Interest groups and presidential approval.
Cohen, Jeffrey E.
Job approval ratings are important to presidents and scholars of
the presidency for a number of reasons. First, presidents view approval
as a source of bargaining leverage with Congress (Edwards 2009): Recent
scholarship supports this contention (Cohen 2011). Approval, too, has
been used to predict the president and his party's prospects in the
upcoming elections (Holbrook 2010). Most important for this article are
the representational implications of approval. Approval can be thought
of as an interelection barometer of public support for the president and
his policies. Given the importance of presidential approval, presidents
spend a lot of time on activities geared at enhancing and/or preserving
their standing with the public, and scholars have invested considerable
energy into studying the sources of presidential approval.
For the most part, research has centered on the macro sources of
presidential approval. Commonly, studies look at temporal movement in
national approval, asking such questions as whether changes in economic
indicators or major presidential addresses to the nation affect the
level and trend in approval (Gronke and Newman 2003, 2009). Cohen (2008,
2010), however, argues that recent changes in the political system,
notably the rise in party polarization and the new media, has undermined
the effectiveness of a national "going public" strategy
(Kernell 2006) for building presidential support. Presidents now
supplement going national with a going narrow strategy, in which they
target their party base, certain localities, and interest groups. There
has been very little research on either the efficacy of the going narrow
strategy or the importance of interest groups to presidential support.
The lack of attention to interest groups for presidential approval is
also somewhat odd given their importance in politics and to the
presidency (Baumgartner and Leech 1998; Baumgarmer et al. 2011; Walker
1991).
Interest Groups and Presidential Approval
This article asks whether interest groups can affect their
members' evaluation of the president. Recent research suggest the
growing importance of interest group support to presidents, especially
as the parties have polarized and the media have fragmented, two trends
that limit the ability of presidents to generate support in the public
at-large (Cohen 2008, 2010; Holmes 2008). Party polarization reduces the
likelihood that opposition party members will support the president
(Bond and Fleisher 2001; Jacobson 2007; Newman and Siegle 2010). A
fragmented media reduces the size of the audience for presidential
communications, simultaneously forcing the president to compete with
other media voices for the public's attention (Baum and Kernell
1999; Young and Perkins 2005).
As a consequence, presidents have turned to other sources for
support, like interest groups. The literature on presidential-interest
groups ties, however, focuses primarily on institutional-level linkages,
for instance, the willingness of interest groups as organizations to
support the president, to coordinate their congressional lobbying
efforts with the president's, etc. (Loomis 2009; Peterson 1992;
Pika 1991, 1999, 2009). Although knowledge of the institutional linkages
between presidents and interest groups is scant, even less is known
about the factors associated with the ability of the president to
generate support from members of interest groups, or even whether
interest group membership is relevant to member opinions about the
president.
Scholars believe that interest groups may influence opinions on
issues: "Citizens' judgments about ... issues rely crucially
on the descriptions and rhetorical representations of political elites
and other information sources, including the media and interest
groups" (Joslyn and Haider-Markel 2002, 690). Several studies
suggest a linkage between group membership and approval of the
president.
For instance, Mueller's (1970) "coalition of
minorities" concept hints at a possible linkage between interest
groups and presidential approval. The coalition of minorities concept is
built upon spatial theory assumptions. Mueller argues that "the
possibility that an administration, even if it always acts with majority
support on each issue, can gradually alienate enough minorities to be
defeated. This could occur when the minority on each issue feels so
intensely about its loss that it is unable to be placated by
administration support on other policies it favors." (1970, 64).
Over the course of an administration, the number of presidential
decisions that alienate some "minority" will mount, leading
Mueller to hypothesize a linear decline in presidential approval.
Although Mueller does not formally define minority, most equate the term
with "issue publics" and/or "interest groups."
Mueller's coalition of minorities concept, however, is vague about
the specific conditions under which a presidential decision will
translate into support withdraw among these intense minorities. Does
this reaction happen on all issues or only some? Do formal interest
groups play any role in this process, for instance, by publicizing a
presidential decision to its members and either endorsing or criticizing
the president's decision? (1)
Claibourn and Martin (2007) look at the impact of voluntary group
membership on confidence in the executive, using a series of General
Social Survey studies. Using a social capital framework, they find that
presidential liberalism affects confidence in the executive for those
who belong to voluntary groups but has no effect on those who do not
belong to a voluntary group. However suggestive, Claibourn and
Martin's study is limited for our purposes. Although confidence in
the executive and presidential job approval are related, they are not
the same, which they acknowledge (Claibourn and Martin 2007, 195). (2)
Plus, their measure of group membership tallies the number of types of
groups to which a person belongs, not a particular group, the objective
of this article. Third, they do not distinguish among types of issues,
looking at general presidential liberalism rather than presidential
actions or positions on specific issues. Still, their results are
suggestive of a group membership-approval linkage.
Below I develop and test a theory of interest group influence on
member approval of the president. Interest groups affect member approval
of the president by linking the issue important to the group to
evaluations of the president. This linkage between issue and approval is
conditional on issue type, specifically whether the issue is
"easy" or "hard" (Carmines and Stimson 1980, 1989;
Cobb and Kuklinski 1997; Joslyn and Haider-Markel 2002; Leege et al.
2002). Group effects on member approval will be greater for
"hard" than "easy" issues. I test this with data
from the 2006 Cooperative Congressional Election Study (CCES), which
asked respondents about membership in 11 interest groups. Variance in
the type of groups and issues provides an opportunity to test this
conditional effects hypothesis. The next section presents the theory,
followed by a discussion of the data, and then the analysis, which
demonstrates support for the conditional interest group-issue effects
theory. The conclusions put the findings into perspective, speak to the
implications for presidential leadership, and suggest directions for
future research.
A Theory of Interest Groups Influence on Presidential Approval
The theory linking interest group membership to presidential
approval contains two elements: (1) special services that interest
groups provide to its members and (2) characteristics of the issue
around which the group is organized. Those special services, especially
interest group media, provide the foundation for possible group effects
on member opinion. The type of issue, easy versus hard, interacts with
this group environment, linking the issue of importance to the group to
the president. This linkage is likely to be stronger for harder as
opposed to easier issues.
Information Flow within Interest Groups
First, consider the special services that interest groups may
provide members. Of particular importance here are specialized media,
such as newsletters, magazines, and online materials that members
receive from the group. Two attributes of these specialized interest
group media stand out. One, they often cover the issue of concern to the
group and its members in greater breadth and depth than found in the
general news media. Two, this coverage blatantly takes an advocacy
posture, endorsing or criticizing the president's actions. The
general news media, in contrast, may feel professional and other
constraints to be more "objective" in their reporting.
