foundation for the deliberative discussions 
because exchanging different ideas does not take place unless 
exposing oneself to other opinions precedes.
One important point to be made at this point is that exposure to 
diverse views does not happen frequently. People try to avoid 
political discussions that are likely to be controversial (Graber, 
1984). Many express that politics and religion are topics they avoid 
because they considered them potentially divisive. At the same time, 
people are talking with same clusters of people and chance of meeting 
someone who has completely different political view is decreasing as 
the social network filters the kind of people we interact with in a 
daily basis (Waldman, 2001; Huckfeldt, 2004).
Exposure to diverse views through mass media also educates and 
prepares the public for deliberative discussion. However, whether 
people actually look for media messages that contradict their 
opinions is questionable. People tend to selectively expose 
themselves to information and media outlets that confirm their 
attitudes and opinions (Jonas, Schulz-Hardt, Frey, & Thelen, 2001). 
Exposure to diverse views provide basis for the deliberation, but it 
doesn't happen frequently.
Frequency of political discussion
The frequency of political discussion differs from deliberation in 
that it just records the frequency of political discussion without 
considering the quality of discourse. Carpini & Keeter (1996) suggest 
that political discussion is related to political knowledge. It is 
also suggested that the impact of hard news use on political 
knowledge is moderated by a person's interpersonal discussion about 
politics (Schuefele, 2002). The frequency of political discussion may 
predict deliberation, because those who frequently engage in 
political discussions are more likely to encounter opponents who have 
different political views and exchange ideas with them. Thos who 
avoid political discussion are not likely to be encounter discussants 
who have different political views, and thus, deliberation can not occur.
Political Knowledge
In this study, political knowledge is conceptualized as factual or 
current-event knowledge, which is knowledge about political issues or 
politicians that are relevant in a contemporary political context. 
Political knowledge is considered as a key form of citizenship and, 
without this, citizens are unable to understand key issues (Perloff, 
1998). With respect to participatory behavior, knowledge fulfills at 
least three purposes (Wolfinger & Rosenstone, 1980): allows citizens 
to make informed decisions, creates a sense of civic duty among 
citizens, and increases familiarity with bureaucratic institutions 
and political process (Scheufele, 2002). This political knowledge 
will influence deliberation because without political knowledge, 
people cannot make valid rational arguments, which is a key concept 
in deliberation.
Demographics
Studies (Huckfeldt & Sprague, 1995; Verba et al., 1993) suggests that 
man more than women, whites more than blacks or Hispanics, the 
wealthy more than the poor, Republicans more than Democrats engage in 
political discussion and exchange their views with others. Also those 
with strong political ideologies are more likely to talk than those 
with weak political ideologies.

Hypotheses
We propose that political interest, political knowledge, news 
exposure, frequency of political discussion, and exposure to diverse 
views influence deliberation directly and indirectly. The summary of 
the hypotheses is presented in a proposed model of the deliberation 
process (Figure 1).
[Figure 1 inserted about here]
H1: Political interest will have direct positive influence on 
political knowledge, frequency of political discussion, and exposure 
to diverse views.

H2: News exposure will mediate the relationship between political 
interest and the three variables, political knowledge, frequency of 
political discussion, and exposure to diverse views.

H3: Political knowledge, frequency of political discussion, and 
exposure to diverse views will have direct positive influence on deliberation.

Finally, a research question about the online deliberation process is asked.

R1: How does the online deliberation process different from the 
face-to-face deliberation?



Methods

In order to test the proposed theoretical model, an online survey was 
conducted on the international journalism and mass communication 
educators and their students. The survey was conducted as a part of 
the World Inventory Media Use Survey Project conducted at a 
northeastern university, where seven research teams 
participated.[1]  The online questionnaire was designed and 
distributed using an online survey website called the Surveymonkey 
(http://www.surveymonkey.com).

Sampling
Snow ball sampling method was used in order to reach broad levels of 
participants in the world. The survey questions were sent to sampled 
members affiliated with communication associations, IAMCR, WAPOR, 
ICA, NCA, and AEJMC, and then the receivers were asked to forward the 
messages to their students and colleagues. For IAMCR and WAPOR, the 
entire list of members was included in the sample. For ICA and NCA, 
every third email address in the alphabetized directory list was 
included in the sample. For AEJMC, every second address in the 
alphabetized directory list with an American suffix (.com, .net, 
.org, .edu), and every address with a non-American suffix, such as 
.il, .mx, .de, were included in the sampling. When email addresses 
were not available, the next available address was used.
In order to increase the sample size for the international 
respondents, websites for journalism and communication schools 
located in countries other than the U.S. were searched and faculty 
and students email addresses found from those websites were included 
in the sample. As the results a total of 4610 survey invitation 
emails were sent, and 807 people completed the survey.

Measurement
The analyses involved control exogenous variables and endogenous variables.

