Asian Journal of Communication

(1)

On: 28 February 2014, At: 21:29 Publisher: Routledge

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Asian Journal of Communication

Publication details, including instructions for authors and subscription information:

http://www.tandfonline.com/loi/rajc20

The impact of political party support on interpersonal political discussion:

survey evidence from two Chinese societies

Francis L.F. Lee^a & Frank C.S. Liu^b

a School of Journalism and Communication, Chinese University of Hong Kong, Hong Kong, China

b Institute of Political Science, National Sun Yat-Sen University, Kaohsiung, Taiwan

Published online: 06 Aug 2013.

To cite this article: Francis L.F. Lee & Frank C.S. Liu (2014) The impact of political party support on interpersonal political discussion: survey evidence from two Chinese societies, Asian Journal of Communication, 24:2, 109-127, DOI: 10.1080/01292986.2013.827731

To link to this article: http://dx.doi.org/10.1080/01292986.2013.827731

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the

“Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

(2)

and-conditions

Downloaded by [National Sun Yat-Sen University] at 21:29 28 February 2014

(3)

ORIGINAL ARTICLE

The impact of political party support on interpersonal political discussion: survey evidence from two Chinese societies

Francis L.F. Lee^a* and Frank C.S. Liu^b

aSchool of Journalism and Communication, Chinese University of Hong Kong, Hong Kong, China;^bInstitute of Political Science, National Sun Yat-Sen University, Kaohsiung, Taiwan

(Received 11 January 2013; final version received 18 July 2013)

This study examines whether and how political party support shapes interpersonal political discussion. Drawing upon existing research, party support is hypothesized to lead to more frequent political discussion and lower levels of disagreement within discussion networks. Party support is also hypothesized to moderate the relationship between news consumption and discussion frequency and the relationship between discussion frequency and disagreement. The analysis further explores if the impact of party support varies according to the parties being supported. The hypotheses and research question were examined using data from representative surveys conducted in Hong Kong and Taiwan. The findings illustrate the importance of political party support in structuring citizens’

interpersonal political discussions in the consolidated democracy of Taiwan and, though to a lesser extent, in the semi-democratic environment of Hong Kong. In Taiwan, the impact of discussion frequencies on disagreement in discussion network varies according to the party being supported. Implications of the findings are discussed.

Keywords: political party support; political discussion; disagreement; news consumption; Hong Kong; Taiwan

Introduction

What motivates people to talk about politics in their everyday life? What factors affect the extent to which people are willing to talk to disagreeing others? These questions are highly important given ample research evidence about the influence of interpersonal political discussions and disagreement on people’s opinions, participation behavior, and adoption of various democratic values and beliefs (for critical reviews, see Carpini, Cook, & Jacobs, 2004; Eveland, Morey, & Hutchens, 2011;

Mutz,2008). However, most of the studies on interpersonal political discussion in the past two decades have treated discussion and disagreement as independent variables. As Eveland et al. (2011) pointed out, questions of how frequently and in what contexts people discuss politics ‘‘have not garnered nearly the scholarly attention that other areas have’’ (p. 1084).

At the same time, political communication researchers have exhibited renewed interests in the notion of partisan selective exposure in recent years (e.g., Garrett, 2009; Iyengar & Hahn,2009; Mutz & Young,2011; Stroud,2010). The proliferation

*Corresponding author. Email:francis_lee@cuhk.edu.hk

Vol. 24, No. 2, 109127, http://dx.doi.org/10.1080/01292986.2013.827731

#2013 AMIC/SCI-NTU

(4)

of channels in the new media environment has led many news outlets to ‘‘go niche’’

by adopting a clear partisan or ideological stance (Jamieson & Cappella, 2008).

Within media systems with high levels of political parallelism (Hallin & Mancini, 2004), partisanship is an important factor structuring people’s consumption of and responses to the media (Holbert & Benoit,2009; Stroud,2010).

This study brings together the core concerns of these two streams of research and examines how political party support structures interpersonal political discussions.

While some past studies have illustrated the correlation between party support and discussion frequencies, few attempted to interrogate into the relationship more thematically and systematically. This study should therefore contribute to research on interpersonal political discussion as well as research on the impact of partisanship on political communication.

Moreover, this study examines two Chinese societies Taiwan and Hong Kong simultaneously. Taiwan has arguably been successfully developed into a consolidated democracy in the past two decades, whereas the democratization process in Hong Kong is slow and fundamentally constrained. The party systems in the two places also differ substantially due to the different political systems. The analysis of the two cases can generate insights into the generalizability of the examined relationships and how contextual variations may affect the ways partisanship structures interpersonal talk.

Partisanship and interpersonal political discussions

Political parties constitute one of the most important institutions in modern democracies mediating the relationships between the polity and the society. Besides elite recruitment and campaign organization, one of the main functions of political parties is ‘‘issue structuring’’ (Gunther & Diamond, 2001, p. 8). That is, political parties organize public choices such that a wide range of complicated policy issues are articulated into a limited number of more or less coherent ‘‘ideologies’’

(Sniderman, 2000). This work done by political parties simplifies the choices that citizens have to make and hence allows citizens to use party affiliation as a heuristic when forming political judgments (Popkin, 1991). In fact, party identification has long been regarded as a fundamental factor determining people’s vote choice and policy stances (e.g., Campbell, Converse, Miller & Stoke, 1960; Keith, Magleby, Nelson, Orr, & Westlye,1992).

Notably, political scientists have observed a decline in party support in many democratic countries around the world in the past decades (Diamond & Gunther, 2001, pp. xxiv). The increasing personalization and mediatization of politics (Wattenberg, 1991), the change toward a post-materialist culture in developed countries (Inglehart,1990), and widespread disaffection toward the operation of the political system in general have led to lower levels of trust and confidence in parties.

Schmitter (2001) argued that political parties are not as strong as they once were when it comes to structuring electoral competition, providing symbolic identity, and aggregating interests and passions.

However, these arguments about the decline of partisanship are arguably applicable more at the collective level than at the individual level. That is, the number of people who behave as strong and loyal partisans might have declined over time in many countries, and this has weakened the capability of political parties to

(5)

mediate between the society and the polity. But at the individual level, partisans and non-partisans, or supporters of different parties, still differ from each other substantially in their opinions and behavior. In a recent review of the American situation, for example, Hetherington (2012) stated that ‘‘[knowing] one’s party identification provides more information than any other single source about a person’s political behaviors, positions on issues, or feelings about groups’’ (p. 115).