Objectivity in journalism usually requires presenting both sides of the
issue. Thus, while the general news media may present voters with a
two-direction information flow, interest group media will present voters
with a one-direction information flow (Zaller 1992).
This information flow from interest groups to members may have both
priming and framing effects. For instance, the deeper coverage of an
issue in the specialized group media should raise the importance of the
issue to the group member, in effect priming that issue. The blatant
advocacy in these specialized media may aid members to interpret the
president's position with regard to the issue, a framing effect. In
the analysis below, we will not be able to distinguish between these two
mechanisms. Thus, we should see the issue of concern to the group
carrying greater weight in member approval of the president than for
nonmembers, but for hard as opposed to easy issues, the second major
element of the theory.
Issue Type and Interest Group Effects: Easier versus Harder
Type of issue will mediate the effect of group media on
member's approval of the president. Carmines and Stimson's
(1980) distinction between "easy" and "hard" issues
provides a good starting place for understanding how issue
characteristics mediate the effect of specialized group media on
members. Carmines and Stimson's conceptualization of an easy versus
hard issue refers to the inherent qualities of the issue. Since their
introduction of the concept, studies have offered alternative ways of
distinguishing easy versus hard issues.
To Carmines and Stimson, an easy issue evokes a "gut
reaction" from the voter; little effort or thought is required to
take a stance on it. Easy issues are more likely to be symbolic, have
been on the agenda for a long time, and deal more with policy ends than
means (1980, 80). In contrast, hard issues are more complex and
technical than easy ones, debate often centers on policy means and
implementation rather than ends, and the issue may have arrived recently
on the political agenda. Carmines and Stimson demonstrate that more
informed and politically sophisticated voters weigh hard issues more
heavily in their vote choice than less informed and sophisticated
voters. (3)
We are more likely to observe interest group effects on member
approval on hard as opposed to easy issues. The general news media are
likely to cover easy issues voluminously. In fact, high news coverage
levels help make an issue easier for voters. Hard issues may not receive
as much general news media attention. The technical and complex nature
of many hard issues renders them more difficult for the general news
media to provide in depth and detailed coverage. Although technical
debates may involve intense conflicts, such debates may be very dry or
abstract, attributes that news organizations, concerned about paying
subscribers, do not think will attract subscribers as much as the clash
of personalities or political gain, themes often important in news
coverage. Interest groups communications do not suffer these limitations
on issue coverage and may have incentives, primarily member interest and
the mobilization of members, to cover hard issues in greater depth than
found in the general news medial. By doing so, they provide members with
information not generally available from the general news media.
If the general news media cover an issue profusely and in depth,
which is likely the case for many easy issues, communications from the
interest group's specialized media may not add to the member's
store of information and may not "prime" or "frame"
the issue beyond the effects due to exposure to general news coverage.
If the issue is easy, the advocacy presented in the group's
specialized media may have little impact on member understanding of the
issue and/or the implications of the president's action. However,
when the issue is not covered much or in depth in the general news
media, generally the case for hard issues, interest group media provide
new information to its members, information that they would not
necessarily possess otherwise or would find very costly to collect and
process. And the issue advocacy found in the interest group media may be
especially helpful for the member in interpreting, making sense of, the
issue and the president's action on the issue. This is more likely
for hard than easy issues.
The Conditional Effects Hypothesis
Thus, interest group membership will affect presidential approval
for hard but not for easy issues. Further, for hard issues we should see
an interaction between member position on the issue and presidential
approval as the information from the interest group media tightens the
connection between the issue and the president, as expressed in this
equation:
Presidential Approval = a + [b.sub.1] (Group Membership) +
[b.sub.2] (Issue Position) + [b.sub.3] (Group Membership * Issue
Position)
For easy issues, once controlling for issue position, we expect
[b.sub.1] = 0 and [b.sub.3] = 0. But for harder issues, controlling for
issue position, we expect [b.sub.1] and [b.sub.3] to be statistically
significant.
Data
In addition to a question on presidential approval, the 2006 CCES
asked respondents about membership in 11 interest groups spanning the
political spectrum, from liberal groups like the Sierra Club and the
National Abortion Right Action League, to moderate groups like the
League of Women Voters, to more conservative groups, like the Christian
Coalition. The CCES used the following question to measure membership in
the 11 groups: "Are you a member of any of the following
organizations? Please check all that apply." It is unclear how many
respondents will misrepresent their group membership, either claiming
membership when they are not members of a group, or saying they are not
members when they are. Unlike the vote validation studies, which allow
us to check whether a person actually voted, we have no way to validate
group membership. Still the general size of groups elicited from these
data and the relationship to common variables suggests reasonable
accuracy in measuring group membership.
Table 1 presents some basic descriptive information on the
percentage of respondents claiming membership in each group and the
hypothesized impact of a particular group on presidential approval in
2006 based on the political orientation of the group. Those who answered
"not sure" to the group membership questions are excluded from
this and all subsequent analyses. The groups vary in size from the
massive American Automobile Association (AAA) (26.5%) and American
Association of Retired Persons (AARP) (19.3%) to the much smaller League
of Women Voters and Christian Coalition, with only about 1.5 % of the
sample as members. As an indicator of the validity of this measure, we
find a correlation between group membership with other variables for
which we would expect a relationship such as political activity and
interest. Table 1 also presents correlations between group membership
and a question on political interest. For all groups save the
Parent-Teacher Association (PTA), there is a statistically significant
correlation between individual group membership and political interest.
There is a strong negative correlation between political interest and
not claiming membership in any of the 11 groups.
Having membership questions about so many groups spanning the
political spectrum will allow us to see whether membership has the
directional effects on approval as hypothesized above. In testing the
interest group-approval linkage, it is important that members'
(dis)approval of the president is consistent with the relationship of
the group to the president. Thus, we expect membership in conservative
groups or those with positions that align with the Bush administration,
like the Christian Coalition, to lead to respondent approval of the
president. In contrast, we expect membership in liberal groups, like
National Abortion Rights Action League (NARAL), to lead to disapproval.
Finally, we do not expect membership to affect respondent approval of
President George W. Bush for groups like the League of Women Voters,
which are essentially nonpartisan or politically nonaligned.
Unfortunately, the CCES did not query respondents on the quality of
group membership or strength of association between the member and the
group, for instance, how active the member is in group activities, how
long the member has belonged to the group, why the member joined, nor
how closely the member identifies with the group and its political aims.
Such data may mediate the effects of group membership on political
opinions. Still, the variance in political orientations of the groups
provides an opportunity to test for group effects on member political
opinions.