Exogenous variables
Control variables included age, gender, education, income, and race. 
Age was measured by asking how old the participants were on their 
last birthday. Education was accessed by asking the years of school 
attendance. Income was measured by asking people to rate their house 
hold income in comparison to other people in their country on a 
nine-point scale from one (least income) to nine (most income). Fro 
gender, male was coded 1 and female was coded 2. Whites were coded 1 
and other ethnicities were coded 2.
Endogenous variables
News exposure
News exposure was measured separately for television news, newspaper, 
and internet news. On average, how many days a week people spend 
watching television, reading newspapers and browsing interne was 
asked. On those days, how much time people send getting news from 
those sources were measured in hours and minutes with open-ended 
questions. The total days of week multiplied by the minutes of news 
consumption per day yielded the news exposure score for each medium.
Political interest
Three questions were indexed to measure political interest 
(Cronbach's Alpha = .93). How much respondents are interested in 
politics; how worthy of time they think pursuing political issues 
are, and how closely they follow news about political events were 
measured on a five-point Likert-scale.
Political knowledge
Ten true or false questions on factual knowledge of current political 
events and political figures were asked. Each of the ten items was 
scored 1 for correct answers and 0 for incorrect answers. The 
aggregation of the total correct scores became the political 
knowledge score.
Frequency of political discussion
The frequency of political discussion in the face-to-face situations 
and the online context was measured. For the face-to-face situations, 
questions on how frequently people are engaged in political talks 
with their parents, siblings, friends, teachers, spouse, and 
co-workers or colleagues were asked on a five-point Likert-scale 
raging from strongly agree to strongly disagree. The six questions 
were indexed to yield a single score (Cronbach's Alpha = .75). The 
frequency of political discussion on the internet was measured by 
asking how frequently people share their political views with others 
on the internet on a five-point Likert-scale.
Exposure to diverse views
Exposure to diverse political views was measured in two contexts – 
the face-to-face situation and the online context. In the 
face-to-face situations, the degree to which attention is paid to 
diverse political views and diverse media with different political 
views was measured to gauge exposure to diverse views.
For the online exposure to diverse view, the diversity of political 
discussion forums respondents participate and the diversity of media 
with different political views they consume were asked. Those two 
items were indexed (Cronbach's Alpha = .91).
Deliberation
Six concepts – reciprocity, reasoned discourse, freedom of 
expression, open mindedness, empathy, and public interest – that 
compose the deliberation construct were measured with 14 Likert-scale 
items. The same set of questions was asked twice for the face-to-face 
situations and the online context.
Reciprocity was measured by four questions asking how frequently 
respondents talk about politics with people who have different 
political views and those who support different political views, how 
frequently they try to convince others to change their political 
views, and how frequently others try to convince them to change their 
political views. The three items were indexed for the face-to-face 
(Cronbach's Alpha = .77) and the online situations (Cronbach's Alpha = .61).
Reasoned discourse was measured with two open-ended questions. 
Respondents were asked to write as many reasons they can list as to 
why they feel favorable toward the political party they support or, 
if being independent, the reasons why they are independent. The 
second question asked respondents to write as many reasons as they 
can think of for why others feel unfavorable toward the political 
party they support, or why others might criticize about their being 
independent. The number of reasons they provided and the quality of 
their answers were coded based on Cappella et al. (2002). One point 
was given for every reason the respondent wrote. However, when the 
answers were irrelevant or did not make sense they were coded zero. 
The combined score of the answers for the two questions produced the 
reasoned discourse score.
Freedom of expression was assessed by asking about how freely 
respondents can express their opinion even when others disagree with 
them as well as how much opportunity they have to express their 
opinions. The two questions were combined to produce a single score 
for face-to-face (Cronbach's Alpha = .67) and the online models 
(Cronbach's Alpha = .67).
Open-mindedness is measured by asking how often their opinions have 
been trivialized by those who disagree with them, how often they 
respect other people's opinion, and how often they ignore opinions 
that are opposite to theirs. Due to a low reliability score, one 
question on how often they respect other people's opinion was used 
for both face-to-face and online measures.
Empathy was measured by asking how close their discussion partners 
are to the respondents, how much they understand the perspectives of 
those who disagree with them, and how much talking with people with 
different political views helps them to build mutual respect with 
their discussion partners. Due to a low reliability score, the 
question on how close their discussion partners are to the 
respondents was dropped, and two remaining questions were used to 
produce indexes for both the face-to-face (Cronbach's Alpha = .57) 
and the online (Cronbach's Alpha = .63) measures.
Finally, two questions about public interests were asked, but one 
question on how much respondents are willing to withhold their 
opinion if that can result in a better society was used in the final 
measurement due to a low reliability score among the two items.

Analysis
First, missing values were treated with the method of full 
information maximum likelihood estimates provided by the AMOS 
program. Full information maximum likelihood estimates tends to be 
less biased than other methods such as multiple imputation, listwise 
deletion, and pairwise deletion (Arbuckle & Wothke, 1995). Second, 
bivariate correlation matrices were created to examine zero-order 
correlations for the face-to-face and online deliberation models 
separately (Table 3 & 4). Third, structural equation modeling (SEM) 
was carried out using the AMOS statistical program. Because 
chi-square test is often unreliable with larger samples (Bentler, 
1990), two additional model fit indices – the comparative fit index 
(CFI) and the root mean square error of approximation (RMSEA) were 
used to gauge the model fit. CFI values range from 0 to 1, which .95 
indicating a good fit (Hu & Bentler, 1999). For the RMSEA, a value 
below .06 indicates a good fit.  Path significance was evaluated at 
the .05 level.