Partisanship thus remains an important variable in public opinion research.

The broader concern underlying this study is the capability of partisanship to structure citizens’ communication behavior. As mentioned earlier, few studies have thematically examined how partisanship generates and structures interpersonal political discussions among citizens. However, the centrality of partisanship to individual opinion formation means that the variable does exist in many empirical studies on interpersonal political discussion, if only as a control variable. Hence, it is possible to derive from existing research a set of arguments and hypotheses about how partisanship would relate to political discussions. Specifically, this study focuses mainly on party support (i.e., whether a person supports a political party or not), rather than which party a person supports, as the main independent variable.

Besides, it treats discussion frequency and amount of disagreement within discussion network as the two major dependent variables. The following establishes a number of hypotheses regarding how these three variables should relate with each other.

Party support and frequencies of political discussion

Past research has pinpointed three main sets of factors as the predictors of interpersonal political discussion frequencies. First, citizens who are more interested and involved in politics are more likely to talk (Beck,1991; Bennett, Flickinger &

Rhine, 2000; Schmitt-Beck, 1994). Second, citizens are more likely to engage in political talk when they are stimulated by political communication media, ranging from mundane news consumption (e.g., De Boer & Velthuijsen, 2001; Mondak, 1995) to exposure to electoral campaign communications (Cho,2011; Cho & Choy, 2011; Landreville, Holbert & LaMarre,2010; Pan, Shen, Paek & Sun,2006). Third, people cannot talk about politics unless there are discussants to talk to. Hence, people are more likely to engage in political discussions when they are more socially connected (McLeod, Scheufele & Moy,1999; Schmitt-Beck,1994).

Against such background, the first and most basic hypothesis for this study is that party support should relate positively to discussion frequency. Strong partisans should be more involved in and concerned about politics. Hence, they should engage in discussions more frequently. Indeed, several past studies have documented this relationship between party support and interpersonal discussions (e.g., Beck,1991;

Bennett et al.,2000).

In addition, party support can also moderate the relationship between news consumption and discussion frequencies. Specifically, party support can strengthen the impact of news consumption on discussion. News consumption generates discussion partly because people simply would not discuss politics if they know nothing about it. Nevertheless, sheer news exposure may not result in discussions unless people are motivated to talk. While people may be motivated by different considerations, such as a desire to validate the personal implications of public affairs reported in the news (De Boer & Velthuijsen,2001; Wyatt, Katz & Kim,2000), the

(6)

partisans’ higher levels of political involvement should also motivate them to talk about the news information they obtained. Besides, discussion may also lead to news exposure because people who discuss politics frequently may feel the need to consume the news so that they could better articulate and defend their views (Eveland, 2004). Compared to non-partisans, partisans should feel the need to articulate and defend their own views more keenly. Therefore, party support should strengthen the impact of discussion on news consumption. In other words, no matter how one understands the causal direction involved in the positive relationship between news consumption and interpersonal discussion, party support can be expected to strengthen the relationship.

Based on the above discussions, two hypotheses are set up:

H1: Party support is positively related to interpersonal discussion frequency.

H2: Party support strengthens the positive relationship between news consumption and discussion frequency.

Party support and amount of disagreement in discussion network

Besides discussion frequencies, this study is also concerned with the extent to which people experience political disagreement. As Ikeda, Liu, Aida, and Wilson (2005) pointed out: The dynamical systems theory of groups posits that people who strongly identify with a group would seek out and maintain ties with others belonging to the same group. Hence, strong partisans should tend to seek out like-minded people to talk to. More generally speaking, one basic reason why political disagreement exists in interpersonal networks is that political opinions do not always constitute an important criterion when people select whom to befriend, where to live, and where to work (Huckfeldt, Johnson & Sprague,2004). Yet, for strong partisans who are deeply involved in politics, political views may indeed be a relatively more important factor shaping their interpersonal relationships. Therefore, strong partisans are more likely than non-partisans to find themselves in politically homogeneous networks. Indeed, some past studies have demonstrated the negative relationship between party support and disagreement within discussion networks (e.g., Feldman & Price, 2008;

Brundidge,2010; Huckfeldt, Ikeda & Pappi,2005; Schmitt-Beck,1994).

Beyond the direct relationship, party support may also moderate the relationship between discussion frequency and interpersonal disagreement. Many scholars have acknowledged that disagreeing with one’s significant others constitutes an unpleasant experience that many people tend to avoid (Mutz, 2006; Schudson, 1997).

Besides, people can mutually persuade and influence each other through discussion.

Hence, interpersonal disagreement can be reduced over time if people talk to each other frequently. These reasons account for why the relationship between discussion frequency and disagreement is typically negative (Huckfeldt & Mendez,2008; Kenny, 1991; Schmitt-Beck,1994).

However, party support is likely to strengthen the negative relationship between discussion frequency and disagreement. This hypothesis can be derived from a cognitive dissonance framework. That is, disagreeing with a close friend or family member can generate cognitive dissonance and, hence, psychological discomfort. In one sense, this is just reiterating the point that discussing with disagreeing others can be unpleasant. Yet, the degree of psychological discomfort caused by the dissonance

(7)

can vary. Interpersonal disagreement on a matter would not generate much discomfort for a person who regards the matter as unimportant (Festinger, 1962).

Therefore, people may not have a strong urge to avoid political disagreement if politics does not matter too much to them (Huckfeldt et al., 2005). As partisans are more likely than non-partisans to see politics as important, they should be more motivated to avoid discussing with disagreeing others. Hence, the negative relationship between discussion frequency and disagreement should be stronger among partisans.

Based on the above discussion, the hypotheses concerning the relationship between party support and disagreement are:

H3: Party support relates negatively to political disagreement in interpersonal network.

H4: Party support strengthens the negative relationship between discussion frequency and political disagreement within discussion network.

Does the identity of the party matter?

H1 to H4 are concerned with the implications of whether a person supports any political party or not. Theoretically, the arguments underlying the hypotheses should be generally applicable no matter which political party a person is supporting.

Nevertheless, political parties do differ from each other in various ways, such as size, ideology, and status as the incumbent party or not. While political party support may generally generate more discussion and undermine disagreement within network, it remains possible that the strength of the impact would vary depending on which party one is really supporting. This article, therefore, also examines the following research question:

Q1: Does the impact of party support on political discussion frequencies and disagreement within discussion network vary depending on the identity of the party being supported?