Group Membership and Presidential Approval: A First Cut
The dependent variable is a four-point job approval question:
"Do you approve or disapprove of the way George W. Bush is handling
his job as President? Strongly Approve (4), Somewhat Approve (3),
Somewhat Disapprove (2), Strongly Disapprove (1)." More respondents
disapprove than approve: nearly 50% strongly disapprove, 9.6% somewhat
disapprove, 22% somewhat approve, and 18.4% strongly approve.
The present analysis follows this analytic strategy. First, I build
a baseline model, which incorporates variables known to influence
individual level approval of the president: party identification,
ideological self-placement, national economic assessments, race (Black,
Hispanic), gender, education, age and age-squared, (4) and unemployment
status, and church attendance (5) (Gilens 1988; Greene 2001). Then I add
the interest group membership variable to assess whether membership
affects approval beyond the baseline. (6) Since the dependent variable
is a four-point scale and is not normally distributed, I use ordered
probit. Overall, the baseline model performs quite well, with a
pseudo-[R.sup.2] of .41 (see Table 2). All variables but the Hispanic
and unemployment dummies are statistically significant and signed
correctly. However, at the zero-order, both Hispanic and unemployment
are correctly signed and significant. The sign reversals appear to be a
function of correlations with the other variables.
Next I enter the dummy variables for interest group membership,
simultaneously and one at a time, to the baseline model, for a total of
12 estimations. Multicollinearity appeared among some group membership
items (7) when trying to enter all 11 dummies simultaneously. The
utility of these interest group membership variables is that knowing the
political orientation of the group, we can make directional predictions
about the effect of membership in a particular group on presidential
approval. This is an advantage over the generic items typically used in
surveys that only ask respondents about membership in types of groups
(Baumgarmer and Walker 1988; Claibourn and Martin 2007). We hypothesize
that respondents who are members of the liberal groups--the Sierra Club,
NARAL, and AARP--will be more likely to disapprove of the president than
nonmembers. In contrast, we expect member of conservative groups, the
Veterans of Foreign Wars (VFW), American Legion, National Right to Life,
and the Christian Coalition, to be more likely to approve of Bush than
nonmembers. Finally, three groups--AAA, PTA, and League of Women
Voters--tend not to be as partisan or ideological as the others, and
thus we do not expect membership in them to affect attitudes toward the
president.
Table 3 presents the results. For space reasons, the table only
presents results of the coefficients for the interest group membership
dummies. Rarely does the addition of an interest group membership
variable affect the performance of the baseline variables. The effect of
interest groups in almost all cases conforms to expectations. For the
three liberal/opposition groups (AARP, Sierra Club, NARAL), we find that
membership is associated with disapproval, as hypothesized. The
coefficients for membership all are negative and statistically
significant, in both the saturated and separate estimations. As probit
coefficients are not intuitively interpretable, I convert them to
probabilities. Table 4 presents the probability effects based on the
separate estimations. Membership in the liberal groups tends to have the
strongest effect on the probability of strong disapproval. With controls
for the baseline variables, a member of AARP has a .06 higher
probability of strong disapproval than a non-AARP member. The
probabilities of strong disapproval for NARAL and the Sierra Club are
.26 and .19. Membership in liberal interest groups has pronounced
effects on presidential approval.
CCES asked about membership for five conservative groups: NRA, VFW,
American Legion, National Abortion Rights Action League, and the
Christian Coalition. In all cases, the membership coefficient has a
positive sign, but membership effects are statistically significant for
only two groups in the saturated model and three in the separate
estimations. Members of NRA are more likely to approve of the president
than nonmembers for both the saturated and separate estimations.
Although the probability analysis does not indicate that NRA membership
will increase strong approval of the president, such membership dampens
strong disapproval by about .05. (8) As for veterans groups, membership
in either group becomes significant in the separate estimations. The
high correlation between membership in VFW and the American Legion (r =
.34, p = .000) accounts for their weak effects in the full estimation.
Membership in either group lessens the probability of strong
disapproval, .05 for the American Legion and .04 for the VFW.
Given the strong association of Bush with antiabortion policies, it
is somewhat of a surprise that membership in the two antiabortion
groups, National Right to Life and the Christian Coalition, are not
significant in the saturated or separate models at conventional .05
levels. Members of both groups are likely to attend church frequently
compared to other respondents, one of the baseline variables. (9) When
church attendance is dropped from the estimation membership in National
Right to Life or the Christian Coalition becomes significant. (10)
Finally, we did not expect membership in the nonpartisan,
nonaligned interest groups to affect approval. This is what we find for
the PTA and League of Women Voters, but somewhat surprisingly, members
of AAA are less likely to approve of Bush than nonmembers. For both the
saturated and separate estimations, the coefficients for AAA are
negative and statistically significant. The probability analysis
suggests a .05 higher probability of strong disapproval, similar in
magnitude to AARP, the veterans groups, or the Christian Coalition. It
is not clear why this significant effect of AAA membership. AAA members
are more likely to be liberal and identify as Democrats compared to
respondents overall, but we have already controlled for these variables.
Issue Type, Group Membership, and Presidential Approval
The above analysis demonstrates that membership in specific
interest groups sometimes is associated with presidential approval, but
as that analysis did not control for respondent issue position, those
results may be spurious. This section tests for the possibility of
spurious effects as well as the conditional interest group effects
hypothesis. To test the conditional hypothesis requires classifying
issues as easy or hard. There are four issues associated with the
aligned groups in this analysis: abortion, the Iraq War, the
environment, and reform of Social Security. The first two, abortion and
Iraq, are classic easy issues, while the environment and Social Security
reform, are harder issues for voters.
In Carmines and Stimson (1980), three attributes differentiate easy
from hard issues. Easy issues are primarily symbolic, while hard issues
have a more technical or complex element. Second, easy issues are mostly
about policy ends, where hard issues are debates on policy means.
Finally, easy issues tend to have been on the agenda for a long time
compared to harder issues. Rather than the strict dichotomy between easy
and hard, it may be useful to think of issues as ranging from easier
versus harder for voters. All issues contain some mix of these three
attributes.
Abortion may be the easiest of the four issues dealt with here. It
has been on the agenda since 1973, when the Supreme Court issued the Roe
decision. Since then it has been a major national issue, dividing voters
and the two parties. Abortion is often debated in moral terms, although
more recent related debates, such as on stem cell research and
regulation of abortion access, invoke some technical issues into the
debate.
Although the Iraq War has not been on the agenda as long as
abortion, it is also a relatively easy issue for voters. For much of the
Bush administration, the Iraq War has been the primary issue on the
agenda, receiving massive news coverage and attention from political
leaders. Partisan division on the war also provides voters with
relatively clear cues on how to think about the war. For many voters,
the war, thus, is an issue that defines a major division between the
parties. Admittedly there are technical aspects to the war, such as the
debate over the surge policy, but the partisan attributes, which make
the issue relatively easy for voters, overwhelms these technical
aspects.
