Content-Type: text/html

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
<a href="/cgi-bin/wa?LOGON=A3%3Dind0602A%26L%3DAEJMC%26E%3D0%26P%3D7968687%26B%3D--%26T%3Dtext%252Fplain%3B%2520charset%3Dus-ascii" >[log in to unmask]</a> with just the four words, "get help info aejmc," in the
body (drop the "").

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

Party affiliation, political ad perceptions and political 
involvement: Evidence from the 2004 Presidential campaign


Sangki Lee
Doctoral Student

Fuyuan Shen
Assistant Professor

College of Communications
The Penn State University

Address correspondence to the first author at

115 Carnegie Building
University Park, PA 16802-5101
814.574.1703
<a href="/cgi-bin/wa?LOGON=A3%3Dind0602A%26L%3DAEJMC%26E%3D0%26P%3D7968687%26B%3D--%26T%3Dtext%252Fplain%3B%2520charset%3Dus-ascii" >[log in to unmask]</a>

Paper submitted to the Mass Communication and Society Division, the 
Association for Education in Journalism and Mass Communication Annual 
Meeting, San Antonio, Texas, 2005.
0
Party affiliation, political ad perceptions and political involvement


Party affiliation, political ad perceptions and political 
involvement: Evidence from the 2004 Presidential campaign
  (Abstract)
	This research used a data from a survey during the 2004 presidential 
campaigns found that party affiliation was a significant factor in 
how individuals perceived the negativity and truthfulness of 
political ads. Specifically we found that people respond to negative 
political ads in accordance with their partisanship. Furthermore, it 
was found that party affiliation has significant effects on how 
negative perceptions of ads influence political involvement. 
Theoretical and practical implications of these findings will be addressed.

Party affiliation, political ad perceptions and political 
participation: Evidence from the 2004 Presidential campaign