Contexts: Taiwan vs. Hong Kong

The four main hypotheses stated in the previous section are based mainly on research in Western democracies and premised on the power of partisanship in shaping citizens’ political behavior. However, the centrality of partisanship to public opinion can vary across contexts. As this study examines the cases of Taiwan and Hong Kong, a brief discussion of the political and party systems of the two places is needed.

Before democratization, Taiwan was governed by the Nationalist Party (KMT) for decades, and the formation of new political parties outside the KMT was banned.

But social and economic developments of Taiwan have led to rising popular demands, which led the KMT regime to introduce democratic elections into the national representative bodies beginning in the 1970s (Chu,2001). A new group of political leaders outside of KMT emerged, articulating a platform centered on democratization and a separate Taiwan identity. This new political elite founded the Democratic Progress Party (DPP) in December 1986, even though the ban on political party was still in place.

The party ban was lifted soon afterward, and democratization quickened since the late 1980s. A two-party system (KMT vs. DPP) emerged, though numerous minority parties also exist. The first democratic presidential election was held in 1996

(8)

and won by the KMT. Yet, DPP succeeded in winning the elections in both 2000 and 2004, ending decades of KMT rule. However, KMT returned to power in 2008, and the incumbent Ma Ying-jeou has won the reelection in January 2012. The fact that political power has shifted largely peacefully from KMT to DPP and back again constitutes a sign of democratic consolidation.

Some scholars have argued that Taiwan citizens do not have very widespread and strong party loyalty (Mattlin, 2004). Nonetheless, while the proportion of party identifiers may not be very high, partisanship does have a substantial influence on people’s evaluations of government and electoral choices (Huang, 2010; Wu,2009;

Wu & Huang,2007). It is worth noting that some recent studies have addressed the phenomenon of partisan selective exposure in mass communication in Taiwan (Liu, 2009,2010), but relatively little has been written on the relationship between party support and interpersonal political discussion.

In contrast, the process of democratization in Hong Kong, currently a Special Administrative Region under China, has been slow. The introduction of democratic elections into the representative bodies in the city began in the 1980s. The first local political party, the United Democrats of Hong Kong, was established in 1991. Since then, numerous political parties were established. Nowadays, the city has a multiparty system, even though the major parties are often grouped together by the media into the ‘‘pro-establishment’’ camp and the ‘‘pan-democrats.’’ The development into a multi-party system was partly the result of the adoption of proportional representation in the Legislative Council elections since 1998.

Despite 20 years of development, political parties in Hong Kong have yet to gain strong public support. Public opinion research in the city has shown that local citizens do not place high levels of trust on political parties in general. A survey in 2010, for instance, asked citizens to rate the extent to which various entities could represent public opinion in Hong Kong. Political parties obtained an average rating of only 4.42 on a 0-to-10 scale. It fared worst in a list containing various media platforms, legislators, the government, and social movements (Lee & Chan,2012).

Both cultural and institutional reasons contribute to the lack of trust and confidence in political parties in Hong Kong. Culturally, the post-war development of Hong Kong has led to the formation of a culture of depoliticization. People tended to see politics as dirty and party politics as power struggles among self- interested factions (Lam, 2004; Lau & Kuan, 1988; Lee & Chan, 2011).

Institutionally, the current election laws in Hong Kong stipulated that the Chief Executive of the government cannot belong to any political party. In other words, political parties have no chance to govern; they can only become either the government’s helpers or oppositional parties within the legislature. As a result, as Ma (2007, p. 124) stated, ‘‘political parties in Hong Kong are plagued by meager resources, a low level of participation from the public, and a paucity of policy influence.’’

Against such background, one might expect partisanship to have bigger influence on people’s political opinions and communication behavior in Taiwan than in Hong Kong. This is not set up as a hypothesis though, since this study is not a formal comparative analysis (due to differences in the operationalization of a couple of variables). But we will return to the contextual features outlined above when interpreting the findings.

(9)

Method and Data Survey method

The Taiwan survey examined below was conducted in November, 2011, by the Survey Center of National Sun Yet-San University, Kaohsiung, Taiwan. The population was all eligible voters in Taiwan (above 20 years old). Sampling was based on the telephone book published by Chung-Hua Telecom in 2010. The last two digits of all numbers were removed and replaced by a full set of 100 double-digit figures from 00 to 99. Specific numbers were then randomly selected from the database by computers. The survey has 1074 completed interviews, with a response rate of 17.8% following American Association of Public Opinion Research (AAPOR) formula 1.¹ The major sample-population discrepancies reside in the underrepresentation of young adults and overrepresentation of highly educated citizens. With the adjustment of raking, proper weights were applied to the sample to ensure that the distributions of age, gender, educational level, and geographical location do not substantially differ from the population.

The Hong Kong survey was conducted in October and November 2011 by the Center of Communication Research at the Chinese University of Hong Kong. The population was all Chinese-speaking city residents between 18 and 70. Sampling was based on telephone numbers from the residential directories in 2005, 2007, and 2009.

The last two digits of all numbers were replaced by a full set of 100 double-digit figures. Specific numbers were then randomly selected from the database by computers. The most recent birthday method was used to select the target respondent from a household. The survey has 845 completed interviews and a response rate of 42.0% following AAPOR RR3. The sample does not differ substantially from the population in gender ratio and age distribution. But educated people are over- represented. The sample was weighted by education when conducting the analysis.

Operationalization

Party support was examined in the two surveys in the same way through two questions. The first asked the respondent if there is a political party in Hong Kong / Taiwan that s/he supported. If the respondent said no, a second question asked if there is a party that the respondent ‘‘leaned toward.’’ Based on these two items, a three-point scaled index of party support was created with party supporters 2, party-inclined 1, and non-partisan 0. In Hong Kong, 32.9% of the respondents acknowledged that they supported a political party, while 20.2% were party-inclined and 46.9% were non-partisans. Slightly fewer respondents in Taiwan acknowledged that they were partisans, as the corresponding percentages in Taiwan were 27.9%, 20.2%, and 51.9%. When mean scores are concerned, respondents in Hong Kong did score significantly higher on the variable than the Taiwan respondents (M 0.88 and 0.76, S.D. 0.90 and 0.86, t 2.97, p B.01 in an independent-samples t-test).

Political discussion frequency was measured by a question in each survey. The question was phrased somewhat differently in Hong Kong and Taiwan though. The Hong Kong survey asked: ‘‘Usually, how frequently do you discuss government policies with friends or family members?’’ The answers were recorded with a five- point scale ranging from 1 not at all to 5 very frequently (M 2.52, S.D. 0.92).