Results
Means and standard deviations for the variables used in the study are 
presented in Table 1. The average age for the sample is 39.89 years 
old.  The average education level is very high (Mean = 17.02 years, 
SD= 5.59), which indicates that a lot of participants had above 
Bachelor's degree. About the same proportion of male and female 
respondents completed the survey (Table 2) and the majority of people 
who participated in the survey are white (74.2%).
[Insert Table 1 about here]
[Insert Table 2 about here]
Two correlation matrixes were presented for face-to-face deliberation 
model and the online deliberation model (Table 3 and Table 4).  In 
both matrixes, exposure to diverse view is correlated with many other 
variables while political knowledge is not very well correlated with 
other variables.
[Insert Table 3 about here]
[Insert Table 4 about here]
Two structural equation models were tested – one for the face-to-face 
deliberation process (Figure 2, Table 5 and Table 6), and the other 
for the online deliberation (Figure 3, Table 5 and Table 6). For the 
online deliberation, three variables, exposure to diverse view, 
frequency of political discussion, and deliberation construct, were 
replaced with online exposure to diverse views, frequency of online 
political discussion, and online deliberation respectively. The two 
models did not produce satisfying goodness of fit values, so both 
models had to be modified. As the result, the modified model of the 
face-to-face deliberation has the following goodness of fit values: 
chi-square = 188.16, d.f.= 116, p < .001; CFI = .96; and RMSEA = .02. 
The modified online deliberation model has the following goodness of 
fit values: chi-square = 167.43, d.f. = 112, p < .01; CFI = .967; 
RMSEA = .02. Thus the two modified models has a very good model fit. 
The chi-square is significant which doesn't suggest a goodness of 
fit, but this could be related to the big sample size (Bentler, 1990).
[Insert Figure 2 about here]
[Insert Figure 3 about here]
[Insert Table 5 about here]
[Insert Table 6 about here]
The squared multiple correlations indicates that the face-to-face 
model explains 50% of variance in deliberation, 11% in political 
interest, 11% in internet news, 63% in exposure to diverse views, 24% 
in frequency of political discussion, 10% in knowledge, 1% in 
television news, and 0% in newspaper. The square multiple 
correlations show that the online deliberation model explains 30% of 
variance in online deliberation, 11% in political interest, 11% in 
internet news exposure, 1% in television, 23% in exposure to diverse 
views on the internet, 10% in online political discussion, 10% in 
knowledge, and 0% in newspaper news exposure.

Deliberation measurement model
One goal of this study is to identify sub-dimensions that compose the 
deliberation construct and see how those variables contribute. As 
shown in Figure 2, a large variance in the deliberation construct is 
explained by the six variables. For the face-to-face deliberation, 
freedom to express ideas, reciprocity and the common good dimension 
contribute to the deliberation construct better than four other 
dimensions. As shown in Figure 3, freedom of expression and 
reciprocity dimensions contribute to the online deliberation 
construct as well.  The coefficient for reasoned discourse in the 
face-to-face deliberation measurement model is higher than the online 
deliberation model, while coefficients for freedom of expression and 
reciprocity are higher in the online deliberation.

Findings related to hypotheses
Three hypotheses about the face-to-face deliberation process has been 
proposed. H1 predicted that the people with high political interest 
has more factual political knowledge, participate in political 
discussion more frequently and expose themselves to more diverse 
political views.  This hypothesis was supported. In the model, 
political interest has direct paths to political knowledge, frequency 
of political discussion, and exposure to diverse political views, and 
all three paths are statistically significant.
The second hypothesis predicted that the relationships between 
political interest and the three variables, political knowledge, 
frequency of political discussion, and exposure to diverse views are 
mediated by news exposure variables – newspaper, television news, and 
internet news. This hypothesis is partly supported. Internet news is 
found to mediate the relationship between political interest and the 
frequency of political discussion. Those who are interested in 
politics tend to get news from internet, and those who read a lot of 
news from internet tend to involve more frequently in political discussions.
The third hypothesis proposed that political knowledge, frequency of 
political discussion, and exposure to diverse views have positive 
relationship with deliberation. This hypothesis is partially 
supported. Frequency of political discussion, and exposure to diverse 
political views have direct path to deliberation. As implied in the 
correlation matrix, political knowledge which did not show many 
correlations with other variables did not have influence on the 
deliberation process. This indicates that frequency of political 
discussion and exposure to diverse views mediate the relationship 
between political interest and the deliberation.
The research question asks about the difference between face-to-face 
and online deliberation process. As shown in Figure 3, a similar 
model is proposed for the online deliberation process. Political 
interest predicts political knowledge, the frequency of online 
political discussion, and exposure of diverse views on internet. News 
exposure also meditates this process similar to the face-to-face 
deliberation model. Internet news exposure mediates the relationship 
between political interest and the frequency of online political 
discussion, and between political interests and the exposure to 
diverse views on the internet. Those who are interested in politics 
are more likely to read news from the internet, and those who read 
news from the internet are more likely to engaged in online 
discussion about political issues and view more diverse political 
views on the internet. Unlike the face-to-face process, television 
news exposure mediates the relationship between political interest 
and the frequency of online discussion participation. However, a 
negative coefficient is linked between television news exposure and 
the online discussion. Those who are interested in politics are more 
likely to watch television news, but those who watch television news 
are less likely to participate in online political discussions. 
Similar to the face-to-face situation, the frequency of online 
discussion and the exposure to diverse views on the internet have 
direct paths to the online deliberation construct, but knowledge does 
not affect the online deliberation.

