|
This paper was presented at the Association for Education in Journalism and
Mass Communication in Toronto, Canada, August 2004.
If you have questions about this paper, please contact the author
directly. If you have questions about the archives, email
[log in to unmask] For an explanation of the subject line, send email to
[log in to unmask] with just the four words, "get help info aejmc," in the
body (drop the "").
(Oct 2004)
Thank you.
Elliott Parker
************************************************************************
What If Everyone Were Fully Informed and Talked about Politics Everyday?
Political Knowledge, Political Conversation, and Public Opinion
Nam-Jin Lee (Doctoral Student)
School of Journalism and Mass Communication
University of Wisconsin-Madison
Contact information
6401 Offshore Drive, APT 215
Madison, WI 53705
Phone: (608) 661-9320
E-mail: [log in to unmask]
What If Everyone Were Fully Informed and Talked about Politics Everyday?
Political Knowledge, Political Conversation, and Public Opinion
Abstract
The simulation modeling technique has often been used to estimate
statistically what public opinion would look like if people were fully
informed. By correcting not only the low levels and uneven distribution of
political knowledge but also those of political conversation in simulating
higher-quality opinion, this study found that people's everyday
conversation with family and friends has a substantial influence on policy
preferences. The simulation results also suggest that the influences of
conversation do not merely supplement the information effects but often
times operate as countervailing forces on opinion changes.
What If Everyone Were Fully Informed and Talked about Politics Everyday?
Political Knowledge, Political Conversation, and Public Opinion
One of the ongoing questions among public opinion researchers is whether
the apparent public ignorance of public affairs is a problem for democratic
politics. This question, according to Schlesinger and Lau (2000), concerns
how to resolve the paradox between the democratic ideal of an informed
citizenry and the reality of public ignorance and apathy. This paradox is
also represented in competing empirical findings that characterize the
public as, on the one hand, knowing "virtually nothing about the public
issues" (e.g., Delli Carpini & Keeter, 1996) and, on the other hand, being
able to make sense of complex public issues (e.g., Gamson, 1992).
Underlying these competing views on citizen competence are diverging
visions of public opinion in a democracy that range from, for example,
Habermas' (1989) view that public opinion, which the author claims
comprises "the critical reflections of a public competent to form its own
judgments" (p. 90), is the only legitimate source of justification for
political domination to that of Lippmann (1922) who argues that the
democratic model in which pubic opinion is a governing force is "not
workable. And…it is not even thinkable" (p. 363).
Also complicating the aforementioned paradox are public opinion polls and
surveys that are currently considered to be the dominant indicators of
public opinion. That is because, while most researchers agree that
individual opinions recorded in opinion surveys are mostly ill-informed,
unstable, fragmented, and inconsistent (Converse, 1964; Entman & Herbst,
2002), the very ability of surveys to convey the voice of the people has
been widely debated. It is hardly new to cast doubts on whether opinion
surveys can tell us reliably what the people really think and feel. Most
familiar to researchers in the field are such methodological problems as
question wording effects, sampling problems, and nonresponse errors, to
name just a few. Some researchers also doubt that standard opinion surveys
can adequately capture what citizens bring in when they think or talk about
politics (e.g., Gamson, 1992). Further, other researchers criticize these
opinion surveys as having perpetuated a particular conception of public
opinion—that is, the individualistic and measurement oriented conception
(see Price, 1992)—and criticize this conception as being atheoretical in
nature and as failing to situate public opinion as a social and political
construction (Blumer, 1948; Bourdieu, 1972; Herbst, 1998). All these
critiques question ultimately the extent to which general opinion surveys
tap something close to the classic conception of informed and deliberative
opinion (Herbst, 1993; Price, 1992).
In response to the questions raised on the quality of public opinion
recorded in the opinion surveys, two quite different approaches have
emerged to provide alternative measures of public opinion—more correctly,
aggregated policy preferences—that are of higher quality than those
recorded in typical mass opinion surveys. Both the deliberative opinion
polls (e.g., Fishkin, 1991) and the simulation approach conducted by
Bartels (1996), Delli Carpini and Keeter (1996), and Althaus (1998, 2003)
share the aim of providing the estimates of, though hypothetical in nature,
"higher-quality" opinions.[1] The former method models "what the electorate
would think, if, hypothetically, could be immersed in intensive
deliberative process" (Fishkin, 1991, p. 81). In the typical deliberative
polls, a national random sample is drawn, contacted, and then brought to
gather at a specific location for participation in, typically, a
weekend-long deliberative forum on some policy issues. Combining scientific
random sampling and deliberation, Fishkin and Luskin (2002) argue, this
procedure "ensures (with extremely high probability) a microcosm of the
interests that need to be articulated—and responded to—in any serious
deliberation on policy issues" (p. 26). While not offering a mirror of
actual "raw" opinion, Fishkin (2000) further argues that this "refined"
opinion through deliberative forums represents a "picture of counterfactual
yet more informed opinion" (p. 17). Owing to these features of deliberative
polling, Fishkin (1991) maintains that a deliberative opinion poll is
"prescriptive, not predictive. It has a recommending force" (p. 81).
In contrast, the simulation methods, instead of conducting an actual
deliberation of a representative sample of people, statistically correct
the low levels and uneven distribution of political knowledge. This
correction is done by manipulating survey respondents' levels of political
information—more concretely, by replacing the actual knowledge scores of
each survey respondent with the highest possible value so that respondents
are made "fully informed," at least statistically. In particular, this
method is intended to reveal "information effects," which Althaus (1998)
defines as a "bias in the shape of collective opinion caused by the low
levels and uneven social distribution of political knowledge in a
population" (p. 545).
Of particular concern in this paper is the certain tendency among
researchers to directly equate these two approaches with each other despite
their apparent differences. That is, one may think that the simulation
modeling technique, with its relatively simple statistical techniques and
little expenditure of money, does essentially the same job as the
deliberative polling. Recent attempts to directly compare the result
attained from the two methods may reinforce this tendency. Sturgis (2002),
for example, attempted to test the validity and reliability of these two
methods by comparing the two methods performed on the same sample of
respondents. Based on the deliberative opinion poll data gathered in
Britain, the author found that opinion changes derived from the
deliberative opinion poll and from the simulation methods were essentially
the same, providing some evidence for "concurrent validity." That is,
opinion changes obtained from statistical simulation actually mirrored
those from the deliberative forum.
As Sturgis admitted, however, by comparing just the outcomes from the two
methods, this way of validation sidesteps rather than confronts the
so-called "black-box problem" that is attributed to different ways of
estimating higher-quality opinion (Price & Neijens, 1997, 1998). That is,
given that several intervention methods—e.g., provision of information,
private deliberation, public discussion—were used in the deliberative
opinion polls, one cannot identify reliably whether the opinion changes in
the deliberative forums, if any did exist, were caused specifically by
deliberation itself or by the information that participants acquired in the
forums. By leaving this problem untouched, Sturgis's (2003) direct
comparison of deliberative opinion polls and simulation methods may provide
the impression that the opinion changes that had surfaced in the
deliberative opinion polls were essentially comparable to "information
effects" rather than to "deliberation effects."