Negative political ads and their effects on elections and candidates 
has been a subject of research and public debate almost since its 
inception. Over the years, there has been a substantial body of 
evidence showing that under various conditions, negative political 
ads can have serious and damaging effects on the targeted candidate . 
Several studies have focused on how political attack ads influence 
voters' intention to vote for the target candidate, with a majority 
of the results indicating that negative ads might lower voting 
intentions . Negative political advertising might also lead to 
backlash causing decreased liking and voting intent for the candidate 
sponsoring the negative ads . More recently, scholars have probed the 
unintended effect of negative advertising asserting that negative 
political ads could drive people away from polls and engendered 
distrust in government and the democratic processes .
Although a substantial body of studies found negative consequences of 
attacks ads on voters, overall finings regarding influences of 
negative political ads have generated somewhat inconsistent results. 
In contrast of Ansolabehere and Iyengar's (1995) findings, some 
studies, which analyzed NES voting data, indicated that exposure to 
negative political ads did not produce any demobilized effects on 
voters . This inconsistency might be emerged because one of important 
variables which significantly moderates the effect of negative 
political advertising. One of the variables might be voter's party 
affiliation .
Past studies found that party-bias effects substantially moderate 
voters' evaluation of competing candidates. By reinforcing support 
for affiliated candidates and reducing support for opponent 
candidates, voters' party identification leads their preference 
toward candidates more polarized . In the similar vein, 
Chang  suggested that voters' party affiliation decide how they 
respond to political ads which sponsored by competing candidates. He 
found that respondents showed consistent patterns of perceptions 
pertaining to evaluations of political ads in accordance with their 
party orientation.  In particular, voters indicated that they favored 
the negative ads which sponsored by their affiliated party than the 
ads which sponsored by opponent party .
Based on the reviewed evidence regarding party-bias effects on 
voters, this study attempts to explore how voters' party affiliation 
moderate their perceptions of negative political ads as well as the 
relative influence of negative perceptions of political ads on 
political involvement. In particular, this study predicts that party 
affiliation facilitates people's selective information processing and 
facilitated biased information processing makes people respond to 
negative political ads in selective way. Furthermore, it is also 
predicted that reinforced attitude through selective information 
processing consequently influences people's level of political involvement.
To examine these propositions, we analyzed a data from a survey 
during the 2004 presidential campaigns. By taking major advantages of 
survey method, we hope that we can test ongoing conflicting results 
regarding the influence of negative political advertising in real 
campaign environments and we can provide a clearer picture of the 
consequence of negative political advertising through examining a 
moderating role of party affiliation in political campaign.
Party identification and negative perceptions of political advertising
Bartels  indicated that party identification has strong effects on 
perceptions of the political world. According to him, party 
affiliation is "a pervasive dynamic force shaping citizens' 
perceptions of, and reactions to, the political world. Partisan bias 
in political perceptions plays a crucial role in perpetuating and 
reinforcing sharp differences in opinion between Democrats and 
Republicans." (Bartels, 2002, p. 138). Regarding the influence of 
partisanship in perceptions of political information, Campbell, 
Converse, Miller, and Stokes  argued that party identification 
facilitate a perceptual screen through which the individual tends to 
see is favorable to his party orientation. In the similar vein, 
Zaller  claimed that individual has a tendency to accept information 
which is congenial to their partisan values and to reject information 
which is not.
The effects of partisanship also have been explored under political 
advertising framework. Some studies regarding the role of 
partisanship in political advertising indicated that voter's party 
identification has strong influences on voter's candidate preference 
. Regardless of using positive or negative political ads, voter's 
evaluation or preference of competing candidates is to be polarized. 
This phenomena might be attained through strengthening support for 
affiliated candidate and reducing support for opposed candidates .
  Ansolabehere & Iyengar  argued that individual's vote-preference 
was reinforced in the line with their party identification after they 
were exposed to political ads. After exposure to campaign ads, 
participants showed reinforced vote-preference toward their 
supporting candidate. Similarly, Faber, Tims, and Schmitt (1990) 
found that participants showed their voting support for competing 
candidates was polarized in accordance with their party orientation.
Party identification also moderated the way people to evaluate 
political advertising messages. Ansolabehere & Iyengar  suggested 
that voters' party orientation has been reported to influence how 
they process political campaign information including political ads. 
According to Chang , the tendency to accept pro-attitudinal 
information and avoid counter-attitudinal information induce biased 
information processing of political advertising message, and this 
biased processing can be explained by the notion of 
selective-processing. Basically, this notion argued that processing 
bias is a function of perceiver's existing attitude . Empirical 
studies on biased information search in decision making have been 
predominantly carried out within the framework of dissonance theory . 
Festinger (1957) suggested that people have a tendency to avoid or 
evade threatening messages for averting cognitive dissonance. This 
premise has been supported by several studies which found that people 
actively attempt to seek out complementary information and avoid 
dissonance-producing information .  This type of biased information 
processing helps people to reinforce their readily established 
attitude or preference as well as to be comfortable without cognitive 
dissonance. To achieve this status under exposure to political 
advertising, it has been predicted that people may pay more attention 
to evaluative consistent information and avoid evaluative 
inconsistent information .
Regarding the party bias effects on perception of political ads, in 
particular, Chang  found that when participants were exposed to 
political adverting, they liked the ads were sponsored by a 
supporting candidate more than the ads sponsored by a competing 
candidate. Participants also perceived that the ads sponsored by 
their affiliated party were more persuasive than the ads sponsored by 
their competing party. In line with Chang's specific findings and 
general evidence of selective-processing literature, the current 
study predicts that respondents are more likely perceive that 
political ads which are critical of a supporting candidate are more 
negative and less truthful than the ads which are critical of a 
competing candidate. Driven by existing preference and attitude, 
people are likely to bias their perceptions of political ads in the 
line with their political affiliation. As the research evidence 
indicated, when people have preexisting positive attitude toward 
their supporting candidate, they are less likely to accept the 
negative messages about their supporting candidate and may perceive 
the attack messages as more negative because this negative messages 
are evaluative inconsistent information and produce cognitively 
dissonant information. Similarly, this biased response to political 
ads may influence participants' evaluation of message truthfulness. 
That is, it is plausible to predict that voters are likely to 
perceive the attack ads against their supporting candidate are less 
truthful than the ads against a competing candidate. Based on these 
predictions, this study proposes the following hypotheses:
H1: Respondents are likely to perceive that the ads attacking a 
candidate nominated by their supporting party are more negative than 
the ads against a candidate nominated by competing party.
H2: Respondents are likely to perceive that the ads attacking a 
candidate nominated by their supporting party are less truthful than 
the ads against a candidate nominated by competing party.
Negative perceptions of political advertising and political involvement
Past studies indicated that negative advertising might be more 
remembered than positive advertising (Faber et al., 1993). One 
plausible explanation of this phenomenon might be found from Gestalt 
principles. Lau  suggested that negative information may be 
inconsistent to what people are accustomed to and expect from 