The Taiwan survey asked: ‘‘How frequently did you discuss government policies with

(10)

your friends and family members last week?’’ The answers were also recorded with a five-point scale ranging from 1 not at all to 5 almost every day (M 2.05, S.D. 1.27).

Interpersonal disagreement was measured in exactly the same way in the two surveys. The respondents were asked if they regarded their opinions as consistent with their friends’ and family members’ when discussing: (1) politics and politicians, and (2) government policies. The answers were recorded with a five-point Likert scale ranging from 1 very consistent to 5 very inconsistent. The two items are highly correlated with each other in both surveys (r .59 and .44 in Taiwan and Hong Kong respectively). They were averaged to form an index of amount of disagreement within discussion network. The mean scores in the two samples do not differ from each other significantly (M 2.77, S.D. 0.99 in Taiwan, and M 2.69, S.D. 0.69 in Hong Kong, t 1.85, p .05 in an independent-samples t-test).

News consumption was measured in terms of level of attention that people paid to the news media. In media effects research, the general notion of news consumption has been operationalized in terms of exposure, attention, reception, or reliance (e.g., Eveland, Hutchens & Shen, 2009; McLeod & McDonald, 1985; Price & Zaller, 1993). This study opts for attention because concepts such as reception and reliance imply more than sheer consumption of news. At the same time, when compared to news exposure, news attention was shown to be more capable of registering news effects (Chaffee & Schleuder, 1986; McLeod & McDonald, 1985). Specifically, the variable was measured in both surveys by asking the respondents to indicate, by means of a five-point Likert scale (1 very inattentive to 5 very attentive), how much attention they paid to political news when: (1) reading newspapers and (2) watching TV news. The two items are highly correlated in both surveys (r .65 and .74 for Taiwan and Hong Kong respectively) and were averaged to form the index.

Throughout the analysis, we label this variable news attention, but it conceptually represents the more general notion of news consumption in this study. Hong Kong people reported slightly but significantly higher levels of news attention when compared to the Taiwanese (M 3.37, S.D. 1.22 in Taiwan, and M 3.53, S.D. 0.80 in Hong Kong, t 3.54, p B.001 in an independent samples t-test).

Control variables included sex, age, education, and four attitudinal factors in each survey. Three of the attitudinal factors are the same in the two surveys: external efficacy, conflict avoidance, and perceived polarization of the society. External efficacy was measured through respondents’ agreement, expressed through a five- point Likert scale, with the statement ‘‘the government is responsive to citizens’

demands.’’ Conflict avoidance was similarly measured with the statement ‘‘debating with others makes me feel uncomfortable.’’ Perceived polarization was the average of agreement with three Likert-scaled statements (details omitted due to space concern).

The seventh control variable is political interests in Hong Kong, measured by agreement with the statement ‘‘I am interested in politics,’’ and policy concern in Taiwan, measured by agreement with the statement ‘‘I am concerned about the government’s public policies.’’

The control variables are included for different reasons. Political interests, policy concern, and external efficacy are basic political orientations that are frequently included as controls in much political communication research. Conflict avoidance is controlled because of its potential relevance to discussion and disagreement (Mutz, 2006). Perceived polarization is controlled because group polarization has been an

(11)

important issue in both Hong Kong and Taiwan politics, and the issue may impinge on how people relate to political parties.²

Analysis and Results

Predicting discussion frequency

Multiple regression analysis was conducted to tackle the first two hypotheses regarding how political party support and news attention relate to discussion frequency. The independent variables include all the controls, news attention, party support, and the interaction between news attention and party support (for testing H2). The interaction term was centered to alleviate multicollinearity. The regression was conducted among the Hong Kong and Taiwan samples separately

The first column ofTable 1shows the results for the Taiwan sample. Among the control variables, concern with policies has a positive relationship with discussion frequency, and people who were more conflict-avoidant tended to discuss politics less frequently. News attention is, unsurprisingly, strongly related to discussion frequency.

More importantly, party support has a moderately strong relationship with discussion frequency. Partisans in Taiwan indeed tended to discuss politics with their family members and friends more frequently. H1 is therefore supported in Taiwan. At the same time, H2 is also supported in Taiwan, as the two-way interaction between news attention and party support obtains a statistically significant and positive regression coefficient. The positive coefficient means that the generally positive relationship between news attention and discussion frequency would be even more strongly positive when party support is strong.

The meaning of the interaction effect can be more clearly discerned graphically as inFigure 1. The figure was produced based on the regression results, with sex set at 1 (i.e., male) and all other variables set at their respective means. The figure shows that discussion frequency goes up as one moves from low levels of news attention (set at 1, the lowest value in the scale) to high levels of news attention (set at 5, the highest value in the scale), no matter whether one is a non-partisan, a party-inclined, or a

Table 1. Party support and interpersonal political discussion frequency.

Taiwan Hong Kong

Sex 0.05 0.05

Age 0.01 0.06

Education 0.00 0.11**

Political interests / Policy concern 0.11*** 0.10**

External efficacy 0.04 0.06*

Conflict avoidance 0.06* 0.01

Perceived polarization of society 0.01 0.07*

News attention 0.29*** 0.45***

Party support 0.21*** 0.07*

Party support X news attention 0.09** 0.00

Adjusted R2 0.205*** 0.295***

N 1074 845

Notes. Entries are standardized regression coefficients. For sex, men 1 and women 2. Missing values were replaced by mean scores. p B0.10.*** p B0.001; ** p B0.01; * p B0.05.

(12)

party supporter. Yet, the effect of news attention is, as predicted, stronger for party supporters. Party support, in other words, serves as a contributory condition (Eveland,1997) for the positive relationship between news attention and discussion frequency.

The second column ofTable 1shows that more educated people and people more interested in politics in Hong Kong discussed politics more frequently. There is also a positive relationship between perceived polarization of the society and discussion frequency, while people who regarded the government as irresponsive to public opinion were also more likely to engage in interpersonal political discussion. News attention is very strongly and positively related to discussion. Although formal statistical comparisons between the Hong Kong and Taiwan findings cannot be done due to differences in how the discussion frequency variable was operationalized, the coefficient obtained by news attention in Hong Kong does seem to be substantially larger than the coefficient for news attention in Taiwan.

In contrast, party support is not a strong predictor of discussion frequency in Hong Kong. The party support variable does have main effect on discussion frequency; thus, H1 is also supported. But the size of the relationship is small.

Besides, party support does not moderate the relationship between news attention and discussion frequency. Hence, H2 is not supported in the Hong Kong data.