This paper represents an attempt to test, though quite indirectly, the
validity of the thesis that opinion changes in simulation models are
comparable to those in opinion forums. In other words, this paper asks
whether deliberative opinion polls and simulation methods gauge essentially
the same thing. For this aim, this study simulates another version of
"higher quality opinion," by using the same simulation techniques that were
used previously. This new version of simulated opinion is based on the
theoretical arguments, and recent empirical evidence, that, in Katz's
(1992) words, "conversation is missing" in the previous simulation
modeling. My version of simulation asks what would happen if increased
political knowledge comes hand in hand with increased political talk. That
is, this study tries to estimate what the collective preference would look
like if people were fully informed and talked about politics everyday with
their family and friends. By correcting not only the low levels and uneven
distributions of political knowledge but also those of political
conversation in simulating higher-quality opinion, this study tries to
estimate what might be called "conversation effects" above and beyond
information effects.
The basic logic for testing the aforementioned thesis is that, if we find
just a minimal degree, if any, of "conversation effects" (which many
political theorists and empirical researchers suggest have a deliberative
quality with real political consequences), or if the conversation effects
play mainly a supplementary role to the information effects, then this
would suggest that the aforementioned thesis may not so widely differ from
the reality. But if the conversation effects are found either to be
substantial or to operate in complicated ways, we should call into question
the validity of the aforementioned thesis. In addition, it is hoped that,
by shedding some light on the roles of "talking politics" in the estimating
of collective policy preferences, this paper will contribute to efforts to
better gauge high-quality opinion.
Simulation of Higher-Quality Opinions
The simulation modeling techniques, first used by Delli Carpini and Keeter
(1996) and Bartels (1996), and most comprehensively employed by Althaus
(1998, 2003), represent attempts to answer statistically the question of
what the public would think if it could be fully informed. As Converse
(2000) commented, these simulation models provide much more "elegant" ways
of estimating informed opinion than, say, models that gather estimates of
informed opinion simply by discarding a segment of the respondents who can
be considered ill-informed. The simulation method isolates an informed
preference within relevant categories of demographic characteristics that
may reflect competing interests in the real political world, thereby
preserving, rather than discarding, the actual numeric weights of those
categories (Althaus, 1998). That is, this method imputes the preference of
the most highly informed members of a given demographic group to all
members of that group.
Typically, this is done through the following procedures:
1. The simulation modeling estimates a baseline equation from regressing
opinion items on an index of political knowledge, demographic variables,
and the interaction terms between the two. The inclusion of interaction
terms between political knowledge and demographic variables ensures that
the impact of increased knowledge on preferences acts differently in
different demographic groups;
3. each respondent's score on political knowledge is changed to the highest
possible value;
4. Individual-level fully informed predicted probabilities on each opinion
item are obtained by plugging the coefficient values obtained from the
baseline equation into each respondent's actual scores on the demographic
variables, substituting only the new values of political knowledge and the
interaction terms; and
5. Collective level fully informed opinions are attained by taking the mean
of individual predicted probabilities.
In this way, the mean values provide the estimates of what collective
preference would look like if everyone were fully informed but still
different in terms of their demographic characteristics.
As is clear in the above procedure of estimating fully informed opinion,
this technique is based on the assumptions that citizens who share similar
socio-democratic characteristics also share the same interests and that, if
the ill-informed had (possibly to the maximum) more information, they would
develop the same preferences as those who are currently informed. While one
can challenge these assumptions, validity of simulated informed opinion
largely depends on the specification of the baseline models used to
estimate the information effect and the measurement of political knowledge
(Sturgis, 2003).
In particular, this technique has been used to challenge two prominent
views on public opinion recorded in surveys. First, it has been used to
counter the so-called "low-information-rationality" thesis by showing that
there exist substantial information effects. The basic tenet of
low-information rationality is that "rationally ignorant" citizens, though
they lack factual information and political interests, nevertheless manage
to approximate "enlightened preferences" by using various heuristics or by
taking cues from the elites (e.g., Popkin, 1991; Sniderman, Brody, &
Tetlock, 1991). If the low-information rationality works, the simulated
fully informed opinions should not be much different from the actual
opinions. Althaus (1998, 2003), Bartels (1996), Delli Carpini and Keeter
(1996), Gilens (2001), and Sturgis (2003) all demonstrated that this is not
the case. Information does matter. More often than not, correcting the low
levels and the uneven distribution of political information significantly
skewed the collective preference in one way or another (Althaus, 1998,
Bartels, 1996; Gilens, 2001).
Second, this technique has also been used to challenge the views
corresponding to the so-called "collective rationality" (e.g., Page &
Shapiro, 1992). Basically, the advocates of the collective rationality
thesis argue that, although individual-level opinions are unstable and full
of "noise" associated with the measurement error of surveys, the collective
level opinions are stable and rational because of the power of aggregation.
That is, aggregation of individual responses "cancels out" noises that are
said to be random, thereby leaving only the underlying meaningful component
of public opinion. However, using the simulation models, both Althaus
(2003) and Duch, Palmer, and Anderson (2000) demonstrated that error terms
are usually not random and that they are related systematically with
various factors including political knowledge.
Whereas this method provides a sophisticated way of obtaining the
estimates of higher-quality opinion, there also exists a potential danger
in this approach. While not changing the demographic characteristics of the
respondents, this approach essentially confers elites' mode of reasoning to
the general public. In other words, as Althaus (1998) stated, "this method
imputes to all respondents the information processing strategies and
cognitive styles employed by well-informed people" (p. 549). That is, this
approach employs "elite" opinions, or rather the elite's mode of forming
opinions (i.e., full information and rational reasoning based on such
information) as the criteria with which to show what the opinion
distribution would look like if the public as a whole operated in that mode.
However, as Gamson (1992) demonstrated, average people, though deficient in
factual knowledge, are nonetheless capable of sophisticated, coherent, and
meaningful conversations about public issues. Schlesinger and Lau (2000)
further argue that previous efforts to reconcile seemingly inconsistent
findings on citizen competence are limited in that they mainly rely on how
the formation of public opinion is constrained relative to elite judgment.
What is needed, these authors argue, is an identification of the modes of
reasoning that can explain "how the average citizen can assess policy
matters without resource to much knowledge of the political context or the
sort of ideological framework that help shape elite reasoning" (p. 611).
This paper identifies people's everyday conversations with their family and
friends as one possible candidate for people's own modes of forming opinion.
Political Conversation and Deliberative Democracy
There seems to exist a renewed interest in "political talk" among scholars
in communication and other related fields. With the widespread use of
electronic, interactive media such as the Internet and radio and television
talk shows, talk as a mode of communication exerts a significant impact on
the practice of political communication (Pan & Kosicki, 2001). Emphasizing
the significance of talk, Katz (1995) states that "the elementary building
block of participatory democracy is conversation" (p. xxx). Chambers (2003)
also maintains that "Talk-centric democratic theory replaces voting-centric
democracy" (p. 308).