On the one hand, the environment might appear an easy
issue--everyone wants a clean environment. Too, the environment has been
on the nation's agenda, at least since the early 1970s, if not
earlier. But this issue can be thought of as harder for voters than
abortion or the war for several reasons. While everyone wants a clean
environment, the crux of the debate on environmental issues is how to
produce a clean environment, how much and what types of regulation are
necessary. This instills a technical aspect to the debate, where
economics and environmental science often are employed by advocates in
making their case. And even though environmental issues have been on the
national agenda for decades, in general that issue does not receive much
media attention unless an environmental catastrophe occurs, like the
Gulf Sea oil spill. In contrast, the Kyoto Accords, important to the
environmental debate, received much less news coverage, and voters in
general were unaware of the accords, the Bush administration on them, or
how they relate to the environmental debate. These attributes make the
environment a harder issue than abortion or the Iraq War.
Seemingly, the popularity and visibility of Social Security would
render it an easy issue, and in some regards like the environment, it
has easy aspects. But the specter of future insolvency, which seems to
have arisen periodically over the past quarter century, injects an
element of difficulty into this issue. Joslyn and Haider-Markel (2002)
classify Social Security reform as a hard issue. For instance, it may
not be clear to voters how a particular reform proposal relates to the
insolvency issues or to the benefits that they would receive. George W.
Bush's initiative in 2005 to privatize Social Security, and his
arguments that his proposal would strengthen Social Security and average
voters' retirements, is a case in point. Again economics,
demographic projections, and other fields, often arcane to voters, were
used to justify both support and opposition to his proposal.
Thus, while it may be difficult to argue that any one issue is
absolutely easy or hard, we can rank issues as being easier or harder
for voters. Here, abortion and the Iraq War are viewed as easier, with
the environment and Social Security reform as harder. For the easy
issues, we do not expect membership in the relevant interest groups, the
three abortion and two veterans groups, to affect member approval of the
president. In contrast, for the harder issues, we expect membership in
the environmental (Sierra Club) and senior group (AARP) to affect
approval by tightening the relationship between opinion on the hard
issue and approval. In other words, there should be an interaction
effect between group membership and relevant issue positions.
Two Easier Issues: Abortion and the Iraq War
The 2006 CCES asked respondents the following question on abortion
policy: "There has been some discussion about abortion during
recent years. Which one of the opinions on this page best agrees with
your view on this issue? By law, abortion should never be permitted (1),
The law should permit abortion only in case of rape, incest or when the
woman's life is in danger (2), The law should permit abortion for
reasons other than rape, incest, or danger to the woman's life, but
only after the need for the abortion has been clearly established (3),
By law, a woman should always be able to obtain an abortion as a matter
of personal choice (4)."
To test the conditional effects hypothesis, I interact this
variable with membership in one of the three abortion groups, NARAL,
National Right to Life, and the Christian Coalition. I then enter this
interaction term, along with the group membership variable and the
abortion opinion variable, into the baseline equation. We expect that
opinion on abortion to affect approval, but neither the group membership
nor interaction variables should be significant (see Table 5).
As expected, abortion opinion strongly colors evaluations of Bush,
but neither the group membership nor the interaction terms reach
statistical significance for any of the abortion groups. Also, if the
interaction term is deleted, group membership falls short of statistical
significance for National Right to Life and the Christian Coalition
dummies, although group membership in NARAL is statistically significant
(b = -.65, SE =. 16, p < .000). Probability analysis indicates that
NARAL membership increases the likelihood of strong disapproval of Bush
by .25 controlling for abortion attitudes. NARAL membership acts on
member attitudes toward Bush independent of opinion on abortion, but
membership does not appear to strengthen the impact of abortion
attitudes on approval of the president.
Turning to veterans group membership and the Iraq War, I use the
following question: "Do you think it was a mistake to invade Iraq?
Yes (1), not sure (2), No (3)." Again, the analysis includes the
group membership and the Iraq War variables, as well their interaction.
Like for abortion, we expect only attitudes on the war to be
statistically significant (see Table 6). For the VFW, membership alone
increases support for the president, while war opinion also affects
presidential approval, but the interaction term points in the wrong
direction, with a .053 significance level. With such a large n, we may
not want to make much of the .053, which nearly attains the conventional
.05 significance level. However, we can speculate about the reasons
underlying the wrong sign, assuming a possibly significant affect here.
With a nearly even split in opinion among VFW members on the war, the
VFW may not have to provide members with much official group
communication on the war. Perhaps too, war opponents within the
membership held their opinions more strongly and were more willing to
voice their war opposition to other members. Through face-to-face
contact, these "anti-war opinion leaders" may have been able
to influence the opinion of other members. (11) Unlike VFW, neither
membership in the American Legion nor the interaction of membership with
war opinion affects member approval of the president. The lack of
American Legion effects may be due to the less political nature of the
group and its membership compared to the VFW.
Two Harder Issues: Social Security Reform and the Environment/Jobs
Trade-Off
The 2006 CCES asked respondents about their attitudes toward social
security privatization reform. Importantly, the question does not
reference President Bush, thereby mitigating any contamination effects
between Social Security opinion and attitudes toward Bush:
Now, we'd like to ask you about Social Security. A proposal has
been made that would allow people to put a portion of their Social
Security payroll taxes into personal retirement accounts that would
be invested in private stocks and bonds. Do you favor or oppose
this idea? Strongly favor (1), Somewhat favor (2), Neither favor
nor oppose it (3), Somewhat oppose (4), Strongly oppose (5).
Again, I enter AARP membership, issue position, and their
interaction into the baseline model from Table 2. Results on Table 7
strongly support the conditional effects hypothesis. The interaction
coefficient is negative and statistically significant. With the
interaction term, AARP membership now falls to insignificance, while
attitudes toward the reform are significant. Each one-step shift on the
five-point Social Security reform scale affects the probability of
strongly disapproving of Bush by .09. Those who strongly oppose the
reform are .36 more likely to strongly disapprove of Bush than those who
strongly favor the reform. The interaction term indicates that each
one-step shift on the reform scale for members of AARP leads to a .03
probably effect on strong disapproval. For members, those who strongly
oppose the reform are .12 more likely to strongly disapprove than those
who strongly favor the reform.
Figure 1 plots the complete interactive effect, taking into account
the three component variables, membership, reform opinions, and their
interaction. The figure clearly reveals the interactive effect of
membership and opinion on the reform. As respondents become opposed to
the reform, not only does the probability of strong disapproval rise,
but the gap between AARP members and nonmembers widens. Membership in
AARP tightens the linkage between Social Security opinion and
presidential approval consistent with the conditional effects hypothesis
for a hard issue.