Predicting disagreement within discussion network

We can now turn to examine the hypotheses regarding disagreement within discussion networks. Multiple regression was again conducted, and the independent variables included the demographics and attitudinal controls, news attention, discussion frequency, party support, and an interaction term between party support and discussion frequency. The interaction term was again centered.Table 2summarizes the findings.

Figure 1. Interaction effect between party support and news attention on discussion frequencies (results from the Taiwan survey).

(13)

In both Hong Kong and Taiwan, better educated people tended to be located within more politically homogeneous networks. Besides, females and people who paid higher levels of attention to the news in Hong Kong also tended to find themselves in more homogeneous discussion networks. More importantly, discussion frequency relates negatively and significantly to disagreement in both societies,³i.e., people who reported higher levels of perceived disagreement in social networks tended to discuss politics less frequently. Partisanship does not moderate the relationship between disagreement and discussion though. In both Hong Kong and Taiwan, the two-way interaction term obtains a negative coefficient in the analysis.

The negative sign of the coefficient is consistent with the prediction of H4 it means that the negative relationship between disagreement and discussion would be even more strongly negative when party support becomes stronger. But the coefficients fail to achieve statistical significance. Hence, H4 is not supported.

H3 is supported in both Taiwan and Hong Kong. Party support is significantly and negatively related to disagreement within discussion network. But similar to Table 1, the negative coefficient obtained by the party support variable is substantially larger in Taiwan. That is, partisanship seems to have stronger effects in structuring interpersonal political discussion in Taiwan than in Hong Kong.

Effects of party support for supporters of different parties

Q1 asks whether the effects of party support would vary across supporters of different parties. We examine the cases of Taiwan and Hong Kong in turn. For Taiwan, to simplify the analysis, we grouped together the supporters of the KMT and the People First Party into the supporters of the ‘‘blue camp.’’ We also grouped supporters of the DPP and the Taiwan Solidarity Union into supporters of the

‘‘green camp.’’ Only three respondents named a party beyond the four parties, and 25 party supporters or party-inclined did not name a party that they supported.

Although these respondents were party supporters, we nonetheless grouped them Table 2. Party support and interpersonal political disagreement.

Taiwan Hong Kong

Sex 0.04 0.11**

Age 0.01 0.01

Education 0.13** 0.08*

Political interests / Policy concern 0.01 0.00

External efficacy 0.01 0.02

Conflict avoidance 0.05 0.01

Perceived polarization of society 0.06 0.01

News attention 0.07 0.11*

Discussion frequency 0.14** 0.16***

Party support 0.18*** 0.08*

Party support X discussion frequency 0.05 0.04

Adjusted R2 0.098*** 0.083***

N 739 766

Notes. Entries are standardized regression coefficients. For sex, men 1 and women 2. Missing values were deleted pairwise. p B0.10.*** p B0.001; ** p B0.01; * p B0.05.

(14)

together with the non-partisans in the current analysis. This procedure essentially divided the respondents into three groups the blue supporters, the green supporters, and ‘‘others.’’ The third category ‘‘others’’ is a residual, reference category including all respondents who were not supporters of the two major camps.⁴ With this regrouping, tackling Q1 simply means replacing the party support variable by the blue supporter and green supporter variables in the regression analysis. This implies that we ignore the difference between the party-inclined and the party supporters in order to focus on whether the blue supporters and the green supporters differ significantly from the non-partisans (treated as the reference group). We also created relevant interaction terms (e.g., news attention X blue supporter) to examine the moderating impact of supporting specific parties.

The top half ofTable 3summarizes the main results from the Taiwan survey. The main effects of party support on both dependent variables remain intact. Yet the coefficients obtained by the green supporter variable are slightly larger in both columns. Moreover, in both columns, only the green supporter variable exhibits significant moderating influence. WhileTable 1shows that party support strengthens the relationship between news attention and discussion frequency,Table 3shows that this effect is significant only among green supporters. More interestingly, whileTable 2shows that party support does not significantly strengthen the negative relationship between discussion frequency and disagreement, Table 3 shows that this hypothesized moderating influence does exist among green supporters. Although the differences between the corresponding coefficients are not always large, the consistency of the pattern that the coefficient of the green supporter variable is

Table 3. Effects of identity of parties being supported.

Discussion frequency Disagreement Taiwan

News attention 0.30*** 0.07

Discussion 0.16***

Blue supporter 0.12*** 0.08*

Green supporter 0.14*** 0.11**

News attention X blue supporter 0.05

News attention X green supporter 0.09**

Discussion X blue supporter 0.03

Discussion X green supporter 0.09*

Adjusted R2 0.191*** 0.090***

Hong Kong

News attention 0.45*** 0.11***

Discussion 0.15***

Pan-democrat supporter 0.06 0.01

Establishment supporter 0.06 0.11**

News attention X pan-democrat supporter 0.02

News attention X establishment supporter 0.03

Discussion X pan-democrat supporter 0.06

Discussion X Establishment supporter 0.01

Adjusted R2 0.295*** 0.076***

Notes. Entries are standardized regression coefficients. The other control variables were included in the regression but omitted from the table. *** p B0.001; ** p B0.01; * p B0.05.

(15)

larger than the corresponding coefficient of the blue supporter variable in all cases is highly suggestive. The impact of party support on interpersonal political discussion is seemingly more substantial among green supporters in Taiwan.

The Hong Kong data were analyzed following the same procedure. The respondents were first regrouped into: (1) supporters of the pan-democrats, (2) supporters of the pro-establishment parties, and (3) the remaining others. The pan- democrat supporter and establishment supporter variables, together with the relevant interaction terms, were then used in the regression in place of party support and the original interaction terms. The bottom half of Table 3 shows that, when the two groups of supporters were separated, neither group was significantly more likely to discuss politics when compared to the non-partisans. Besides, the pan-democrat supporters were not more likely than others to encounter less disagreement in their discussion networks. Nevertheless, supporters of the pro-establishment parties were less likely to encounter disagreement in their discussion networks. The findings show that the main effects of party support on discussion frequencies do not vary according to the parties that one supports, but the effect of party support on disagreement does apply only to supporters of the pro-establishment camp.

There is no significant interaction effect between news attention and both party support variables on discussion frequencies. At the same time, there is also no significant interaction effect between discussion frequencies and both party support variables on disagreement.