This renewed interest in political talk, in large part, accompanied a
growing interest in the deliberative theory of democracy. The deliberative
theory of democracy is presented as a normative theory according to which
"the public deliberation of free and equal citizens is the core of
legitimate political decision making and self-government" (Bohman, 1998, p.
401). Advocates of the theory hold that a citizenry's deliberation on
public issues will produce more reasonable conclusions, resulting from more
informed, rational, and impartial opinions. As a promising mode of public
deliberation, conversation has attracted the enthusiastic attention of
scholars. Much earlier, Dewey (1939) championed the virtue of conversation,
saying that "the heart and final guarantee of democracy is in free
gathering of neighbors on the street corner to discuss back and forth what
is read in uncensored news of the day, and in gathering of friends in the
living rooms of houses and apartments to converse freely with one another"
(¶ 10). Habermas (1974) wrote, "A portion of the public sphere comes into
being in everyday conversation in which private individuals assemble to
form a public body" (Habermas, p. 198). Kim, Wyatt, and Katz (1999) even
contended that conversation is the prototypical and ideal form of
deliberation in the public sphere.
Whereas the virtue of conversation has been widely recognized, scholars
disagree wildly over the question of what type of talk serves democracy.
According to Wyatt, Katz, and Kim (2000), the typical disagreement can be
found between those who hold that "casual conversation and informal
conversation include political topics that breed public opinion" and those
who hold that "conversation and political talk are different things" (p.
73). Both of these competing arguments distinguish the two types of talk.
Borrowing Wyatt, Kim, and Katz's terms (2000, pp. 99-100), I call these two
conceptions of political talk as follows: formal deliberation
(rule-governed and goal-oriented talk) and ordinary political conversation
(causal, voluntary talk among people who typically know each other).
Schudson (1997), for example, argues that ordinary political
conversation—"sociable talk" in his words—has little to with democracy.
According to the author, political conversation in a democracy should have
a certain purpose: solving problems, deciding public policies, or
protecting one's own interests. The author further argues that democratic
talk is not spontaneous but essentially rule governed. Democratic talk is
inherently uncomfortable because it involves disagreements and conflicts.
The author thus wrote that "in what we might term 'truly public'
conversation, citizens talk with other citizens who may not share their
views and values" (p. 302). This type of formal problem-solving
conversation, according to the author, is at the heart of democracy, and
everyday sociable conversation is not.
By contrast, Wyatt, Katz, and Kim (2000; see also Kim, Wyatt, & Katz, 1999;
Wyatt, Kim, & Katz, 2000) argue that the life-world is not sharply divided
into the public sphere and the private. Rather, political conversation and
personal conversation form a continuum with bridging topics (e.g.,
education and crime) providing common ground. Political conversation
conducted in private, even among family and friends, has political
consequences. That is, talking about politics has positive effects for both
citizens and the democratic process by improving the quality of public
opinion and by encouraging political action. Thus the authors stated that
"democratic culture receives its most concrete realization" in this form of
ordinary political conversation (Wyatt, Katz, & Kim, 2000, p. 72).
In the face of these competing claims about the democratic potential of
everyday conversation, empirical studies may shed some useful light on the
topic in question. Though empirical studies about this topic are still
scarce, a brief survey of those studies would be helpful.
Some studies characterize the state of "talking politics" in a negative
tone. Basing their research on the survey data of the 1994 Eurobarometer
and the 1994 American National Election Study, Bennett, Flickinger, and
Rhine (2000), for instance, summarized that the typical citizen talked
about politics only sporadically. Only a small percentage of citizens (15%
of the British and 13% of the Americans) claimed to talk about politics
nearly every day. But half of the people felt free to speak with most of
the people whom they encountered. Similarly, Conover et al. (2002), using
surveys and focus group methods, made a claim that the low levels and
inequalities characterize political discussion in practice. Moreover, their
focus groups avoided argumentation. The discussants in the focus groups
typically neither gave reasons or justifications for their own preferences
nor wanted to listen to reasons for different perspectives.
Representing a more positive characterization of everyday conversation,
however, Wyatt, Katz, and Kim (2000) showed that political conversation is
greatly related to higher levels of both opinion expression and opinion
quality and to greater political participation. McLeod et al. (1999) also
found that the higher levels of interpersonal discussion have a positive
impact on issue reflection (i.e., respondents' thinking, recalling, and
following up on political information obtained from the media) and
anticipated forum participation.
In addition, according to Dutwin (2003), it is not knowledge or education
but conversation that is the most powerful and important predictor of
further conversation and deliberation. That is, the more one talks about
politics, the more articulate their interests, opinions, and knowledge
become, and thus the more capable such people are of voicing their
interests, opinions, and knowledge in political discussion. So Dutwin
argues that "individuals with high levels of prior political conversation
have more crystallized opinions ready and waiting to be vocalized" (p.
258). By comparing cross-sectional survey data and the public forum survey
data gathered before and after forum participation, Barabas (2003) also
shows that people's everyday political talk essentially yields the same
kind of opinion changes as the talk that occurs through participation in
public forums.
Although there is little definite evidence one way or another on the
competing claims, recent empirical evidence seems to suggest, at least, the
potential of everyday political conversation for providing a promising
venue for public deliberation. It is on the basis of these recent empirical
findings that the following claim is made in this paper: "conversation
effects" may be an important indicator of "deliberation effects." In order
for simulated opinion to resemble deliberative opinions in the public
forums, the role of political conversation should be considered. Therefore,
it is further argued here that, before directly comparing the outcomes from
the deliberative opinion polls with the ones from the simulation models, as
in Sturgis (2003), we need to use the simulation models in order to examine
whether conversation effects exist.
Methods
This study simulates two sets of "higher-quality" opinions: hypothetical
opinions of full-information and those of full-information and
full-conversation. Following Althaus (1998), Gilens (2001), and Sturgis
(2003), the simulation models are two logistic regression models. The first
model includes political knowledge, demographics, and sets of interaction
terms between these two. The second adds political conversation and its
interaction terms with political knowledge and with demographics. The model
for political knowledge takes the following form:
(1)
The model for both political knowledge and political conversation takes the
following structure:
(2)
where is respondent 's dichotomous policy preferences, is respondent 's
level of political knowledge, is respondent 's level of political
conversation, is respondent 's score on the jth democratic variable, and
are the respective interaction terms of political knowledge and political
conversation with the demographic variables, and are the regression
coefficients, and is the error term for the th observation.
The estimated coefficients from the above models were used to simulate the
two sets hypothetical policy preferences for respondents with less than the
maximum level of political knowledge and political conversation. Two sets
of simulated opinions were then compared with each other and with the
actual opinions.