CCES asked respondents a question about the trade-off between
environmental concerns and jobs, a useful question for our purposes
since it raises the often technical aspects of the debate on the
environment. The question does not reference Bush or any specific policy
decision.
Some people think it is important to protect the environment even
if it costs some jobs or otherwise reduces our standard of living.
Other people think that protecting the environment is not as
important as maintaining jobs and our standard of living. Which is
closer to the way you feel, or haven't you thought much about this?
Much more important to protect environment even if lose jobs and
lower standard of living (4), Environment somewhat more important
(3), About the same (2), Economy somewhat more important (1), Much
more important to protect jobs, even if environment worse (0).
Results that include the Sierra Club membership variable, opinion
on the environment/jobs trade-off, and their interaction are presented
on Table 8. As for the AARP, we see interaction effects between
membership in the Sierra Club and the environment-jobs trade-off on
presidential approval. The interaction is statistically significant and
negatively signed, as hypothesized, but the membership variable is no
longer statistically significant, and holds the wrong sign, while
attitude on the environment-jobs question is strongly significant and
negative. Each one-step shift on the five-point environment-jobs
question is associated with a .05 probability shift in strong
disapproval. Those who support environmental protection at the cost of
jobs are .20 more likely to disapprove than those who think jobs are
more important than the environment. For Sierra Club members, the
probability analysis indicates a .44 shift in probability of strong
disapproval with each one-step shift on the environment-job trade-off
question. Sierra Club members who support environmental protection at
the cost of jobs will be .33 more likely to disapprove of Bush than
those who think jobs are more important than the environment. Figure 2
diagrams the dramatic impact of the interaction effect on strong
disapproval.
[FIGURE 1 OMITTED]
Stem Cell Research: A Harder Abortion-related Issue?
The above finding of NARAL membership influences on approval with
controls for abortion opinion counters our expectation. We hypothesized
that for an easy issue, like abortion, once controlling for issue
position, group membership would no longer affect approval. As alluded
to above, aspects of the abortion debate have turned more technical and
complex in recent years, making them harder for some voters. Stem cell
research is one case in point. In contrast to early arguments in flavor
of abortion rights, which focused on the woman's right to choose,
the stem cell debate focuses on the scientific benefits to be gained.
Stem cell research is not entirely a "rights" issue, but it is
tied into attitudes toward science and thus may be more complex for some
voters. Nisbet (2004, 2005) shows that attitudes toward science affect
support/opposition to stem cell research. Some even raised the economic
consequences of restricting stem cell research in the wake of George W.
Bush's 2001 stem cell decision (Karch 2009). Information about stem
cell research appears to affect attitudes toward the issue (Nisbet
2005). Here NARAL may be especially important in providing information
to its members. In contrast, to those already opposed to abortion on
religious grounds, stem cell research may not be a hard issue.
The CCES asked respondents the following stem cell question, which
suits our purposes quite well: "Now we'd like to ask you about
whether the federal government should fund stem cell research. Some in
Congress argue that this research may lead to cures for diseases and
disabilities affecting large numbers of Americans, and should be funded.
Others argue that a potential human life has to be destroyed in order to
use these cells, and funding it would be unethical. What do you think?
If you were faced with this decision, would you vote for or against
federal funds for this research?"
As before, I create interaction terms between membership for the
three abortion groups and opinion on stem cell research. The analysis
keeps the abortion opinion variable as well as group membership, stem
cell opinion, and the interaction of stem cell opinion and group
membership. Since the stem cell issue may be hard for NARAL members, we
expect to see a significant effect for the interaction term, but as stem
cell research may not be as hard for National Right to Life or Christian
Coalition members, the interaction term should fall short of statistical
significance. Results are presented on Table 9.
[FIGURE 2 OMITTED]
As predicted, there is a strong interaction effect between NARAL
membership and the stem cell issue. NARAL members who favor stem cell
research, the NARAL position, are .44 more likely to strongly disapprove
of Bush than non-NARAL members who also support stem cell research.
Figure 3 diagrams the dramatic interactive effects of NARAL membership
and stem cell opinion on strong disapproval of Bush. Results show, as
predicted, no interaction effect of National Right to Life membership
and stem cell attitudes, nor does membership in the group affect
approval. For members of the Christian Coalition, the interaction term
appears significant but points in the wrong direction. And with this
interaction, membership in the Christian Coalition becomes significant.
But these supposed effects are statistical artifacts, a function of the
high collinearity between the interaction term and membership--85% of
Christian Coalition members opposed stem cell research.
[FIGURE 3 OMITTED]
Conclusions
Although primarily interested in the effects of interest groups on
presidential approval, the article also unearthed a direct effect of
issue position on approval in all cases looked at. In all cases dealt
with in this article, respondents' issue position affected their
evaluation of the president even controlling for respondent ideological
self-placement (Gronke 1999; Newman 2003). The direct impact of issue
position on approval in an age of party and ideological polarization is
all the more impressive. With so many important issues aligning with
party and/or ideological divisions, one would not expect issue position
to exert an independent direct effect on presidential approval. In part,
this may reflect the visibility, importance, and perhaps presidential
attention, of the issues studied here in recent national politics.
The direct effect of issue position on approval is important for
another reason. It demonstrates that approval is not solely a function
of valence issues, but also of positional issues, which has implications
for presidential representation and accountability. Presidents, based on
these findings, are held accountable for the policy direction that they
steer as well as whether their chosen policies work well or not. In
selecting policies to pursue, presidents thus have to consider not only
how well the policy works, but who will be receptive to differing policy
approaches. Some issues, in particular positional issues, may not have
uniform impacts on voters--some will win or benefit from the course the
president selects, while others will lose.
In thinking about the course to select on positional issues,
presidents may enlist and target some interest groups. Especially in an
era of polarized parties and fragmented media, presidents may face extra
difficulty in building support coalitions. In this political climate,
interest groups become increasingly important for presidential coalition
building efforts. With this in mind, the second question raised in the
article was whether interest groups can influence member attitudes about
the president.
Using the 2006 CCES, I found that membership in an interest group
that was aligned with the parties or ideology was associated with
approval of the president. However, we should not expect all groups to
affect members' political opinions. The information that interest
groups provide to members will have greater impact on hard as opposed to
easy issues. Comparing group membership for issues characterized as
easier (abortion, the Iraq War) versus harder (Social Reform, the
environment), this is what the analysis found. When an issue is hard,
interest group membership tightens the linkage between that issue and
presidential approval.