Discussion

This article begins with the concern of how partisanship may structure citizens’

political communication behavior not only in terms of media consumption, as analysis of partisan selective exposure has illustrated (Stroud, 2010), but also in terms of interpersonal talk. A set of hypotheses on the implications of partisanship on political discussion and disagreement was developed and tested in two East Asian societies. In Taiwan, three of the four hypotheses regarding the impact of party support were supported, while the fourth hypothesis is also partly supported after further analysis. Certainly, party support is not the only or most important factor behind discussion behavior. For example, consistent with the extant literature (Beck, 1991; Landreville et al., 2010; Schmitt-Beck, 1994), discussion frequency was strongly driven by news consumption (i.e., news attention in the analysis) and interests in politics. Yet, party support does relate to discussion frequency and disagreement even after controlling for news consumption and several basic political attitudinal variables.

Moreover, the analysis adds the new finding that party support can strengthen the linkage between news consumption and political discussion. That is, while one type of political communication tends to breed the other, the mutual reinforcement between media consumption and interpersonal talk is strengthened when people have

‘‘their own team’’ in the political arena. Further, although initial analysis finds no support for the argument that party support would strengthen the negative relationship between discussion and disagreement, additional analysis shows that the moderating influence is indeed significant among supporters of the green camp.

H4, therefore, is also partly supported in the end.

(16)

Being a newly consolidated democracy, the case of Taiwan illustrates the relevance of partisanship in shaping how citizens communicate in a democratic system. It should be noted that the Taiwan findings do not necessarily contradict the claims made by scholars regarding the decline of the centrality of political parties in democracies (Gunther & Diamond,2001; Schmitter,2001). Scholars concerning with the decline of parties typically focused on the phenomenon at the aggregate level.

The weakening role of political parties is often indicated by collective-level phenomena such as low levels of trust in political parties, declining proportions of citizens being party supporters, or higher levels of electoral volatility. In fact, the Taiwan survey finding may be taken as confirming these scholars’ accounts, since as many as 51.9% of the Taiwan respondents reported no party affiliation or inclination. What the analysis confirms, however, is the continual relevance of party support when analyzing citizens’ opinions and behavior at the individual level.

The last sentence of the previous paragraph is less applicable to the case of Hong Kong. Only two of the four main hypotheses received clear support from the data, and the effects in those two cases seem to be substantially weaker than in the case of Taiwan. Party support does not moderate the relationship between news consumption and discussion frequencies, and party support strengthens the negative relationship between discussion and disagreement only among supporters of the pro-establishment parties. The Hong Kong findings, however, can be understood as an illustration of how the characteristics of the political and party systems in place would shape the relevance of partisanship as a factor in explaining public opinion and citizens’ political behavior. As noted earlier, Hong Kong is not a fully democratic society. Local political parties have a relatively short history. They cannot win the right to govern because the Chief Executive of the Hong Kong government is not directly elected. While they could win seats in the legislature, their power to monitor and influence the executive branch of the government is limited by the design of the current political system. These structural constraints meant that political parties in Hong Kong are weak intermediaries between the society and the polity (Ma,2007).

Past research has shown that local citizens do not evaluate political parties highly (Lee & Chan,2012). It is therefore not surprising that partisanship would also play a minor role in shaping public opinion.

Of course, regarding the contrast between Taiwan and Hong Kong, one limitation of the present study is that the two surveys are not directly comparable to each other, mainly due to the slightly different operationalization of the core discussion frequency variable. Nevertheless, the hypothesized impact of party support is consistently and substantially stronger in the case of Taiwan. At the same time, there are variables that have seemingly stronger impact on the dependent variables in Hong Kong than in Taiwan, e.g., the impact of education and news consumption on discussion frequency.

The differential power of party support in the two surveys, though cannot be examined precisely in statistical terms, is nonetheless unlikely to be artifacts of the different operationalization of the discussion frequency variable. In any case, the findings point to the need of more formal comparative analysis of the impact of partisanship on citizens’ communication behavior in different societies.

Our analysis also explores the issue of whether the impact of party support varies depending on the parties that one supports. The analysis helps us to discern some evidences supporting H4. In Taiwan, the hypothesized moderating influence of party support on the relationship between discussion frequency and disagreement does

(17)

exist, though only among supporters of the ‘‘Green camp.’’ In the Hong Kong case, however, the hypothesized relationship does not exist for supporters of both major camps. It is probably beyond the scope of this article to speculate on why supporters of the Green camp in Taiwan would ‘‘stand out’’ in exhibiting a negative relationship between discussion frequencies and disagreement. But generally speaking, the finding suggests that, as different political parties or factions in the society may have different histories and ideologies, interpersonal political discussions of supporters of the different political parties may be structured somewhat differently.

This could be a topic for future research to explore.

In any case, one should not exaggerate the differences between the effect of party support for supporters of different parties. In both Hong Kong and Taiwan, the direct and moderating effects of party support on the dependent variables among supporters of the two main camps are actually comparable in most cases. In other words, there are certain basic effects of being a supporter of a political party on interpersonal political discussion that are largely independent of the size, status, and ideology of the party being supported.

On the whole, this study contributes to our understanding of how interpersonal political discussions among ordinary citizens can be structured along partisan lines.

One underlying presumption of the study and the foregoing discussion is the importance of considering how political institutions, especially those institutions that are supposed to serve as intermediaries between the society and the polity, may structure citizen communications. Future research can extend the analysis to consider how people’s relationships with other intermediaries, such as interest groups, civic associations, and social movements (Schmitter,2001), may shape their discussion behavior. Of course, one can also extend the analysis of the impact of partisanship by taking additional factors into account.

Such research is highly important given the role that interpersonal discussions and disagreement have in shaping individuals’ political opinions and behavior. The findings generated can have normative implications too. Based on the findings in this study, for example, one may argue that partisanship’s implications on public deliberation are paradoxical. Deliberation requires not only discussions, but also discussions across lines of political differences (Mutz,2006,2008). But as both the Taiwan and Hong Kong findings show, being a partisan leads one to engage in more interpersonal talk and to the development of a more homogeneous discussion environment. How to resolve this tension between ‘‘getting involved in the process’’

and ‘‘engaging with others’’ is a question for both theorists and empirical researchers to ponder on.

Notes

1. RR1 refers to the minimum response rate and is the strictest formula. Hence, the figure tends to be small. The low response rate is also a result of proliferation of survey research in contemporary societies and is not atypical in contemporary Taiwan.