The two simulation models were fitted to the 2000 American Nation Election
Study (ANES) data. This dataset was chosen because unlike the other waves
of ANES datasets, it includes a battery of items measuring the respondents'
networks of political discussion, some of which were used in the current
study as measures of political conversation. A total of 28 opinion items
was chosen for simulation. Theses items cover almost all the opinion items
that can be reasonably recoded into dichotomies (for more detailed wordings
and recoding rules of the opinion questions, see the Appendix). Following
Althaus (1998), these opinion questions were classified into four issue
categories: foreign, fiscal, operative, and social issues.[2]
The measure of political knowledge is the 17-item additive index,
originally developed by Delli Carpini & Keeter (1996). The index has a
maximum value of 21, with M = 10.91, SD = 5.17, and Cronbach's a = .874.
The index items tap the knowledge of national political leaders, partisan
control of Congress, and the ideological positions and the policy stances
of major presidential candidates and their parties (for details on this
measure, see the Appendix).
Delli Carpini and Keeter (1996) defined political knowledge as "the range
of factual information about politics that is stored in long-term memory"
(p. 10). Although some skepticism has been raised as to whether these
factual knowledge tests are, indeed, a manifestation of people's political
knowledge (Sturgis, 2003), this measure was comprehensively validated by
Delli Carpini and Keeter. It is widely accepted by many scholars as an
effective "diagnostic of more profound differences in the amount and
accuracy of contextual information voters bring to their judgment"
(Converse, 2001, p. 333).
The measure of political conversation used the following 4 items:
(a) How many days in the past week did the respondent talk about politics
with family or friends? (possible values ranging from 0 to 7)
(b) How many discussants did the respondents identify? (none to four
discussants)
(c) How often does the respondent discuss politics with the first
discussant? (a 4-point scale from "never" to "often")
(d) How often does the respondent discuss politics with the second
discussant? (a 4-point scale from "never" to "often")
Since each measure has different scales, the raw values of each
respondent were first transformed into percentiles and then averaged across
the four items to construct an index of political conversation. The valid
values ranged from .147 to .878 (M = .500, SD = .229), and Cronbach's a was
.846. When the simulated opinions of full-information and full-conversation
were calculated, the maximum value of .878 was assigned to all respondents'
political conversation scores.
Surely, this measure of political conversation, though better than a
single-item measure, is still limited. Rather than tap the extent to which
respondents talked about specific issues, this item measures a general
behavior pattern. Furthermore, one should not expect the respondents to
agree on what constitutes "politics." Thus, Katz (1995) maintained that
"this evokes a too generalized self-portrait that gives too little insight
into the dynamics of everyday talk about politics" (p. xxx). However, the
current index of political conversation was constructed utilizing all items
available in the ANES data and thus can be considered to be a stronger
index than a single-item measure.
The reciprocal relationship between political conversation and political
knowledge is well documented (e.g., Bennett, Flickinger, & Rhine, 2000).
That is, the more that people talked about political affairs, the more
knowledgeable they became, and vice versa. Reflecting this relationship, in
the current data, political conversation and political knowledge are highly
correlated with each other, Pearson r = .502, p < .001. Quite possibly,
this high correlation may have generated the problem of collinearity.
However, both Althaus (1998) and Gilens (2000) suggested that the
collinearity problem, in the case of the simulation modeling, does not
affect the unbiasedness of the estimated coefficients.
Demographic variables were constructed following Althaus (1998). A set of
23 demographic variables reflects respondents' education, income, age,
partisanship, race, gender, region, marital status, parental status,
religious affiliation, union membership, home owner status, children in the
household, financial status, and occupational status. Among the 27
demographic variables that Althaus (1998) used, the following 4 variables
were not included in this study because they were simply not available in
the 2000 ANES dataset: receiving welfare, receiving other benefits, rural,
and urban.
There were 170 item-missing cases on the income variable, but other
independent variables have missing values of less than 10. All missing
values in the independent variables were imputed using the EM algorithm
(Schafer, 1997). Although the multiple imputation method is preferable,
this study, in order to reduce computational complexity, used the single
imputation method provided by the SPSS 11.5 Missing Value Analysis.
Results
In this study, using Equation 1 and Equation 2 (see p. 15) as baseline
models, two sets of hypothetical high-quality opinions were calculated on
28 opinion questions in the 2000 ANES data. Before presenting the detailed
results, a set of likelihood ratio tests was performed in order to see
generally whether "conversation effects" exist. These tests compare the log
likelihoods of the unrestricted model (Equation 2), which includes all the
variables, and the restricted "information effects model" (Equation 1),
which includes demographics, information, and interaction terms between
these two. This is essentially the model comparison between Equation 1 and
Equation 2. So these tests looked at whether "conversation effects" exist
above and beyond information effects.
The test was found to be significant at the .05 level in 8 out of 28
opinion questions. This result suggests that in 8 out of 14 opinion
questions, adding political conversation and its interaction terms with
demographics significantly increased the model fit. Though not significant
at the .05 level, the p values of 3 opinion items were less than .10. Then,
slightly more than a third of all the p values were less than .10.
This result looks disappointing at first glance. However, before drawing a
hasty conclusion, it should be noted that this test looks at only the
additional conversation effects after considering highly significant
information effects. It should also be noted that the significant results
in 8 items were obtained despite the existence of a large number of
statistically insignificant interaction terms between political
conversation and demographics. These results suggest that at least a modest
degree of conversation effects exists in the current data.
[Table 1 about here]
More detailed results are presented in Table 1. The third column (Column B)
in Table 1 lists hypothetically fully informed policy preferences. The
fourth column (Column C) refers to the hypothetical policy preferences of
full-information and full-conversation. The fifth column (Column D) shows
"information effects," defined here as the difference between the fully
informed opinion and the observed opinion (i.e., Column B – Column A). The
last column (Column E) shows what I call here "conversation effects,"
defined as the difference between the simulated fully informed opinion and
the simulated opinion of full-information and full-conversation (i.e.,
Column C – Column B). Thus, the conversation effect was intended to capture
the unique contribution of political conversation to opinion changes above
and beyond the effect of political knowledge.
Table 1 shows the magnitudes and directions of conversation effects (see
Column D and Column E). In terms of the size of the effects, that of
information effects normally exceeded that of conversation effects. But the
size of conversation effects was greater in 9 out of 28 opinion items. The
average opinion shifts were 7.2% for the information effects and 5.2% for
the conversation effects. This result suggests that, compared to the
information effects, the magnitude of the conversation effects is substantial.
The directions of the conversation effects were somewhat surprising. The
signs of the entries in the third and forth columns in Table 1 (i.e.,
Column D and Column E) show the directions of both the conversation effect
and the information effects. In Table 1, opinion items showing the
different directions of the information and the conversation effects have
bold-faced entries. The cases having the same direction of the two effects
were 14 out of 28 opinion items. In the other half of the cases, the
information effects and the conversation effects operated in the opposite
directions. As a result, the simulated opinion of full-information and
full-conversation were located in between the actual observed opinion and
the simulated fully informed opinion.