However, analysis also detected one finding seemingly counter to
the conditional effects of interest group membership on approval. For
the easy abortion issue, membership in NARAL was associated with
presidential approval controlling for abortion opinion, where the theory
elaborated here argued that for easy issues, controls for opinion on the
issue should erase any group effect on opinion. The abortion issue
provided us with a wonderful test case, as some abortion-related
debates, for instance, stem cell research, may be harder for some
voters. Returning to NARAL, I found that membership tightened the
relationship between stem cell opinion and approval. This last test
provided us with a strong test of the theory of conditional interest
group effects on members. Here we were able to hold group membership
constant but allow the easy or hardness of the issue to vary.
This study points to future research directions. First, here we
only possessed simple membership data. It would be useful to collect
data on the quality of membership in the group, such as identification
with the group, credibility of the group for political issues, exposure
to group publications, and participation in the group, all factors that
may moderate the effect of group membership on member opinion. Such data
should also be supplemented with content analyses of group
communications to members. And demonstrating membership effects on
member approval, we may want to address other ways in which interest
groups may affect their members such as on other opinions. As for
positional issues with regard to approval, future studies should aim to
look at a broader array of issues than just the visible and important
ones studied here. These suggestions total to a large and ambitious
future research program on interest groups, positional issues and
presidential approval, and the interest group-approval connection.
References
Baum, Matthew A., and Samuel Kernell. 1999. "Has Cable Ended
the Golden Age of Presidential Television?" American Political
Science Review 93 (March): 99-114.
Baumgartner, Frank R., and Beth L. Leech. 1998. Basic Interests:
The Importance of Groups in Politics and in Political Science.
Princeton, NJ: Princeton University Press.
Baumgartner, Frank R., Heather A. Larsen-Price, Beth L. Leech, and
Paul Rutledge. 2011. "Congressional and Presidential Effects on the
Demand for Lobbying." Political Research Quarterly 64 (March):
3-16.
Baumgartner, Frank R., and Jack L. Walker. 1988. "Survey
Research and Membership in Voluntary Associations." American
Journal of Political Science 32 (November): 908-28.
Bond, Jon R., and Richard Fleisher. 2001. "The Polls:
Partisanship and Presidential Performance Evaluations."
Presidential Studies Quarterly 31 (September): 529-40.
Carmines, Edward G., and James A. Stimson. 1980. "The Two
Faces of Issue Voting." American Political Science Review 74
(March): 78-91.
Carmines, Edward G., and James A. Stimson. 1989. Issue Evolution:
Race and the Trans-formation of American Politics. Princeton, NJ:
Princeton University Press.
Claibourn, Michelle P., and Paul S. Martin. 2007. "The Third
Face of Social Capital: How Membership in Voluntary Associations
Improves Policy Accountability." Political Research Quarterly 60
(June): 192-201.
Cobb, Michael D., and James H. Kuklinski. 1997. "Changing
Minds: Political Arguments and Political Persuasion." American
Journal of Political Science 41 (January): 88-121.
Cohen, Jeffrey E. 2008. The Presidency in an Era of 24-Hour News.
Princeton, NJ: Princeton University Press.
--. 2010. Going Local: Presidential Leadership in the
Post-Broadcast Age. New York: Cambridge University Press.
--. 2011. "Whose Approval Matters? Reelection, Constituency
Approval and Senate Support for President George W. Bush." Congress
& the Presidency 38 (3): 253-70.
Edwards, George C., III. 2009. "Leading the Public." In
Oxford Handbook of the American Presidency, ed. George C. Edwards, III,
and William G. Howell. New York: Oxford University Press, 182-207.
Gilens, Martin. 1988. "Gender and Support for Reagan: A
Comprehensive Model of Presidential Approval." American Journal of
Political Science 32 (February): 19-49.
Greene, Steven. 2001. "The Role of Character Assessments in
Presidential Approval." American Politics Research 29 (March):
196-210.
Gronke, Paul. 1999. "Policies, Prototypes, and Presidential
Approval." Paper delivered at the American Political Science
Association, Washington, DC.
Gronke, Paul, and Brian Newman. 2003. "FDR to Clinton, Mueller
to?: A Field Essay on Presidential Approval." Political Research
Quarterly 55 (December): 501-12.
--. 2009. "Public Evaluations of Presidents." In The
Oxford Handbook of the American Presidency, ed. George C. Edwards, III,
and William G. Howell. New York: Oxford University Press, 232-53.
Holbrook, Thomas. 2010. "Forecasting U. S. Presidential
Elections." In The Oxford Handbook of American Elections and
Political Behavior, ed. Jan E. Leighley. New York: Oxford University
Press, 346-74.
Holmes, Lisa M. 2008. "Why 'Go Public'? Presidential
Use of Nominees to the U. S. Courts of Appeals." Presidential
Studies Quarterly 38 (January): 110-22.
Jacobson, Gary C. 2007. A Divider, Not a Uniter: George W. Bush and
the American People. New York: Longman.
Joslyn, Mark R., and Donald E Haider-Markel. 2002. "Framing
Effects on Personal Opinion and Perception of Public Opinion: The Cases
of Physician-Assisted Suicide and Social Security." Social Science
Quarterly 83 (3): 690-706.
Karch, Andrew. 2009. "Vertical Diffusion and the Policymaking
Process: The Politics of Embryonic Stem Cell Research." Paper
presented at the 2009 Annual Conference on State Politics and Public
Policy, Chapel Hill, NC.
Kernell, Sam. 2006. Going Public: New Strategies of Presidential
Leadership. 4th ed. Washington, DC: CQ Press.
Leege, David C., Kenneth D. Wald, Brian S. Krueger, and Paul D.
Mueller. 2002. The Politics of Cultural Differences: Social Change and
Voter Mobilization Strategies in the Post-New Deal Period. Princeton,
NJ: Princeton University Press.
Loomis, Burdette A. 2009. "Connecting Interest Groups to the
Presidency." In The Oxford Handbook of the American Presidency, ed.
George C. Edwards, III, and William G. Howell. New York: Oxford
University Press, 403-24.
Mueller, John E. 1970. "Presidential Popularity from Truman to
Johnson." American Political Science Review 64 (March): 18-34.
Newman, Brian. 2003. "Integrity and Presidential Approval,
1980-2000." Public Opinion Quarterly 67 (Fall): 335-67.
Newman, Brian, and Emerson Siegle. 2010. "The Polarized
Presidency: Depth and Breadth of Public Partisanship." Presidential
Studies Quarterly 40 (June): 342-63.
Nisbet, Matthew C. 2004. "Trends: Public Opinion about Stem
Cell Research and Human Cloning." Public Opinion Quarterly 68
(Spring): 131-54.