advertising messages. As a result, negative information may stand out 
disproportionately. Furthermore, negative information is more likely 
to be weighted heavily than positive information in forming 
evaluations . Therefore, the information in negative political 
advertising may play an important role in candidate preference 
decision. Although negative ads were expected to have stronger 
effects on voter's preference decisions than were positive ads, it is 
premature to confirm this idea. In general, voters do not like to be 
exposed to negative political ads , and "a backlash effect may result 
if voters disapprove of a candidate using negative appeals." (Faber 
et al., 1993, p. 68). If it is so, what factor needs to be considered 
to explain this potentially bidirectional effects of negative ads? 
Political involvement may be one moderating variable to do it.
Although there are disagreements over how we define involvement , it 
has been generally regarded as an important mediator of the effects 
of political advertising . Faber et al.  argued that political 
involvement is positively related with the degree of impact which 
negative advertising exerts. They found people who are more involved 
and interested are most influenced by negative ads. However, some 
studies in which involvement was conceptualized as a characteristic 
of voters, indicated conflicting evidence regarding the effects of 
involvement under the realm of political advertising. For example, 
while Hofstetter, Zukin, and Buss  argued that voters who less 
involved learn more from political ads, others indicated that highly 
involved voters recall political ads more . To attempt to examine 
mechanisms of this inconsistency, this paper considered party-bias 
effects again. As mentioned before, voter's party affiliation plays a 
role as a perceptual screen. By using this perceptual filter, voters 
tend to accept information which is congenial to their partisan 
values and to reject information which is not .
  Based on this rational, it is plausible to expect that voters might 
pay attention to, and involve in processing negative political 
information differently in accordance of their party affiliation. In 
particular, voters are likely to involve in processing information 
within negative ads against affiliated candidates than the 
information in the ads against opponent candidates because the former 
information is much more critical of their partisan values than is 
the later information.
More specifically, it is expected that the perceived negativity of 
ads attacking affiliated candidates increase voters' political 
involveemnt. However, in the case of perceived truthfulness of ads 
against affiliated candidates, this paper expect a reverse 
relationship because truthful information which describe voters' 
affiliated candidate negatively is counter to what they want to see 
from political ads and their favored partisan values. Due to tendency 
to reject or avoid information which is not congenial to their 
partisan values, perceived truthfulness of ads against affiliated 
candidates may reduce voter's involvement. In contrast, because 
overall information about opponent candidate is relatively not 
important to their partisan values, the effects of negative 
perceptions of ads attacking opponent candidates may not be strong on 
involvement. To explore these predictions, this study proposes the 
following hypotheses.
H3: Perceived negativity of political ads will be significantly 
related with involvement in accordance with party affiliation. 
Specifically, the perceived negativity of ads against affiliated 
candidates will be positively related with involvement, while the 
perceived negativity of ads attacking competing candidates will not 
have strong influence on involvement.
H4: Perceived truthfulness of political ads will be significantly 
related with involvement in accordance with party affiliation. 
Specifically, the perceived truthfulness of ads against affiliated 
candidates will be negatively related with involvement, while the 
perceived truthfulness of ads attacking competing candidates will not 
have strong influence on involvement.
Method
A telephone survey was conducted in State College in Pennsylvania. A 
systematic random sample was drawn to increase the likelihood of 
representative sampling. The survey was completed within 3 days, from 
Nov. 23 to Nov. 25, before the Presidential Election Day in 2004. On 
campus phone numbers and business numbers were eliminated from the 
sample in this community. The sample drawn (after elimination of 
nonassigned numbers) comprised 1,166 numbers, 394 interviews were 
completed, and resulting completion rate was 33.80 %.
Measurements
The questionnaire was designed to include questions on perceptions of 
negative political advertising including perceived frequency, 
negativity, and truthfulness of the ads, political cynicism, party 
affiliation, and demographics.
Negative advertising perceptions. Three variables including perceived 
frequency, negativity, and truthfulness were measured to gauge 
respondents' perceptions of attack advertising. To estimate how a 
party identification influences on the way respondents evaluate or 
interpret attack advertising that were critical of John Kerry or 
George W. Bush, respondents were asked to indicate their perceived 
frequency, negativity, and truthfulness of attack ads, which were 
critical of the both candidates.
Perceived frequency of negative ads was measured by asking 
respondents the following question: "How often would you say that 
you've seen attack ads that are critical of John Kerry (or George W. 
Bush)?" This question employed 4 points scale such as 3 (Frequently), 
2 (Sometimes), 1 (Rarely), and 0 (Never). To assess the perceived 
negativity and truthfulness of negative ads, they were asked to 
answer to the questions, such as "How positive or negative do you 
feel about these ads that are critical of Kerry (or Bush)?" and "In 
your opinion, how truthful or untruthful is the information contained 
in these ads that are critical of Kerry (or Bush)?" Both of these 
questions employed scales ranging from 4 (Very negative) to 1 (Very 
positive) and from 4 (Very truthful) to 1 (Very untruthful).
Party affiliation. Respondents were asked to indicate the party with 
which they identified.  Four options were listed to choose 
respondents' party orientation including Democrats, Republicans, 
Independents, and Other. Among 374 respondents who identified 
themselves with one of categories, 169 respondents (45.2%) answered 
they were Democrats, 156 respondents (41.7%) were Republicans. Only 
33 respondents (8.8%) identified themselves as Independents and 16 
respondents (4.3%) chose other.
Involvement. According to Zaichkowsky , involvement can be understood 
as three different types:  physical, enduring, and situational 
involvements. In the studies of political advertising, physical 
involvement may reflect the level of election race . While enduring 
involvement refer to a relatively long-term, inherent interest in a 
message, situational involvement reflect a temporary relevance of a 
specific object or concern with a short-term outcome . When 
discussing political elections, enduring involvement might refer to a 
voter's general interest in politics while situational involvement 
would reflect concern about the outcome of specific election .
In this study, the following two items were used to measure the 
concept of involvement: (1) "How interested have you been in the 
political campaigns so far this year?"; (2) " How important is the 
outcome of the presidential election to you?" Four-point scales were 
used to measure responses to aforementioned two questions. The both 
items were combined into a single involvement index. The Cronbach's 
alpha for involvement index was .65, indicating that two scores form 
a reliable index.
Results
Hypothesis 1 and 2 predicted that perceived ads negativity and 
truthfulness would vary as a function of people's party 
identification. To explore these hypotheses, separate ANOVA tests 
were run with party affiliation as a fixed factor and the perceived 
negativity and truthfulness of the ads attacking either Bush or Kerry 
as the dependent variables. Results showed that that people who 
identified themselves as Democrats perceived that the ads attacking 
Kerry were more negative, M = 2.11, than did people who identified 
themselves as Republican, M = 1.39, F(1, 233) = 47.72, p < .001. 
Similarly, Republicans perceived that the ads were critical of Bush 
were more negative, M = 2.32, than did Democrats, M = 1.67, F(1, 233) 
= 29.85,  p < .001.
Furthermore, it was found that Democrats perceived the ads attacking 
Kerry as less truthful, M = 1.89, than did Republicans, M = 2.82, 
F(1, 233) = 79.68,  p < .001. Similarly, Republican evaluated the ads 
attacking Bush were less truthful, M = 2.24, than did Democrat, M = 
2.84, F(1, 233) = 28.61,  p < .001. In light these findings, we 
therefore conclude that both Hypothesis 1 and Hypothesis 2 were fully 
supported. The means and standard deviations of above analyses were 
summarized in Table 1.
___________________________
Table 1 about here
___________________________