2. The descriptive statistics of the control variables are not shown due to space concern.

3. There were some respondents in both societies who did not give a valid response to the questions on political disagreement. Some existing studies would recode the ‘‘don’t know’’

and other invalid responses into zero (e.g., Lee, 2009; Scheufele, Nisbet, Brossard, &

Nisbet,2004). Those studies adopted the recoding procedure due to their concern with the impact of actual exposure to disagreement on other political attitudinal variables. In those

(18)

cases, the recoding procedure is reasonable because the invalid answers were given mainly by people who did not discuss politics at all, and people who did not discuss politics were by definition not exposed to disagreement. The present study, however, is not concerned with impact of exposure to disagreement, but to the extent to which the discussion networks are actually homogeneous. Not being exposed to disagreement does not mean that the network is homogeneous. Hence, no recoding was done in this study, and the actual sample sizes for the analysis reported in Table 2 are somewhat smaller.

4. An alternative approach is to treat the 28 respondents who claimed to be partisans and yet did not indicate themselves as a distinctive category ‘‘other party supporters.’’ But we did not create this category, because the category itself is not substantively meaningful within the context of the current analysis. In any case, the different approaches to handle the 28 respondents would not alter the findings substantively.

Notes on contributors

Francis L. F. Lee (PhD, Stanford University, 2003) is Associate Professor and Head of Graduate Division at the School of Journalism and Communication, Chinese University of Hong Kong. His research interests reside in public opinion studies, journalism studies, political communication, and media and social movements.

Frank C. S. Liu (PhD, University of Kansas, 2006) is Associate Professor at the Institute of Political Science at the National Sun Yat-Sen University, Kaohsiung, Taiwan. His research interests mainly reside in political communication, political parties, and political identities.

References

Beck, P. A. (1991). Voters’ intermediation environments in the 1988 presidential contest.

Public Opinion Quarterly, 55(3), 371394. doi:10.1086/269269

Bennett, S. E., Flickinger, R. S., & Rhine, S. L. (2000). Political talk over here, over there, over time. British Journal of Political Science, 30(1), 99119. doi:10.1017/S0007123400000053 Brundidge, J. (2010). Encountering ‘‘difference’’ in the contemporary public sphere: The

contribution of the internet to the heterogeneity of political discussion networks. Journal of Communication, 60(4), 680700. doi:10.1111/j.1460-2466.2010.01509.x

Campbell, A., Converse, P. E., Miller, W., & Stoke, D. E. (1960). The American voter.

New York, NY: Wiley.

Carpini, M. X. D., Cook, F. L., & Jacobs, L. R. (2004). Public deliberation, discursive participation, and citizen engagement: A review of the empirical literature. Annual Review of Political Science, 7(1), 315344. doi:10.1146/annurev.polisci.7.121003.091630

Chaffee, S. H., & Schleuder, J. (1986). Measurement and effects of attention to media news.

Human Communication Research, 13(1), 76107. doi:10.1111/j.1468-2958.1986.tb00096.x Cho, J. (2011). The geography of political communication: Effects of regional variations in

campaign advertising on citizen communication. Human Communication Research, 37(3), 434462. doi:10.1111/j.1468-2958.2011.01406.x

Cho, J., & Choy, S. P. (2011). From podium to living room: Elite debates as an emotional catalyst for citizen communicative engagements. Communication Research, 38(6), 778804.

doi:10.1177/0093650210378518

Chu, Y. H. (2001). The legacy of one-party hegemony in Taiwan. In L. Diamond &

R. Gunther (Eds.), Political parties and democracy (pp. 266298). New York, NY: Johns Hopkins University Press.

De Boer, C., & Velthuijsen, A. S. (2001). Participation in conversations about the news.

International Journal of Public Opinion Research, 13(2), 140158. doi:10.1093/ijpor/13.2.140 Diamond, L., & Gunther, R. (2001). Introduction. In L. Diamond & R. Gunther (Eds.), Political parties and democracy (pp. ixxxxiv). New York, NY: Johns Hopkins University Press.

(19)

Eveland, W. P. (1997). Interactions and nonlinearity in mass communication: Connecting theory and methodology. Journalism & Mass Communication Quarterly, 74(2), 400416.

doi:10.1177/107769909707400211

Eveland, W. P. (2004). The effect of political discussion in producing informed citizens: The roles of information, motivation, and elaboration. Political Communication, 21(2), 177193.

doi:10.1080/10584600490443877

Eveland, W. P., Jr., Hutchens, M. J., & Shen, F. (2009). Exposure, attention, or ‘‘use’’ of news? Assessing aspects of the reliability and validity of a central concept in political communication research. Communication Methods and Measures, 3(4), 223244.

doi:10.1080/19312450903378925

Eveland, W. P., Jr., Morey, A. C., & Hutchens, M. J. (2011). Beyond deliberation: New directions for the study of informal political conversation from a communication perspective.

Journal of Communication, 61(6), 10821103. doi:10.1111/j.1460-2466.2011.01598.x Feldman, L., & Price, V. (2008). Confusion or enlightenment? How exposure to disagreement

moderates the effects of political discussion and media use on candidate knowledge.

Communication Research, 35(1), 6187. doi:10.1177/0093650207309362

Festinger, L. (1962). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.

Garrett, R. K. (2009). Politically motivated reinforcement seeking: Reframing the selective exposure debate. Journal of Communication, 59(4), 676699. doi:10.1111/j.1460-2466.2009.

01452.x

Gunther, R., & Diamond, L. (2001). Types and functions of parties. In L. Diamond & R.

Gunther (Eds.), Political parties and democracy (pp. 338). New York, NY: Johns Hopkins University Press.

Hallin, D., & Mancini, P. (2004). Comparing media systems. New York, NY: Cambridge University Press.

Hetherington, M. (2012). Partisanship and polarization. In A. Berinsky (Ed.), New directions in public opinion (pp. 101118). New York, NY: Routledge.

Holbert, R. L., & Benoit, W. L. (2009). A theory of political campaign media connectedness.

Communication Monographs, 76(3), 303332. doi:10.1080/03637750903074693

Huang, H. H. (2010). Split-voting in Taiwan’s concurrent election and referendum: An exploratory test of social context. Issues & Studies, 46(2), 87126.

Huckfeldt, R., Ikeda, K., & Pappi, F. U. (2005). Patterns of disagreement in democratic politics: Comparing Germany, Japan, and the United States. American Journal of Political Science, 49(3), 497514. doi:10.1111/j.1540-5907.2005.00138.x

Huckfeldt, R., Johnson, P. E., & Sprague, J. (2004). Political disagreement. New York, NY:

Cambridge University Press.