Discussions
This study tried to identify factors that influence deliberation by 
proposing a deliberation model, and compare the face-to-face 
deliberation process with the online deliberation model. First, six 
variables that compose the deliberation construct were identified, 
and how those six concepts are related to the deliberation process 
have been examined. Overall the six concepts explained a lot of 
variance in the deliberation construct for both face-to-face and 
online models. For both models, freedom to express ideas, 
reciprocity, and common good factors are identified as strong 
contributors of deliberation. This is somewhat consistent with the 
literature on the deliberation. Many scholars points out that the 
reciprocity, exchanging ideas with discussants who have different 
political views and trying to convince them to change their views or 
vote, is the key factor that composes the deliberation. In order for 
deliberation to occur, people should freely exchange ideas and try to 
argue for their political views with those who do not agree with 
their political views, and those exchanges should reflect the idea of 
common good rather than group selfishness.
Political interest was identified as a strong predictor of 
deliberation. Those who have a lot of interest in politics are more 
likely to be engaged in political discussion and seek out diverse 
political views that do not necessarily agree with their own views, 
and these frequent discussion and exposure to diverse views lead them 
to be engaged in deliberation more often.
The concept of exposure to diverse political view is introduced in 
this study as a mediating variable between political interest and 
deliberation. So far, concepts of heterogeneity of network (McLeod et 
al., 1999), cross-cutting network (2002), and cross-pressure 
(Berelson et al., 1954) were examined in relation with political 
knowledge, political participation, and political discussion. 
However, the concept of one's willingness to seek out diverse 
political views that contradicts their own political views has not 
yet been measured. This concept of exposure to diverse political 
views may be developed into a valuable concept in the deliberation 
process. In this study, those who are interested in politics are more 
likely to seek opinions that contradict their own belief rather than 
to surround themselves with similar ideas. Those who have a lot of 
political interest are likely to talk with people who have different 
political views and read newspapers that have different political ideologies.
Another finding of this study is that the news exposure through media 
does not contribute much to the deliberation process. This 
contradicts previous notion that newspaper reading and television 
news viewing indirectly influence deliberation when mediated by 
reflection (McLeod, Scheufele, & Moy, 1999), and increase political 
participation by increasing knowledge. (Eveland & Scheufele, 2000). 
Mutz (2002) also identified media news exposure as an important 
factor that lead to exposure to diverse political views. One 
interesting finding is that internet news reading is found to promote 
political discussion and exposure to diverse political views, whereas 
newspaper reading is neither influence by political interest nor 
influence the political discussion. In the online deliberation model, 
television news viewing even has a negative influence on the online 
discussion, which means it has negative indirect impact on the online 
deliberation. Internet media was more influential than the traditional media.
The study compared the face-to-face deliberation with the online 
deliberation. The two models were not very different, but overall, 
the face-to-face model explained more variance in the deliberation 
(50%) than the online deliberation model did for the online 
deliberation construct (30%). Also the path coefficients that depart 
from political interest to deliberation is higher in the face-to-face 
deliberation model than the online model. It can be assumed that more 
factors that are not captured in the online model influence the 
online deliberation process. In order to fully account for the online 
deliberation process, more psychological variables, such as 
self-efficacy and political tolerance should be included in future 
studies. In the deliberation measurement model, reasoned discourse 
dimension contributes more to the face-to-face deliberation model 
than the online model, while freedom of expression and reciprocity 
dimension contribute more to the online deliberation than the 
face-to-face deliberation. This implies that people use more reason 
to argue for their political views in the face-to-face situation than 
on the internet, while on the internet, people are more freely 
exchange ideas and more frequently encountering those who disagree 
with their own political views.
   Limitations of the study should be pointed out. The survey was 
sent to international scholars with high education levels, who 
relatively enjoy a high social status in the country they reside, 
thus the education and income were not successful predictors. 
Perhaps, due to the high educated sample, political knowledge was 
found to have no influence on the deliberation process, which 
contradicts previous findings. This lack of political knowledge 
influence on the deliberation process may reflect the sampling 
difference. Unlike other traditional political communication studies 
that were conducted with the general U.S. population sample, this 
study was conducted with international respondents.  Another 
possibility is that the knowledge score had a low variability. One 
more thing to be mentioned is the cross-sectional nature of the data. 
The SEM model was structured based on the cross-sectional data, 
therefore the paths do not necessarily indicate causal relationships.
This study provided a way to measure deliberation with 
questionnaires. Future studies should improve the deliberation 
measurement based on this study's finding. For future studies, 
deliberation process for specific issues should be studied in order 
to take accounts the nature of discussion topics in the deliberation 
process. Some may participate in deliberation for one topic while 
they may not for another depends on the issues' salience and 
relevance to them. When issues are relevant, people may exchange 
their ideas with those who disagree with them. How issue's features 
influence the deliberation process should be examined in future studies.










References

Arbuckle, J. L., & Wothke, W. (1995). Amos 4.0 user's guide. Chicago, 
Il: Smallwaters corporation.

Bartle, J. (2000). Political awareness, opinion constraint and the 
stability of ideological positions. Political Studies, 48, 467-484.

Berelson, B., Larzarsfeld, P., & McPhee, W. (1954). Voting. Chicago, 
IL: University of Chicago Press.

Bentler, P. M. (1990). Comparative fit indexes in structural models. 
Psychological Bulletin, 107, 238-246.

Bohman, J. (1996). Public deliberation: Pluralism, complexity, and 
democracy. Cambridge, MA: MIT Press.