[Table 2 about here]
Table 2 summarizes the patterns of opinion changes in each topical issue
category (see also Table 4.2 in Althaus, 2003, for the directions of the
information effects in more extensive issue categories). First, in the
foreign issue category, similar to Althaus (1998, 2003), both the simulated
fully informed opinion and the simulated opinion of full information and
conversation shifted to the more interventionist direction relative to the
actual opinion. Whereas 72. 3% of respondents agreed that the U.S. would be
better off trying to solve some problems in other parts of the world, this
response increased to 78.1% in the simulated fully informed opinion and to
81.8% in the simulated opinion of full information and full conversation
(see opinion item 1 in Table 1).
Second, the operative issue category covers the President's and Congress's
approval ratings, party systems and party control of government, the scope
of government, and campaign finance reform. In approval questions, the
Clinton administration's and Congress's approval ratings dropped in both of
the sets of simulated opinion, except for the simulated opinion of full
information and full conversation on approval of Congress (see opinion
items 2 and 3). Congress ratings dropped, to a relatively large degree
(12.3%), from actual to fully informed opinion but rose slightly in the
opinion of full information and full conversation (4.7%). In the question
of party systems and party control of government, fully informed opinion,
relative to the actual opinion, showed a tendency of being less supportive
of a continuation of the two-party system of Democrats and Republicans and
more supportive of divided party control of the presidency and Congress
(see opinion items 5 and 6). In contrast, the simulated opinions of full
information and full conversation, relative to the fully informed opinion,
showed just the opposite tendency. In opinion items about the scope of
governments, whereas Althaus (1998) found that the fully informed opinion
is more opposed than actual opinion to the idea of "big government," the
current simulation showed the mixed pattern. On the one hand, the response
that "the less government the better" increased from 41.8% in the actual
opinion to 52.1% (see opinion item 8 in Table 1). Fully informed opinion is
also more likely to oppose placing new limits on imports (see opinion item
7). On the other hand, the response that "government is getting too strong"
also dropped from 69.3% to 64.3% in the simulated fully informed opinion
(see opinion item 4 in Table 1). Conversation effects were relatively small
in magnitude in this category of opinion questions. In the questions of
campaign finance reform, both of the sets of simulated opinions are more
likely to agree that "we should protect government from excessive influence
from campaign contributors" (see opinion item 9).
Third, in the fiscal issue category, both sets of simulated opinions showed
a similar pattern of opinion change, the only exception being the question
of defense spending. They are largely less supportive of increases in
government spending and less supportive of spending surpluses, both for tax
cuts and for Social Security and Medicare. However, fully informed opinion
showed a greater preference for increases in defense spending than
simulated opinion of full information and full conversation.
Finally, a variety of social issues reveal interesting patterns of
information and conversation effects. In terms of the magnitude of the
effects, conversation effects were particularly large in this issue
category. The average opinion changes were 6.1% for information effects and
6.4% for conversation effects. In the area of affirmative action, fully
informed opinion is generally less supportive both of the principles of
affirmative action (see opinion item 17) and of preferential treatment of
blacks in jobs (see opinion items 25 and 26). To the contrary, simulated
opinion of full information and full conversation showed just the opposite:
More support both of the principles of affirmative action and of
preferential treatment of blacks in jobs. This simulated opinion of full
information and full conversation also revealed more progressive attitudes
toward gay rights than did the fully informed opinion (61.9% versus 49.8%
in support of allowing gay couples to adopt children and 75.9% versus 64%
in support of laws to protect homosexuals against job discrimination). In
the policy area of education, the two sets of simulated opinion also showed
the opposite directions of opinion changes. Simulated opinion of full
information and full conversation was more supportive of school vouchers
and less supportive of school integration than was fully informed opinion
(see opinion items 20 and 22). Although the information effects and the
conversation effects operated in the same directions in the questions of
immigration and the death penalty, they were not unidirectional in the
questions concerning gun control, health insurance, crime, and a law making
English the official language.
Discussion
This paper, using the simulation modeling technique that was used by
Althaus (1998), Gilens (2001), and Sturgis (2003), calculated both
simulated fully informed opinions and the simulated opinion of full
information and full conversation. The results suggest that there exists a
substantial degree of "conversation effects" and that these conversation
effects do not merely supplement the information effects but often times
operate as countervailing forces on opinion changes. This tendency was more
evident in issues concerning the party system and party control and in a
variety of social issues such as affirmative action, gay rights, health
insurance, gun control, and education. Particularly interesting are the
patterns of conversation effects found in the social issue category.
Compared to information effects, the conversation effects found in this
issue category revealed not only greater magnitudes of opinion changes but
also more consistent patterns that were quite different from information
effects. Especially in such issues as affirmative action and gay rights,
the simulated opinion of the full information of full conversation showed
consistently more progressive policy preferences and more sympathetic
attitudes toward minority rights than did fully informed opinion.
This result calls into question the validity of the thesis mentioned at the
outset of this paper. The thesis suggests that opinion changes in
deliberative forums are comparable to information effects. However, this
study demonstrates that, even after controlling for the low levels and
uneven distribution of political knowledge, political conversation produced
a substantial degree of, and complex patterns of, opinion changes. This
suggests that any direct equating of the opinion changes that occurred in
the deliberative opinion polls with those caused by the "information
effects" could mask the actual effects of various factors that produce
opinion changes in the real deliberative forums.
Despite the existence of "conversation effects," it cannot simply be
concluded that the simulation modeling would be improved if political
conversation were included in the simulation models or that the simulated
opinion of full information and full conversation would be more
representative or more deliberative than the simulated fully informed
opinion. However interesting they are, the opinions simulated here are
hypothetical in nature. It is not worth debating which set of simulated
opinions is of a higher quality. In addition, the meanings of "fully
informed" or "full-conversation" should be taken with caution. As Althaus
(2003) stressed, these terms are meaningful only in relation to specific
indexes of general political knowledge tests or of political conversation.
These particular measures of knowledge and conversation may or may not be
related to what people really bring in when they think and talk about
specific policy issues. Indeed, there are some suggestions that standard
political knowledge tests or simple measures of political conversation may
not adequately capture people's own modes of reasoning and sense-making
(Gamson, 1992; Schlesinger & Lau, 2000).
Given the hypothetical nature and limited meanings of simulated opinions,
what the current comparison between the two simulation models reveals is
simply that political conversation is also unevenly distributed across
different social categories of people, that correcting for this uneven
distribution produces substantial changes in policy preferences, and that
these changes have certain patterns that are quite different from those
produced by information effects. To further examine whether or not
conversation effects are supplementary to information effects, it is
helpful to compare the patterns of opinion changes found in the current
study with those in Gilens (2001), which demonstrated a supplementary role
of policy-specific knowledge to general political knowledge. In Gilens'
findings, effects both of general knowledge and of policy-specific
knowledge work in different directions only in one out of 8 opinion
simulations. In the current study, information effects and conversation
effects produced different patterns of opinion changes in 14 out of 28
simulations. This difference suggests that conversation effects do not
merely supplement information effects.