--. 2005. "The Competition for Worldviews: Values,
Information, and Public Support for Stem Cell Research."
International Journal of Public Opinion Research 17 (Spring): 90-112.
Peterson, Mark A. 1992. "The Presidency and Organized
Interests: White House Patterns of Interest Group Liaison."
American Political Science Review 86 (September): 612-25.
Pika, Joseph A. 1991. "Opening Doors for Kindred Souls: The
White House Office of Public Liaison." In Interest Group Politics,
3rd ed., ed. Allan J. Cigler and Burdett A. Loomis. Washington, DC: CQ
Press, 277-98.
--. 1999. "Interest Groups: A Doubly-Dynamic
Relationship." In Presidential Policymaking: An End of the Century
Assessment, ed. Steven Shull. Armonk, NY: M. E. Sharpe, 59-78.
--. 2009. "The White House Office of Public Liaison."
Presidential Studies Quarterly 39 (September): 549-73.
Walker, Jack L., Jr. 1991. Mobilizing Interest Groups in America.
Ann Arbor: University of Michigan Press.
Young, Garry, and William B. Perkins. 2005. "Presidential
Rhetoric, the Public Agenda, and the End of Presidential
Television's 'Golden Age.' Journal of Politics 67
(November): 1190-205.
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New
York: Cambridge University Press.
(1.) In reviewing the presidential approval literature, Gronke and
Newman (2003, 2009) encourage more research on issue position effects on
approval, which has received much little attention.
(2.) The General Social Survey (GSS) question on executive
confidence, which they use, does not mention president. The GSS question
reads: "I am going to name some institutions in this country. As
far as the people running these institutions are concerned, would you
say you have a great deal of confidence, only some confidence, or hardly
any confidence at all in them? The Executive branch of the federal
government?" Some respondents might equate "those running the
executive branch" as top-level cabinet officials, like department
secretaries. The GSS question is vague on this point. In contrast, the
typical job approval question not only references the president by name,
but forces the respondent to think about job performance.
(3.) There are alternative ways of thinking about easy versus hard.
Cobb and Kuklinski (1997) argue that style of argumentation, easy versus
hard arguments, may better discriminate issues than their inherent
qualities. Their results are ambiguous with regard to the effect of hard
versus easy arguments on opinion, finding instead that "con"
arguments were more persuasive than "pro" arguments. Leege et
al. (2002) critique the common assumption that cultural and moral issues
are inherently easy due to their content. They contend that the style of
reasoning that the voter employs affects the "easiness" or
"hardness" of an issue. Cultural issues often invoke social
values and group differences, which makes them easier.
(4.) The linear and quadratic age terms pick up the curvilinear
relationship between age and approval, where younger and older people
are more likely to disapprove than middle age voters.
(5.) Church attendance taps into the impact of religious
orientations on approval. The 2006 CCES included several other religious
variables, but they either resulted in a large loss of cases or did not
add to the performance of model estimation.
(6.) All analyses employ case weights as recommended by the CCES
study directors.
(7.) Specifically, membership in National Abortion Right to Life
and the Christian Coalition.
(8.) The probability analysis suggests that group membership
generally has more impact on strong disapproval than strong approval,
perhaps because strong disapproval is the modal category.
(9.) The correlation are .12 (p < .000) for the Christian
Coalition and .14 (p < .000) National Right to Life.
(10.) The coefficient for National Right to Life becomes b =. 13,
SE = .06, t = 2.10, p = 0.04 and b =.19, SE = .08, t = 2.42, p = 0.02
for the Christian Coalition.
(11.) Claibourn and Martin (2007) suggest that interpersonal
interactions in voluntary groups may affect confidence about the
president. For such a path to operate, the group must provide
opportunities for members' personal interactions, something that
not all groups provide, or that if these opportunities exist, members
attend group activities. In an age of "check book" membership,
large percentages of members fail to attend such group activities.
JEFFREY E. COHEN
Fordham University
Jeffrey E. Cohen. professor of political science at Fordham
University, is the author of several books, including Presidential
Responsiveness and Public Policy-Making, as well as articles in numerous
journals, including the American Political Science Review, the American
Journal of Political Science, and the Journal of Politics.
TABLE 1
Interest Group Membership and Hypothesized Effects on Presidential
Approval
Member
Group Abbreviation % *
American Automobile Association AAA 26.5
American Association of Retired Persons AARP 19.3
National Rifle Association NRA 9.1
League of Women Voters LWV 1.4
Parent-Teacher Association or
Organization PTA 6.2
Veterans of Foreign Wars VFW 4.3
American Legion AL 5.1
Sierra Club SC 4.0
National Abortion Rights Action League NARAL 2.4
National Right to Life NR 2.2
Christian Coalition of America CCA 1.5
No Group Membership 49.2
Correlation
with
Political Hypothesized
Interest Effect on
Group r (p) ** Approval
American Automobile Association .10 (.000) 0
American Association of Retired Persons .08 (.000) -
National Rifle Association .10 (.000) +
League of Women Voters .06 (.000) 0
Parent-Teacher Association or
Organization .01 (.15) 0
Veterans of Foreign Wars .05 (.000) +
American Legion .05 (.000) +
Sierra Club .08 (.000) -
National Abortion Rights Action League .07 (.000) -
National Right to Life .05 (.000) +
Christian Coalition of America .04 (.000) +
No Group Membership -.16 (.000)
Source: 2006 CCES.
* Question: "Are you a member of any of the following organizations?
Please check all that apply".
** Question: How interested are you in politics and current affairs?
Very much interested (2), Somewhat interested (1), Not much
interested (0).
n = 29,877.
TABLE 2
Ordered Probit Baseline Estimation of Presidential Approval
b SE t p
Party ID 0.28 0.01 42.10 0.000
National Economy -0.63 0.01 -60.74 0.000
Black -0.07 0.04 -1.89 0.059
Hispanic 0.12 (a) 0.03 3.77 0.000
Gender 0.35 0.02 19.18 0.000
Education -0.12 0.01 -17.85 0.000
Age 0.02 0.003 6.94 0.000
Age (2) -0.0002 0.00003 -7.41 0.000
Ideology 0.01 0.001 27.15 0.000
Unemployed 0.17 (a) 0.06 2.93 0.003
Church Attendance 0.07 0.01 10.01 0.000
cur1 0.58 0.09
cut2 1.25 0.09
cut3 2.71 0.10
Pseudo-[R.sup.2] .41
Source: 2006 CCES.
n = 31,490 (a) wrong sign.