Cappella, J. N., Price, V., & Nir, L. (2002). Argument repertoire as 
a reliable and valid measure of opinion quality: Electronic dialogue 
during campaign 2000. Political Communication, 19(1), 73-93.

Carpini, D., Michael, X., & Keeter, S. (1996). What Americans know 
about politics and why it matters. New Haven, CT: Yale University Press.

Chaffee, S. H. (2001). Attention to counter-attitudinal messages in a 
state election campaign. Political Communication, vol. 18, no. 3, 247-272.

Cohen, J. (1989). Deliberation and democratic legitimacy. In A. 
Hamlin and P. Pettit (Eds.), The good polity: Normative analysis of 
the State (pp. 17-34). Oxford: Blackwell.

Converse, P. (1964). The nature of belief systems in mass publics. In 
D. Apter (Eds.), Ideology and discontent, New York: Free Press.

Fishkin, J. (1991). Democracy and deliberation. New Haven, CT: Yale 
University Press.

Fishkin, J. S. (1995). The voice of the people: Public opinion and 
democracy. New Haven: Yale University Press.

Fernback, J., Jones, S. G. (Ed). (1997). The individual within the 
collective: virtual ideology and the realization of collective 
principles. In Virtual culture: identity and communication in cyber 
society (pp. 36-54). Thousand Oaks, CA: Sage Publications.

Gastil, J. (1993). Democracy in small group: Participation, decision 
making, and communication. Philadelphia: New Society.

Graber, D. (1984). Mass media and American politics. Washington, 
D.C.: CQ Press.

Gutmann, A. & Thompson, D. (1996). Democracy and disagreement. 
Cambridge, MA: The Belknap Press of Harvard University Press.

Habermas, J. (1984). The theory of communicative action, Vol. 1: 
reason and the rationalization of society. Boston: Beacon Press.

Habermas, J. (1989). The structural transformation of the public 
sphere. Cambridge, MA: MIT Press.

Habermas, J. (1995). Reconciliation through the public use of reason: 
Remark on John Rawls's political liberalism. Journal of Philosophy, 
92, 109-131.

Herbst, S. (1993). Numbered voices. Chicago, IL: University of Chicago Press.

Hu, L-Z, & Bentler, P. M. (1999). Cutoff criteria for fit indexes in 
covariance structure analysis: Conventional criteria versus new 
alternative. Structural Equation Modeling, 6, 1-55.

Huckfeldt, R., & Sprague, J. (1995). Citizens, politics, and social 
communication: information and influence in an election campaign. 
Cambridge, England: Cambridge University Press.

Huckfeldt, R., Mendez, J. M., & Osborn, T. (2004). Disagreement, 
ambivalence, and engagement: The political consequences of 
heterogeneous networks. Political psychology, 25(1), 65-95.

Jonas, E., Schulz-Hardt, S., Frey, D., & Thelen, N. (2001). 
Confirmation bias in sequential information search after preliminary 
decisions: An expansion of dissonance theoretical research on 
"selective exposure to information". Journal of Personality and 
Social Psychology, 80, 557-571.

Jones, D. A. (2002). The polarization effect of new media message. 
International Journal of Public Opinion Research, 14(2), 158-174.

Kim, J., Wyatt, R. O., & Katz, E. (1999), News, talks, opinion, 
participation: The part played by conversation in deliberative 
democracy. Political Communication, 16, 1-33.

Lazarsfeld, P. Berelson, B., & Gaudet, H. (1944). The people's 
choice. New York: Columbia University Press.

McKuen, M. (1990). Speaking of politics: Individual conversational 
choice, public opinion, and the prospects for deliberative democracy. 
In J. Ferejhon & J. Kuklinski (Eds.), Information and democratic 
process, (pp. 59-99). Urbana: University of Illinois Press.

McLeod, J. M., Daily, K., Guo, Z., Eveland, W. P., Bayer, J., Yang, 
S., & Wang, H. (1996). Community integration, local media use and 
democratic processes. Communication Research, 23, 179-209.

McLeod, J. M., Sotirovic, M., & Holbert, R. L. (1998). Values as 
sociotropic judgments influencing communication patterns. 
Communication Research, 26(6), 623-654.

McLeod, J. M., Scheufele, D. A., Moy, P., Horowitz, E. M., Holbert, 
R. L., Zhang, W., Zubric, S.. & Zubric, J, (1999). Understanding 
deliberation: The effects of discussion networks on participation in 
a public forum. Communication Research, 26, 743-774.

McLeod, J. M., Scheufele, D. A., Moy, P., Horowitz E. M., et al. 
(1999). Understanding deliberation: The effects of discussion 
networks on participation in a public forum. Communication Research, 
vol. 26, no. 6, 743.

Mendelberg, T. & Oleske, J. (2000). Race and public deliberation. 
Political Communication, 17, 169-191.

Moy, P., Scheufele, D. A., & Holbert, R. L. (1999). Television use 
and social capital: Testing Putnam's time displacement hypothesis. 
Mass Communication & Society, 2(1/2), 27-45.

Moy, P., McCluskey, M. R., McCoy, K. (2004). Political correlates of 
local news media use. Journal of Communication, vol. 54, no. 3, 532-546.

Mutz, D. C., & Martin, P. S. (2001). Facilitating communication among lines of
difference: the role of the mass media. American Political Science 
Review, 95(1), 97-114.

Mutz, D. C. (2002). The consequences of cross-cutting networks for 
political participation. American Journal of Political Science, 46, 838-855.