Concerning the relationship between information effects and conversation
effects, one might wonder why the two effects quite often operate in
opposite directions, as revealed in the current analysis. Given the
reciprocal positive relationship between political knowledge and political
conversation, one can easily expect that the two effects would operate in
the same direction. The result was just the opposite. To further explore
the reason for which the two effects are counteractive, the estimated
coefficients of "political knowledge" and "conversation" in the baseline
models were examined. In terms of the main effects, the coefficients of
political knowledge and political conversation have the same valence sign
most of the time. This suggests that the counteractive nature of the two
effects may have been caused by the complex interrelationship between
political conversation and sets of demographic variables. The information
effects and conversation effects seem to operate in a way that is more
complicated than one might expect.
Several limitations of this study should be noted. First, the analyses were
conducted on a limited number of opinion items. To meaningfully generalize
the current findings, the further analysis of more cases in more extensive
issue categories is necessary. Second, there is still uncertainty
concerning the extent to which the simulation methods that were used in
this study reliably isolated conversation effects from information effects.
Given the complex relationship between political knowledge, political
conversation, and public opinion, a more sophisticated way of separating
conversation effects should employed. Third, there still remains more room
for improvement, especially in the measures of political conversation.
Various ways of forming an index should be tested for improvement. Finally,
this study did not conduct the patterns of opinion changes along various
subgroups of the sample. Given the complex patterns that emerged in the
current analysis, comparisons between various categories of subgroups would
be fruitful for our understanding of the underlying mechanism through which
conversation effects work.
References
Althaus, S. L. (1998). Information effects in collective preferences.
American Political Science Review, 92, 545-558.
Althaus, S. L. (2003). Collective preferences in democratic politics:
Opinion surveys and the will of the people. New York: Cambridge University
Press.
Barabas, J. (2002, April). How deliberation affects public opinion. Paper
presented at the Midwest Political Science Association's conference, Chicago.
Bartels, L. (1996). Uninformed votes: Information effects in presidential
elections. American Journal of Political Science, 40, 194-230.
Bennett, S. E., Flickinger, R. S., & Rhine, S. L. (2000). Political talk
over here, over there, over time. British Journal of Political Science, 30,
99-119.
Blumer, H. (1948). Public opinion and public opinion polling. American
Sociological Review, 13, 542-554.
Bohman, J. F. (1998). The coming of age of deliberative democracy. Journal
of Political Philosophy, 6, 399-423.
Bourdieu, P. (1972). Public opinion does not exist. In A. Mattelart & S.
Siegelaub (Eds.), Communication and class struggle, Part 1: Capitalism,
imperialism (pp. 124-130). New York: International General.
Conover, P. J., Searing, D. D., & Crewe, I. M. (2002). The deliberative
potential of political discussion. British Journal of Political Science,
32, 21-62.
Converse, P. E. (1964). The nature of belief system in mass publics. In D.
Apter (Ed.), Ideology and discontent (pp. 206-261). New York: Free Press.
Converse, P. E. (1987). Changing conceptions of public opinion in the
political process. Public Opinion Quarterly, 51, S12-S24.
Converse, P. E. (2000). Assessing the capacity of mass electorate. Annual
Review of Political Science, 3, 331-353.
Chambers, S. (2003). Deliberative democracy theory. Annual Review of
Political Science, 6, 307-326.
Delli Carpini, M. X., & Ketter, S. (1996). What Americans know about
politics and why it matters. New Haven, CT: Yale University Press.
Duch, R. M., Palmer, H. D., & Anderson, C. J. (2000). Heterogeneity in
perceptions of national economic conditions. American Journal of Political
Science, 44, 635-652.
Dewey (1939). Creative democracy: The task before us. Retrieved September
15, 2003, from http://www.beloit.edu/~pbk/dewey.html.
Dutwin, D. (2003). The character of deliberation: Equality, argument, and
the formation of public opinion. International Journal of Public Opinion
Research, 15(3), 239-264.
Entman, R. M., & Herbst, S. (2001). Reframing public opinion as we have
known it. In W.L. Bennett & R. M. Entman (Eds.), Mediated politics:
Communication in the future of democracy (pp. 203-225). New York: Cambridge
University Press.
Fishkin, J. S. (1991). Democracy and deliberation. New Haven: Yale
University Press.
Fishkin, J. S. (2000, November). Virtual democratic possibilities:
Prospects for Internet democracy. Paper presented at the conference on
"Internet, Democracy and Public Goods," Belo Horizonte, Brazil.
Fishkin, J. S., & Luskin, R. C. (2002). The quest for deliberative
democracy. In M. Saward (Ed.), Democratic innovations: Deliberation,
representation and association (pp. 17-28). New York: Routledge.
Gamson, W. A. (1992). Talking politics. New York: Cambridge University Press.
Gilens, M. (2001). Political ignorance and collective policy preferences.
American Political Science Review, 95, 379-396.
Habermas, J. (1972). The public sphere. In A. Mattelart & S. Siegelaub
(Eds.), Communication and class struggle, Part 1: Capitalism, imperialism
(pp. 198-201). New York: International General.
Habermas, J. (1989). The structural transformation of the public sphere: An
inquiry into a category of bourgeois society (T. Burger, Trans.).
Cambridge, MA: MIT Press.
Herbst, S. (1993). Numbered voices: How opinion polling has shaped American
politics. Chicago: University of Chicago Press.
Katz, E. (1992). On parenting a paradigm: Gabriel Tarde's agenda for
opinion and communication research. International Journal of Public Opinion
Research, 4, 80-85.
Katz, E. (1995). Introduction: The state of the art. In T. L. Glasser & C.
T. Salmon (Eds.), Public opinion and the communication of consent (pp.
xxi-xxxiv). New York: The Guilford Press.
Lippmann, W. (1922). Public opinion. New York: Free Press.
Kim, J. Wyatt, R., & Katz, R. (1999). News, talk, opinion, participation:
The part played by conversation in deliberative democracy. Political
Communication, 16, 361-385.
McLeod, J. M., Scheufele, D. A., Moy, P., Horowitz, E. M., Holbert, R. L.
Zhang, W., Zubric, S., & Zubric, J. (1999). Understanding deliberation: The
effects of discussion networks on participation in a public forum.
Communication Research, 26, 743-774.
Page, B., & Shapiro, R. (1992). The rational public: Fifty years of trends
in Americans' policy preferences. Chicago: University of Chicago Press.
Pan, Z., & Kosicki, G. M. (2001). Framing as a strategic action in public
deliberation. In S. D. Reese, O. H. Gandy & A. E. Grant (Eds.), Framing
public life: Perspectives on media and our understanding of the social
world (pp. 35-66). Mahwah, NJ: LEA.
Popkin, S. (1991). The reasoning voter. Chicago. University of Chicago Press.
Price, V. (1992). Communication concepts 4: Public opinion. New York: Sage.
Price, V. & Neijens, P. (1997). Opinion quality in public opinion research.
International Journal of Public Opinion Research, 9, 336-360.
Price, V. & Neijens, P. (1998). Deliberative polls: Toward improved
measures of "informed" public opinion? International Journal of Public
Opinion Research, 10, 145-176.