To test the relative influence of ad perceptions on respondents' 
political involvement, we ran two separate multiple regressions with 
political involvement as a dependent variable. To test Hypothesis 3, 
the perceived negativity against Kerry and Bush ads was entered as 
independent variables, and for testing Hypothesis 4, the perceived 
truthfulness of ads against Kerry and Bush was used as independent 
variables. In addition, we used a block of control variables that 
included the perceived frequency of ads against both candidates. The 
control variables were entered in the first, and it was followed by 
the predictor variable.
To explore the effects of partisanship on relationships between ads 
perceptions and involvement, this study analyzed the relationships 
based on their party identification. For respondents who identified 
themselves as Democrats, Hierarchical Multiple Regression revealed 
that the full regression equation with control variables and 
perceived campaign negativity was significant on political 
involvement, F(4, 127) = 5.40, p < .001, R2 = .15. The results also 
indicated that the same regression model with two variables of 
perceived truthfulness of ads as independent variables was 
significant on the involvement, F(4, 132) = 2.80, p < .05, R2 = .08. 
There was no significant control variable identified. As predicted, 
respondents' negative ads perceptions are significantly related with 
involvement. The results indicated that negativity of ads against 
Kerry was positively related with involvement, while truthfulness of 
ads against Kerry was negatively related with involvement. Negativity 
and truthfulness of ads against Bush were not significantly related 
with involvement. The results of all above analysis were displayed in Table 2.
___________________________
Table 2 about here
___________________________