Park, S. G. (2000). The significance of civility in deliberative 
democracy. Paper presented at the conference of the Public Opinion 
Research in the Digital Age (PORDA) project, Seoul, Korea.

Perloff, R. M. (1998). Political communication: Politics, press, and 
the public in America. Mahwah, NJ: Lawrence Erlbaum Associates.

Putnam, R. D. (2000). Bowling alone. New York: Simon & Schuster.

Rawl, J. (1971/1999). Chapter VII. Goodness as rationality. In A 
theory of justice. (pp. 347-396). Cambridge, MA: Havard University Press.

Rogers, C. (1980). A way of being. Boston, MA: Houghton Mifflin.

Ryu, S. (2004) When online meet offline: A new approach to the 
operating mechanism of the ideal public sphere for deliberative 
democracy. Paper presented at the annual meeting of Association for 
Education in Journalism and Mass Communication, Toronto, Canada.

Scheufele, D. A. & Shah, D. V. (2000). Personality strength and 
social capital: the role of dispositional and informational variables 
in the production of civic participation.  Communication Research, 
vol. 27, no. 2, 107-131.

Scheufele, D. A. (2002). Examining differential gains from mass media 
and their implications for participatory behavior. Communication 
Research, vol. 29, no. 1, 46-65.

Sears, D. O., & Freedman, J. L. (1967). Selective exposure to 
information: a critical review. Public Opinion Quarterly, 31, 194-213.

Steenbergen, M. R., Bachtiger, A., Sporndli, M., & Steiner, J. 
(2003). Measuring political deliberation: A discourse quality index.

Stewart, J. & Logan, C. (1998). Together: Communicating 
interpersonally. Boston, MA McGraw-Hill.

Sunstein, C. R. (2001). Republic.com. New Jersey: Princeton University Press.

Tilley, J., Sturgis, P., & Allum, N. (2004, May 7-8). Political 
information and motivation: A case of reciprocal causality? Paper 
presented at symposium on 'perceptions preferences and 
rationalization: overcoming the problem of causal inference in the 
study of political behaviour', Nuffield College, Oxford.

Van Deth, J. W. (1990). Interest in politics. In M.K. Jennings & J. 
W. Van Deth (Eds.), Continuities in political action. A longitudinal 
study of political orientations in three western democracies (pp. 
275-312). Berlin: Walter de Gruyter.

Verba, S., Schlozman, K. L., Brady, H., & Nie, N. H. (1993). Race, 
ethnicity and political resources: participation in the United 
States. British Journal of Political Science, 23, 453-497.

Waldman, P. (2001). Deliberation in practice: Connecting theory to 
the lives of citizens. In R. P. Hart and B. H. Sparrow (Eds.), 
Politics, discourse, and American society: New agendas (pp. 151-171). 
Lanham, MD: Rowman & Littlefield.

Weger, H. & Askhus, M. (2003). Arguing in Internet chat rooms: 
Argumentative adaptations to chat room design and some consequences 
for public deliberation at a distance.

Wolfinger, R. E., & Rosenstone, S. J. (1980). Who votes?. New Haven: 
Yale University Press.

Zaller, J. (1992). The nature and origins of mass opinion. New York: 
Cambridge University Press.







Table 1	 Means and standard deviations
Variables
Mean
SD
N
Age (in years)
39.89
15.26
458
Education (in years)
17.02
5.59
441
Income*
4.36
1.62
455
Newspaper**
384.28
618.41
632
Television news**
401.15
667.24
640
Internet news**
455.24
792.77
645
Political interest ***
4.16
.89
564
Political knowledge ****
6.77
1.80
478
Frequency of political discussion (face-to-face)***
3.55
.74
431
Frequency of political discussion (online)***
2.08
1.11
560
Exposure to diverse views***
4.09
.92
565
Exposure to diverse views (online)***
3.07
1.13
558
Deliberation (Face-to-face)
Reciprocity***
3.46
.76
498
Reasoned discourse*****
7.03
3.90
394
Freedom of expression***
4.03
.79
505
Open mindedness ***
4.19
.71
501
Empathy***
3.92
.68
496
Public Interest***
3.94
.84
497
Deliberation (Online)
Reciprocity***
2.85
.97
482
Reasoned discourse*****
7.03
3.90
394
Freedom of expression***
3.29
1.29
492
Open mindedness***
3.87
.91
490
Empathy***
3.49
.85
478
Public Interest***
3.59
1.03
478
* Responses were coded 1= least income to 9= most income.
** Days a week X minutes a day spent with the media
*** Ranges from 1 to 5. 5 is the highest.
**** Number of correct answers, ranging from 0 to 10.
***** Number of reasoned provided.