Schafer, J. L. (1997). Analysis of incomplete mulitivariate data. London:
Chapman and Hall
Schlesinger, M. & Lau, R. R. (2000). The meaning and measure of policy
metaphors. American Political Science Review, 94, 611-626.
Schudson, M. (1997). Why conversation is not the soul of democracy.
Critical Studies in Mass Communication. 14, 297-309.
Sniderman, P., Brody, R., & Tetlock, P. (1991). Reasoning and choices:
Exploration in political psychology. New York: Cambridge University Press.
Sturgis, P. (2003). Knowledge and collective preferences: A comparison of
two approaches to estimating the opinions of a better informed public.
Sociological Methods & Research, 31, 453-485.
Wyatt, R. O., Katz, E., & Kim, J. (2000). Bridging the spheres: Political
and personal conversation in public and private spaces. Journal of
Communication, 50, 71-92.
Wyatt, R. O., Kim, J., & Katz, E. (2000). How feeling free to talk affects
ordinary political conversation, purposeful argumentation, and civic
participation. Journalism & Mass Communication Quarterly, 77, 99-114.
Table 1
Effects of Political Knowledge and Political Conversation by Opinion Questions
Opinion Question
Var. No.
A: Actual Opinion
B: Full Information
C: Full Information and Full Conversation
D: Effects of Information (B–A)
E: Effect of Conversation (C–B)
Foreign Issue
1. Oppose isolationism
v514
72.3
78.1
81.8
5.8
3.7
Operative Issues
2. Clinton approval
3. Congress approval
4. Government not getting too strong
5. Favor the two-party system
6. Prefer divided government
7. Oppose import limits
8. More things government should do
9. Oppose campaign finance reform
v339
v356
v1513
v1650
v397
v512
v1420
v1489
66.0
59.6
30.7
38.0
52.4
52.0
58.2
24.4
62.4
47.3
35.7
36.1
62.2
57.6
47.9
19.5
61.1#
52.0
36.4*
41.3*
58.6**
60.3*
49.8#
17.7#
–3.6
–12.3
5.0
–1.9
9.8
5.6
–10.3
–4.9
–1.3
4.7
.7
5.2
–3.6
2.7
1.9
–1.8
Fiscal Issues
10. Increase government spending/services
11. Defense spending should be increased
12. Approve using surplus for tax cuts
13. Disapprove surplus for Social Security
v550
v587
v688
v691
39.0
39.5
63.2
17.8
37.1
44.9
45.7
39.9
31.3
43.6
37.7
48.6
–1.9
5.4
–17.5
22.1
–5.8
–1.3
–8.0
8.7
Social Issues
14. Immigrants should be decreased
15. Favor private health insurance
16. Black should help themselves
17. Oppose affirmative action
18. Allow homosexuals in the military
19. Favor more gun control
20. Favor school voucher program
21. Oppose English as official language
22. Oppose School integration
23. Oppose child adoption by gay couples
24. Oppose death penalty
25. Government should ensure blacks fair treatment in jobs (Percent opposed)
26. Against preferences for blacks in jobs
27. Against laws to protect homosexuals
28. Address social problem to reduce crime
v508
v614
v645
v674
v724
v728
v742
v745
v747
v748
v749
v800
v803
v1478
v1486a
44.5
30.9
43.8
45.3
75.7
58.8
52.7
25.0
46.1
55.0
26.6
44.6
81.0
32.6
48.2
26.3
35.8
41.0
53.1
78.0
49.9
55.0
22.7
55.7
50.2
36.0
43.9
82.4
36.0
60.9
22.6
31.9*
37.1*
47.7
79.9
52.7
52.4
24.3
49.2
38.1
46.6
26.1**
73.0
24.1**
58.7
–18.2
4.9
–2.8
7.8
2.3
–8.9
2.3
–2.3
9.6
–4.8
9.4
–.7
1.4
3.4
12.7
–3.7
–3.9
–3.9
–5.4
1.9
2.8
–2.6
1.6
–6.5
–12.1
10.6
–17.8
–9.4
–11.9
–2.2
Notes: Bold-faced entries denote the information effects and the
conversation effects operate in the same direction.
# p < .10, * p < .05, ** p < .01 in the likelihood ratio test between
Equation 1 and 2 (see p. ).
Table 2
Comparisons between Surveyed Opinion, Simulated Opinion of Full
Information, and Simulated Opinion of Full Information and Full Conversation
Opinion Topics
Full Informationa
Full Information and Full Conversationb
Foreign Issues
Isolationism
Operative Issues
Clinton approval
Congressional approval
Party system and party control
Scope of government
Campaign finance reform
Fiscal Issues
General
Surplus
Social Issues
Affirmative action
Gay rights
Education
Crime
Immigration
Health insurance
Gun control
English as official language
More interventionist
Less approving
Less approving
Less supportive of the two-party system and more supportive of divided
party control of the presidency and Congress
More agree that "government is not getting too strong," but favor less
government
More supportive
Less supportive of increase in government spending and services
Less supportive of using surplus both for tax cuts and for Social Security
and Medicare
Less supportive both of the principles of affirmative action and of
preferential treatment of blacks in jobs
Mixed
More supportive of school voucher and less supportive of school integration
more supportive of addressing the social problems that causes crime
Less agree that immigrants should be decreased
More supportive of private insurance
Favor less gun control
More supportive
More Interventionist
Less approving
More approving
More supportive of the two-party system and less supportive of divided
party control of the presidency and Congress
About the same
More supportive
About the same
Less supportive of using surplus both for tax cuts and for Social Security
and Medicare
Less supportive both of the principles of affirmative action and of
preferential treatment of blacks in jobs
More progressive
Less supportive of school voucher and less supportive of school integration
More supportive of punitive solution
Less agree that immigrants should be decreased
Less supportive of private insurance
Favor more gun control
Less supportive
Notes: a Comparisons was made to the actual surveyed opinion (i.e., Column
A and B in Table 1)
b Comparisons was made with the simulated fully informed opinion (i.e.,
Column B and C in Table 1).
Appendix
Measures of Political Information
The additive index of political knowledge include following items:
Post-interview interviewer rating (v1745); the offices held by Trent Lott
(v1447), William Rehnquist (v1450), Tony Blair (v1453), and Janet Reno
(v1456); the majority party in the House (v1356) and Senate (v1357); the
relative ideological positions of Republicans and Democrats (v1382, v1383),
Gore and Bush (v1372, v1374); and the relative policy position of Gore and
Bush on government spending (v562, v568), the parties on government
spending (v574, v580), Gore and Bush on defense spending (v592, v597), the
parties on defense spending (v602, v607), 3710, v3711), Gore and Bush on
job assurance (v625, v630), the parties on job assurance (v635, v640), Gore
and Bush on abortion(v696, v698), Gore and Bush on environmental regulation
(v783, v790), the change in budget deficit in Clinton presidency (v807, v1590)
Opinion Items
v397: Preference for divided government
Do you think it is better when one party controls both the presidency and
Congress, better when control is split between the Democrats and
Republicans, or doesn't it matter?