For respondents who identified themselves as Republican, Hierarchical 
Multiple Regression showed similar patterns of results which we found 
above. The full regression equation with control variables and 
perceived campaign negativity was significant on political 
involvement, F(4, 122) = 5.46, p < .001, R2 = .15. The results also 
indicated that the same regression model with two perceived 
truthfulness of ads as independent variables was significant on the 
involvement, F(4, 116) = 5.72, p < .001, R2 = .17. The results 
indicated that two control variables, perceived frequencies of ads 
against Kerry and against Bush, as significant predictors of 
Republican's involvement. However, frequency of attacks ads against 
Kerry was negatively related with involvement, while frequency of 
attacks ads against Bush was negatively related with it. As 
predicted, Republicans' negative perceptions of political ads are 
significantly related with involvement. The results indicated that 
negativity of ads against Bush was positively related with 
involvement, while truthfulness of ads against Bush was negatively 
related with it. Negativity and truthfulness of ads against Kerry 
were not significantly related with involvement. The results of all 
above analysis were displayed in Table 3. Therefore, Hypothesis 3 and 
4 were fully supported.
___________________________
Table 3 about here
___________________________

Discussion
This paper set out to examine how party affiliation affects audience 
evaluations of political ads by political candidates. Using data from 
a survey during the 2004 presidential election, we found that voter's 
party affiliation moderated their perceptions of negative political 
ads. When responses of participants who identified with either 
Democrats or Republicans were analyzed separately, interesting and 
consistent results emerged. The results indicated that party 
affiliation was a significant factor in how individuals perceived the 
negativity and truthfulness of political ads. Specifically we found 
that Democrats rated attack ads by the Bush campaigns significantly 
more negative and less truthful than Republicans. Similarly, 
Republicans evaluated ads by Kerry as more negative and less 
truthful. It is important to note that these results confirm the idea 
that voters evaluate or perceive negative political ads in selective 
way and this selective information processing might reinforce their 
preexisting attitude. These results further support Zaller 's 
idea  that party affiliation plays a role as a perceptual screen 
which people tend to evaluate information in favorable way to their 
party orientation. In addition, these results were consistent with 
Chang (2003)' findings that voters' preexisting attitude or 
preference toward candidates who they support and who they don't 
influences how they response to political ads.
This paper also explored how party affiliation affects the 
relationship between negative perceptions of ads and the level of 
political involvement. The results indicated that party affiliation 
is a significant factor to moderate how negative political 
advertising influence voters' political involvement. As expected, 
when we analyzed responses separately by their party affiliation, 
consistent patters of results emerged. Specifically perceived 
negativity of ads attacking affiliated candidates increased voters' 
political involvement and perceived truthfulness of ads against 
affiliated candidates reduced voter's involvement, while both 
perceived negativity and truthfulness of ads against opponent 
candidates did not have significant effects on involvement. This 
evidence also support the role of party orientation as "a perceptual 
screen", and the idea that party affiliation make people process 
information in selective way . It is important to note that when 
responses of all subjects without considering their party 
affiliation, were analyzed, perceived negativity and perceived 
truthfulness of attack ads did not exert significant influences on 
political involvement. Only when the responses were analyzed 
separately by their party identification, significant effects which 
were consistent were emerged. Therefore, if party affiliation was not 
considered in analyses, this study might not found any significant 
effects of negative political advertising on political involvement.
	Taken together, these findings mean that the nature or tone of 
political ads variable as people's political affiliations differ. 
This indicates negativity or truthfulness is a relative concept. As 
this study found, subjectivity emerged from individual party 
affiliation, substantially moderated voters' responses to negative 
advertising and the influence of negative advertising on involvement 
under the realm of political campaign. Therefore, these finings 
recommend that future research attempting to explore the consequence 
of political ads may need to take into account various moderating 
variables including party affiliation to explain variance of the 
effects of political ads on voters. We hope that theses results may 
contribute to explain inconsistent findings regarding how political 
ads affect voter's decisions.
	Although the findings from the present research are important, there 
are several limitations we need address. First, we used self-reported 
exposure. Although we controlled perceived frequency of exposure to 
negative ads in all analyses we did, the self-reported score was 
likely to be exaggerated. Because negative information is more likely 
to be weighted heavily than positive information in forming 
evaluations  and people tend to more actively accept information 
which is critical of their partisan value , depending on how strongly 
people identify themselves with their affiliated party, their 
self-reported score might be elevated substantially. Therefore, this 
elevated score might affect our findings. In the similar vein, 
findings that were related with all negative ads perceptions 
variables including frequency, negativity, and truthfulness should be 
carefully interpreted.
Second, although we controlled for two variables of perceived 
frequency in the regression models, other factors could account for 
variance of involvement we have observed. In light of these findings, 
we suggest that future research replicate these findings in different 
campaigns or explore other factors that might be significant 
mediators of political involvement.
References