Table 2	Percentages of gender and race
Variables
   %
Gender
Male
49.7
Female
50.3
100.00%
(N = 461)
Race
African American
1.3
American Indian or Alaska Native
.2
Arabic
1.3
Black, but not African American
1.1
Central Asian
.4
East Asian
7.1
Hispanic
1.3
South Asian
4.8
White
74.2
Others
8.2
100.00%
(N = 462)
















 
Deliberation Process
Table 3 	Pearson correlations for variables in the face-to-face 
deliberation model
Variables
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Age
- .30**
(455)
.44**
(436)
-.35**
(449)
-.10*
(456)
.32**
(453)
.05
(404)
.29**
(454)
.11*
(439)
-.14**
(442)
.06
(414)
.05
(410)
.07
(438)
.09
(436)
-.03
(437)
-.03
(434)
.01
(436)
.15**
(361)
Gender*
-
-.07
(438)
.14**
(451)
.02
(459)
-.20**
(455_
.03
(405)
-.23*
(456)
-.21
(442)
-.05
(445)
.07
(418)
-.01
(410)
-.07
(440)
-.06
(437)
-.03
(438)
-.09
(435)
.02
(437)
-.03
(361)
Education
-
-.15**
(436)
.09
(438)
.11*
(436)
.04
(389)
.09
(437)
.12**
(423)
-.08
(426)
.03
(397)
-.02
(392)
-.05
(424)
.10*
(420)
-.08
(422)
-.05
(420)
-.02
(422)
.09
(352)
Income
-
.02
(453)
-.22**
(451)
-.07
(402)
-.16**
(452)
-.14**
(437)
.002
(440)
.02
(412)
-.07
(404)
-.05
(438)
-.11*
(434)
.04
(435)
.05
(433)
.04
(435)
-.001
(361)
Race**
-
-.10*
(457)
-.08
(407)
-.03
(458)
-.03
(458)
.13**
(447)
.04
(418)
.10*
(411)
-.10*
(442)
-.10*
(439)
-.03
(440)
.004
(437)
-.03
(439)
-.16**
(363)
Political interest
-
.49**
(428)
.79**
(561)
.24**
(474)
.11**
(544)
.10
(515)
.05
(504)
.32**
(501)
.35**
(494)
.10*
(497)
.17**
(492)
.22**
(493)
.27**
(391)
Discussion (offline)
-
.46**
(430)
.13**
(417)
.002
(415)
.05
(394)
.04
(383)
.30**
(419)
.40**
(419)
.10*
(419)
.14*
(416)
.32**
(417)
.26**
(341)
Exposure to diverse views (offline)
-
.23**
(475)
.16**
(545)
.12**
(515)
.12*
(505)
.39**
(504)
.36**
(497)
.19**
(500)
.28**
(495)
.30**
(496)
.24**
(394)
Political knowledge
-
.04
(460)
-.01
(435)
.07
(428)
.14**
(457)
.21**
(435)
.09
(455)
.04
(453)
.10*
(454)
.11*
(375)
Internet news
-
.15**
(583)
.15**
(572)
.07
(490)
-.03
(484)
.06
(486)
.03
(481)
.11*
(482)
.01
(382)
Television news
-
.10*
(576)
.06
(460)
.01
(453)
.04
(456)
.10*
(452)
.10*
(453)
-.03
(359)
Newspaper
-
.06
(447)
.06
(441)
-.02
(444)
.03
(439)
.03
(440)
-.05
(352)
Freedom of expression
-
.25**
(494)
.27**
(497)
.26**
(492)
.28**
(493)
.13*
(389)
Reciprocity
-
.16**
(494)
.27**
(489)
.34**
(490)
.20**
(386)
Open mindedness
-
.47**
(493)
.41**
(494)
-.02
(389)
Empathy
-
.56**
(493)
.04
(387)
Public interest
-
.13**
(389)
Reasoned discourse
-
* Male was coded 1 and female was coded 2.
** Whites were coded 1 and others were coded 2.



Table 4 	Pearson correlations for variables in the online deliberation model
Variables
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Age
-.30**
(454)
.44**
(436)
-.35**
(449)
-.09*
(456)
.32**
(453)
.00
(451)
.10*
(452)
.12*
(439)
-.14**
(442)
.06
(414)
.05
(410)
-.08
(427)
.03
(420)
-.04
(428)
-.03
(418)
-.02
(420)
.15**
(361)
Gender*
-
-.07
(438)
.14**
(451)
.02
(459)
-.20**
(455)
-.08
(453)
-.14**
(454)
-.21**
(442)
-.05
(445)
.07
(418)
-.01
(410)
-.10*
(429)
-.10
(421)
-.04
(429)
-.04
(419)
.02
(421)
-.03
(361)
Education
-
-.15**
(436)
.09
(438)
.12*
(436)
.03
(434)
.06
(435)
.16**
(423)
-.07
(426)
.03
(397)
-.02
(392)
.001
(415)
.10
(409)
.02
(414)
-.01
(408)
-.002
(409)
.09
(352)
Income
-
.02
(453)
-.22**
(451)
-.05
(449)
-.06
(450)
-.13**
(437)
.002
(440)
.02
(412)
-.07
(404)
-.001
(427)
-.05
(419)
-.04
(426)
-.01
(417)
-.01
(419)
-.01
(361)
Race**
-
-.10*
(457)
.11*
(455)
.07
(456)
-.03
(444)
.13**
(447)
.04
(418)
.10*
(411)
.04
(431)
.03
(423)
.06
(431)
.02
(421)
-.05
(423)
-.16*
(363)
Political interest
-
.21**
(557)
.40**
(555)
.24**
(474)
.11*
(544)
.10*
(515)
.05
(504)
.09*
(488)
.13*
(478)
.04
(486)
.10*
(474)
.17**
(474)
.27**
(319)
Discussion (offline)
-
.41**
(554)
.09*
(470)
.17**
(542)
-.01
(509)
.07
(500)
.44**
(489)
.46**
(479)
.20**
(487)
.20**
(475)
.29**
(475)
.15**
(390)
Exposure to diverse views (offline)
-
.13**
(471)
.23**
(539)
-.02
(508)
.13**
(498)
.20**
(487)
.29**
(477)
.24**
(485 )
.26**
(474)
.26**
(474)
.15**
(390)
Political knowledge
-
.04
(460)
-.09
(435)
.07
(428)
.12*
(448)
.17**
(438)
.08
(446)
.06
(439)
.06
(440)
.11*
(375)
Internet news
-
.15**
(583)
.15**
(572)
.09
(478)
.12**
(468)
.07
(475)
.08
(463)
.07
(463)
.01
(382)
Television news
-
.10*
(576)
-.05
(448)
-.06
(437)
-.08
(445)
-.02
(434)
-.11*
(437)
-.03
(359)
Newspaper
-
.04
(436)
.07
(426)
.05
(434)
.07
(422)
.05
(423)
-.05
(32)
Freedom of expression
-
.66**
(479)
.42**
(485)
.38**
(476)
.41**
(476)
.11*
(382)
Reciprocity
-
.34**
(480)
.37**
(471)
.39**
(471)
.22**
(376)
Open mindedness
-
.55**
(476)
.47**
(477)
-.01
(383)
Empathy
-
.69**
(474)
-.02
(375)
Public interest
-
.05
(377)
Reasoned discourse
-
* Male was coded 1 and female was coded 2.
** Whites were coded 1 and others were coded 2.
 