1. better when control is split
0. better when one party control both; it doesn't matter
v508: Increase/decrease immigration
Do you think the number of immigrants from foreign countries who are
permitted to come to the United States to live should be increased,
decreased, or left the same as it is now?
1. decreased 0. increased; let the same as it is now
v512: Favor/oppose import limits
Some people have suggested placing new limits on foreign imports in order
to protect American jobs. Others say that such limits would raise consumer
prices and hurt American exports. Do you favor or oppose placing new limits
on imports, or haven't you thought much about this?
1. oppose 0. favor
v514: Isolationism
Do you agree or disagree with this statement: This country would be better
off if we just stayed home and did not concern ourselves with problems in
other parts of the world.
1. disagree 0. agree
v550: Government services/spending
Where would you place yourself on this scale, or haven't you thought much
about this?
A combined 5-point scale from "reduce government spending and services a
great deal" to "increase government spending and services a great deal."
Values of 4 and 5 were coded as 1, and the others as 0.
v587: Defense spending
Do you feel the government should decrease defense spending, increase
defense spending, or is the government spending on defense about the right
amount now?
A combined 5-point scale from "decrease defense spending a great deal" to
"increase defense spending a great deal." Values of 4 and 5 were coded as
1, and the others as 0.
v614: Health insurance
Where would you place yourself on this scale, or haven't you thought much
about this?
A combined 5-point scale from "strongly-government insurance plan" to
"strongly-individual plan." Values of 4 and 5 were coded as 1, and the
others as 0.
v645: Aid to blacks
Which is closer to the way you feel, or haven't you thought much about
this? Should the government help blacks to a great extent or only to some
extent? Should blacks have to help themselves to a great extent or only to
some extent?
A combined 5-point scale from "Government help blacks to great extent" to
"Should help themselves to a great extent." Values of 4 and 5 were coded as
1, and the others as 0.
v674: Affirmative action
Some people think that if a company has a history of discriminating against
blacks when making hiring decisions, then they should be required to have
an affirmative action program that gives blacks preference in hiring. What
do you think? Should companies that have discriminated against blacks have
to have an affirmative action program?
1. No, they should not have to have affirmative action
0. Yes, they should have to have affirmative action
v688: Using surplus for tax cuts
Recently, there has been a lot of talk about how to spend the extra money
the federal government is likely to have in the near future. Some people
have proposed that most of the expected federal budget surplus should be
used to cut taxes. Do you approve or disapprove of this proposal?
1. Approve 0. Disapprove
v691 Using surplus for Social Security/Medicare
Some people have proposed that most of the expected federal budget surplus
should go to protecting social security and Medicare. Do you approve or
disapprove of this proposal?
1. Disapprove 0. Approve
v724: Homosexuals in the military
Do you think homosexuals should be allowed to serve in the United States
Armed Forces or don't you think so?
1. Homosexuals should allowed to serve
0. Homosexuals should not be allowed to serve
v728: Gun control
Do you think the federal government should make it more difficult for
people to buy a gun than it is now, make it easier for people to buy a gun,
or keep these rules about the same as they are now?
1. More difficult 0. Make it easier; keep these rules
about the same
v742: School vouchers
Do you favor or oppose a school voucher program that would allow parents to
use tax funds to send their children to the school of their choice,
even if it were a private school?
1. Favor school voucher program 0. Oppose school voucher program
v745: English official language
Do you favor a law making English the official language of the United
States, meaning government business would be conducted in English only, or
do you oppose such a law?
1. Oppose 0. Favor; Neither favor nor oppose
v747: School integration
Do you think the government in Washington should see to it that white and
black children go to the same schools or stay out of this area as it is not
the government's business?
1. Stay out of this area as it is not the government's business
0. See to it that white and black children go to the same schools; Other,
depends
v748: Child adoption by gay couples
Do you think gay or lesbian couples, in other words, homosexual couples,
should be legally permitted to adopt children?
1. No 0. Yes
v749: Death penalty
Do you favor or oppose the death penalty for persons convicted of murder?
1. Oppose 0. Favor
v800: Government should ensure Blacks equal treatment in jobs
How do you feel? Should the government in Washington see to it that black
people get fair treatment in jobs OR is this not the federal government's
business?
1. This is not the federal government's business
0. Government in Washington should see to it that black people get fair
treatment in jobs; other
v803 Preferences for blacks in jobs
Some people say that because of past discrimination, blacks should be given
preference in hiring and promotion. Others say that such preference in
hiring and promotion of blacks is wrong because it gives blacks advantages
they haven't earned. What about your opinion -- are you for or against
preferential hiring and promotion of blacks?
1. Against preferential hiring and promotion of blacks
0. For preferential hiring and promotion of blacks
v1420: Less government, or more things government should do
Next, I am going to ask you to choose which of two statements I read comes
closer to your own opinion. You might agree to some extent with both, but
we want to know which one is closer to your own views.
1. there are more things that government should be doing?
0. the less government, the better
v1478: Laws to protect against job discrimination of homosexuals
Recently there has been a lot of talk about job discrimination. Do you
favor or oppose laws to protect homosexuals against job discrimination?
1. Oppose 0. Favor
v1486a: Ways to reduce crime
Do you think that the best way to reduce crime is to address social
problems or to make sure criminals are caught, convicted, and punished, or
that we should do something in between, or haven't you thought much about this?
A combined 7-point scale from "Much better to address social problem" to
"Much better to punish criminals" Should help themselves to a great
extent." Values of 1, 2, and 3 were coded as 1, and the others as 0.
v1489: Campaign finance
Recently, there has been a lot of talk about campaign finance reform. Some
people believe that we should protect government from excessive influence
by campaign contributors. Others think that protecting the freedom of
individuals to financially support political candidates and parties is more
important. Which is closer to the way you think, or haven't you thought
much about this?
1. Protect government from excessive influence
0. Protect the freedom of individuals
v1513: Government is getting too powerful or not too strong
What is your feeling, do you think the government is getting too powerful
or do you think the government is not getting too strong?
1. Government not getting too strong
0. Government too powerful; Other
v1650 Two party system
Which of the following outcomes regarding political parties best represents
what you would like to see happen? You can just tell me the number of the
outcome you choose.
1. Continuation of the tow party system
0. Elections in which candidates run as individuals without party labels;
the growth of one or more parties that could challenge the Democrats and
Republicans
[1] Opinion quality means many things (Price & Neijens, 1997, 1998). Here,
"higher-quality opinion" refers to the measures of those opinions that are
better informed, more deliberative, and/or more representative. Specific
meanings of these features differ across these two approaches.
[2] The foreign issue category covers whether the U.S. should be involved
in solving problems around the world; the fiscal issues include such issues
as government spending and taxation; the operative issues include the size
and the scope of the federal government, government regulation of the
economy, and congressional and presidential approval; and the social issues
deal with such issues as affirmative action, the death penalty, and
minority rights (Althaus, 1998).
|