Huckfeldt, R., & Mendez, J. M. (2008). Moths, flames, and political engagement: Managing disagreement within communication networks. Journal of Politics, 70(1), 8396. doi:

10.1017/S0022381607080073

Ikeda, K., Liu, J. H., Aida, M., & Wilson, M. (2005). Dynamics of interpersonal political environment and party identification: Longitudinal studies of voting in Japan and New Zealand. Political Psychology, 26(4), 517542. doi:10.1111/j.1467-9221.2005.00429.x Inglehart, R. (1990). Culture shift in advanced industrial society. Princeton, NJ: Princeton

University Press.

Iyengar, S., Hahn, K. S. (2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59(1), 1939. doi:10.1111/j.1460-2466.2008.01402.x Jamieson, K. H., & Cappella, J. N. (2008). Echo chamber. New York, NY: Oxford University

Press.

Keith, B. E., Magleby, D. B., Nelson, C. J., Orr, E., & Westlye, M. C. (1992). The myth of the independent voter. Berkeley, CA: University of California Press.

Kenny, C. B. (1991). Partisanship and political discussion. Political Geography Quarterly, 10(2), 97109. doi:10.1016/0260-9827(91)90014-L

Lam, W. M. (2004). Understanding the political culture of Hong Kong. New York, NY: M.E.

Sharpe.

Landreville, K. D., Holbert, R. L., & LaMarre, H. L. (2010). The influence of late-night TV comedy viewing on political talk: A moderated-mediation model. The International Journal of Press/Politics, 15(4), 482498. doi:10.1177/1940161210371506

(20)

Lau, S. K., & Kuan, H. C. (1988). The ethos of the Hong Kong Chinese. Hong Kong: Chinese University of Hong Kong Press.

Lee, F. L. F. (2009). The impact of political discussion in a democratizing society: The moderating role of disagreement and support for democracy. Communication Research, 36(3), 379399. doi:10.1177/0093650209333027

Lee, F. L. F., & Chan, J. M. (2011). Media, social mobilization and mass protests in post-colonial Hong Kong. London: Routledge.

Lee, F. L. F., & Chan, J. M. (2012). Exploring the movement society in Hong Kong: The formation and development of collective contentious actions in Hong Kong. In S. K.

Cheung, K. C. Leung, & K. M. Chan (Eds.), Hong Kong, discourse, media (pp. 243270).

Hong Kong: Oxford University Press (in Chinese).

Liu, F. S. C. (2009). Partisan orientation and selective exposure during Taiwan’s 2008 presidential election campaign. Electoral Studies, 16(2), 5170 (in Chinese).

Liu, F. S. C. (2010). Polarized news media and the polarization of the electorate. International Journal of Artificial Life Research, 1(1), 3550. doi:10.4018/jalr.2010102103

Ma, N. (2007). Political development in Hong Kong. Hong Kong: Hong Kong University Press.

Mattlin, M. (2004). Nested pyramid structures: Political parties in Taiwanese elections. The China Quarterly, 180, 10311049. doi:10.1017/S0305741004000736

McLeod, J. M., & McDonald, D. G. (1985). Beyond simple exposure: Media orientations and their impact on political processes. Communication Research, 12(1), 333. doi:10.1177/

009365085012001001

McLeod, J. M., Scheufele, D. A., & Moy, P. (1999). Community, communication, and participation: The role of mass media and interpersonal discussion in local political participation. Political Communication, 16(3), 315336. doi:10.1080/105846099198659 Mondak, J. J. (1995). Media exposure and political discussion in U.S. elections. The Journal of

Politics, 57(01), 6285. doi:10.2307/2960271

Mutz, D. C. (2006). Hearing the other side. New York, NY: Cambridge University Press.

Mutz, D. C. (2008). Is deliberative democracy a falsifiable theory? Annual Review of Political Science, 11(1), 521538. doi:10.1146/annurev.polisci.11.081306.070308

Mutz, D. C., & Young, L. (2011). Communication and public opinion: Plus c¸a change?. Public Opinion Quarterly, 75(5), 10181044. doi:10.1093/poq/nfr052

Pan Z., Shen, L., Paek, H.-J., Sun, Y. (2006). Mobilizing political talk in a presidential campaign: An examination of campaign effects in a deliberative framework. Communication Research, 33(5), 315345. doi:10.1177/0093650206291478

Popkin, S. (1991). The reasoning voter. Chicago, IL: University of Chicago Press.

Price, V., & Zaller, J. (1993). Who gets the news? Alternative measures of news reception and their implications for research. Public Opinion Quarterly, 57(2), 133164. doi:10.1086/269363 Scheufele, D. A., Nisbet, M. C., Brossard, D., & Nisbet, E. C. (2004). Social structure and citizenship: Examining the impacts of social setting, network heterogeneity, and informa- tional variables on political participation. Political Communication, 21(3), 315338. doi:10.

1080/10584600490481389

Schmitt-Beck, R. (1994). Intermediation environments of West German and East German voters: Interpersonal communication and mass communication during the first all-German election campaign. European Journal of Communication, 9(4), 381419. doi:10.1177/

0267323194009004002

Schmitter, P. C. (2001). Parties are not what they once were. In L. Diamond & R. Gunther (Eds.), Political parties and democracy (pp. 6789). New York, NY: Johns Hopkins University Press.

Schudson, M. (1997). Why conversation is not the soul of democracy. Critical Studies in Mass Communication, 14(4), 297309. doi:10.1080/15295039709367020

Sniderman, P. M. (2000). Taking sides: A fixed choice theory of political reasoning. In A. Lupia, M. D. McCubbins, & S. L. Popkin (Eds.), Elements of reason (pp. 6784).

New York, NY: Cambridge University Press.

Stroud, N. (2010). Niche news. Oxford: Oxford University Press.

Wattenberg, M. (1991). The rise of candidate-centered politics. Cambridge, MA: Harvard University Press.

(21)

Wu, C.-l. (2009). Semi-Presidentialism and divided government in Taiwan: Public perceptions of government performance. Issues & Studies, 45(4), 134. Retrieved from:http://iiro.nccu.

edu.tw/attachments/journal/add/4/1-34.pdf

Wu, C.-l., & Huang, C. (2007). Divided government in Taiwan’s local politics - Public evaluations of city/county government performance. Party Politics, 13(6), 741760. doi:10.

1177/1354068807080085

Wyatt, R. O., Katz, E., & Kim, J. (2000). Bridging the spheres: Political and personal conversation in public and private spaces. Journal of Communication, 50(1), 7192. doi:10.

1111/j.1460-2466.2000.tb02834.x