Deliberation Process


Table 5	 Direct and indirect effects of exogenous variables
Source of Influence
Income
Age
Gender
(Male=1)
Education
Race
(white=1)
Political interest
Offline
Direct
-.13
.21
-.13
.00
-
Indirect
.00
.00
.00
.00
-
Online
Direct
-.14
.19
-.13
.00
.00
Indirect
.00
.00
.00
.00
.00
Frequency of discussion (offline)
Offline
Direct
.00
.13
.00
.00
-
Indirect
-.07
.11
-.06
.00
-
Frequency of discussion (online)
Online
Direct
.00
.00
.00
.00
.11
Indirect
-.03
-.01
-.06
.00
.00
Exposure to diverse views (offline)
Offline
Direct
.00
.00
-.07
.00
-
Indirect
-.10
.01
-.11
.00
-
Exposure to diverse views (online)
Online
Direct
.00
.00
.00
.00
.09
Indirect
-.06
.01
-.105
.00
.00
Political knowledge
Offline
Direct
.00
.00
-.16
.12
-
Indirect
-.03
.04
-.02
.00
-
Online
Direct
.00
.00
-.16
.12
.00
Indirect
-.03
.04
-.03
.00
.00
Deliberation (offline)
Offline
Direct
.00
.00
.00
.00
-
Indirect
-.07
.05
-.11
.00
-
Deliberation
(online)
Online
Direct
.00
.00
.00
.00
.00
Indirect
-.02
-.003
-.04
.00
.07
Television news
Offline
Direct
.00
.00
.00
.00
-
Indirect
-.01
.02
-.01
.00
-
Online
Direct
.00
.00
.00
.00
.00
Indirect
-.02
-.003
-.04
.00
.07
Internet news
Offline
Direct
.00
-.31
-.17
.00
-
Indirect
-.02
.04
-.02
.00
-
Online
Direct
.00
-.31
-.17
.00
.00
Indirect
-.02
.03
-.02
.00
.00











Table 6	 Direct and indirect effects of endogenous variables
Source of Influence
Interest
Discussion (offline)
Discussion (online)
Exposure to diverse views (offline)
Exposure to diverse views (online)
Internet news
TV news
Political knowledge
Offline
Direct
.20
.00
-
.00
-
.00
-
Indirect
.00
. 00
-
.00
-
.00
-
Online
Direct
.20
-
.00
-
.00
.00
.00
Indirect
.00
-
.00
-
.00
.00
.00
Frequency of discussion (offline)
Offline
Direct
.52
.00
-
.00
-
.00
-
Indirect
.00
.00
-
.00
-
.00
-
Frequency of discussion (online)
Online
Direct
.20
-
.00
-
.00
.17
.00
Indirect
.03
-
.00
-
.00
.00
.00
Exposure to diverse views (offline)
Offline
Direct
.76
.00
-
.00
-
.09
-
Indirect
.02
.00
-
.00
-
.00
-
Exposure to diverse views (online)
Online
Direct
.40
-
.00
-
.00
.23
-.08
Indirect
.03
-
.00
-
.00
.00
.00
Internet news
Offline
Direct
.18
.00
-
.00
-
.00
-
Indirect
.00
.00
-
.00
-
.00
-
Online
Direct
.17
-
.00
-
.00
.00
.00
Indirect
.00
-
.00
-
.00
.00
.00
Television news
Offline
Direct
.20
.00
-
.00
-
.00
-
Indirect
.00
.00
-
.00
-
.00
-
Online
Direct
.10
-
.00
-
.00
.00
.00
Indirect
.00
-
.00
-
.00
.00
.00
Deliberation (offline)
Offline
Direct
.00
.41
-
.45
-
.00
-
Indirect
.56
.00
-
.00
-
.04
-
Deliberation (online)
Online
Direct
.00
-
.49
-
.12
.00
.00
Indirect
.17
-
.00
-
.00
.11
-.01









Figure 2.  Modified Deliberation Model



[1]  For more information, contact XXX at XXX University.