Table 1.
Negativity and truthfulness of Ads against Kerry or Bush
Ads against Kerry
Ads against Bush
Negativity
Truthfulness
Negativity
Truthfulness
Democrat
M
2.11
1.89
1.67
2.84
SD
.07
.07
.08
.07
Republican
M
1.39
2.82
2.32
2.24
SD
.07
.07
.08
.08


Table 2.
Regression equations predicting influence of negative perception on 
Democrat's involvement
Equation 1
Equation 2
Equation 3
Control variables
Frequency of attacks ads against Kerry
.15
.13
.13
Frequency of attacks ads against Bush
-.02
.01
-.01
Predictor Variable
Negativity of Kerry ads
.34***
Negativity of Bush ads
-.18
Truthfulness of Kerry ads
-.23**
Truthfulness of Bush ads
.09
R2
.02
.15
.08
    Note: *p <.05, **p <.01, or *** p < .001.   N = 131
	



Table 3.
Regression equations predicting influence of negative perception on 
Republican's involvement
Equation 1
Equation 2
Equation 3
Control variables
Frequency of attacks ads against Kerry
-.27**
-.21*
-.26**
Frequency of attacks ads against Bush
.24**
.19*
.18
Predictor Variable
Negativity of Kerry ads
-.12
Negativity of Bush ads
.28**
Truthfulness of Kerry ads
.11
Truthfulness of Bush ads
-.25**
R2
.08
.15
.17
    Note: *p <.05, **p <.01, or *** p < .001.   N = 131