TABLE 3
Impact of Interest Group Membership on Presidential Approval
Saturated Model (a)
Group b (SE) p
American Automobile Association -.11 (.02) 0.000
American Association of Retired Persons -.14 (.03) 0.000
National Rifle Association .14 (.03) 0.000
League of Women Voters -.08 (13) 0.54
Parent-Teacher Association .04 (.04) 0.24
Veterans of Foreign Wars .10 (.05) 0.04
American Legion .07 (.04) 0.13
Sierra Club -.44 (.09) 0.000
National Abortion Rights Action League -.59 (.16) 0.000
National Right to Life .05 (.07) 0.41
Christian Coalition of America .12 (.08) 0.17
Separate Estimations (a)
Group b (SE) P
American Automobile Association -.12 (.02) 0.000
American Association of Retired Persons -.15 (.03) 0.000
National Rifle Association .13 (03) 0.000
League of Women Voters -.19 (.13) 0.14
Parent-Teacher Association .04 (.04) 0.33
Veterans of Foreign Wars .14 (.05) 0.004
American Legion .11 (.04) 0.01
Sierra Club -.49 (09) 0.000
National Abortion Rights Action League -.68 (.16) 0.000
National Right to Life .08 (.06) 0.17
Christian Coalition of America .15 (.08) 0.06
Source: 2006 CCES.
(a) Estimations control for all variables from the baseline model
from Table 2. The saturated model enters all interest group
membership variables simultaneously. The separate estimations
include the interest group membership variables one at a time.
n = 26,080.
TABLE 4
Probability Impact of Interest Group Membership on Presidential
Approval Presidential Approval Category
Strongly Somewhat
Group Disapprove Disapprove
American Automobile Association .05 -.01
American Association of Retired Persons .06 -.01
National Rifle Association -.05 .00
League of Women Voters .07 -.01
Parent-Teacher Association -.01 .00
Veterans of Foreign Wars -.05 .00
American Legion -.04 .00
Sierra Club .19 -.04
National Abortion Rights Action League .26 -.07
National Right to Life -.03 .00
Christian Coalition of America -.06 .00
Somewhat Strongly
Group Approve Approve
American Automobile Association -.03 -.01
American Association of Retired Persons -.04 -.01
National Rifle Association .04 .01
League of Women Voters -.05 -.01
Parent-Teacher Association .01 .00
Veterans of Foreign Wars .04 .01
American Legion .03 .01
Sierra Club -.13 -.02
National Abortion Rights Action League -.17 -.02
National Right to Life .02 .01
Christian Coalition of America .05 .01
Source: 2006 CCES.
Based on results from separate estimations, Table 3.
TABLE 5
Impact of Membership in Abortion Groups, Attitudes toward Abortion,
and Their Interaction on Presidential Approval
NARAL
b SE z p Impact
Party ID 0.27 0.01 34.89 0.00 -.11
Nat. Economy -0.63 0.01 -51.08 0.00 .25
Black -0.07 0.04 -1.58 0.11 .03
Hispanic 0.06 0.04 1.56 0.12 -.02
Gender 0.35 0.02 16.66 0.00 -.14
Education -0.10 0.01 -13.41 0.00 .04
Age 0.02 0.00 5.59 0.00 -.01
Age2 0.00 0.00 -6.00 0.00 .00
Ideology 0.01 0.00 18.6 0.00 -.005
Unemployed 0.15 0.07 2.22 0.03 -.06
Church Attend. 0.02 0.01 3.08 0.00 -.01
Group 1.12 1.14 0.99 0.33 -.34
Abortion -0.17 0.01 -13.25 0.00 .07
Interaction -0.47 0.30 -1.55 0.12 .18
cur1 -0.14 0.12
cut2 0.54 0.12
cut3 2.02 0.12
Pseudo-[R.sup.2] .42
N 24303
National Right to Life
b SE z p Impact
Party ID 0.27 0.01 34.99 0.00 -.11
Nat. Economy -0.63 0.01 -51.25 0.00 .25
Black -0.06 0.04 -1.38 0.17 .03
Hispanic 0.06 0.04 1.58 0.11 -.02
Gender 0.34 0.02 16.53 0.00 -.14
Education -0.10 0.01 -13.6 0.00 .04
Age 0.02 0.00 5.72 0.00 -.01
Age2 0.00 0.00 -6.16 0.00 .00
Ideology 0.01 0.00 18.86 0.00 -.005
Unemployed 0.16 0.07 2.26 0.02 -.06
Church Attend. 0.03 0.01 3.10 0.00 -.01
Group -0.01 0.16 -0.06 0.96 .003
Abortion -0.17 0.01 -13.33 0.00 .07
Interaction 0.01 0.09 0.16 0.87 -.01
cur1 -0.13 0.12
cut2 0.54 0.12
cut3 2.02 0.12
Pseudo-[R.sup.2] .42
N 24303
Christian Coalition
b SE z p Impact
Party ID 0.27 0.01 35 0.00 -.11
Nat. Economy -0.63 0.01 -51.22 0.00 .25
Black -0.06 0.04 -1.41 0.16 .03
Hispanic 0.06 0.04 1.54 0.13 -.02
Gender 0.34 0.02 16.56 0.00 -.14
Education -0.10 0.01 -13.55 0.00 .04
Age 0.02 0.00 5.71 0.00 -.01
Age2 0.00 0.00 -6.16 0.00 .00
Ideology 0.01 0.00 18.83 0.00 -.005
Unemployed 0.16 0.07 2.26 0.02 -.06
Church Attend. 0.02 0.01 3.00 0.00 -.01
Group 0.04 0.21 0.19 0.85 -.02
Abortion -0.17 0.01 -13.38 0.00 .07
Interaction 0.05 0.12 0.40 0.69 -.02
cur1 -0.13 0.12
cut2 0.54 0.12
cut3 2.02 0.12
Pseudo-[R.sup.2] .42
N 24303
Source: 2006 CCES.
TABLE 6
Impact of Membership in Veterans Groups, Attitudes toward the Iraq
War, and Their Interaction on Presidential Approval
VFW
b SE z p Impact *
Party ID 0.20 0.01 24.9 0.00 -0.07
Nat. Economy -0.48 0.01 -37.95 0.00 0.18
Black 0.05 0.04 1.20 0.23 -0.02
Hispanic 0.10 0.04 2.54 0.01 -0.04
Gender 0.27 0.02 13.00 0.00 -0.10
Educ. -0.09 0.01 -10.96 0.00 0.03
Age 0.01 0.004 2.42 0.02 0.004
Age2 -0.0001 0.00004 -2.32 0.02 0.00
Ideology 0.01 0.001 15.33 0.00 -0.004
Unemployed 0.13 0.07 1.80 0.07 -0.05
Church Attend. 0.06 0.01 7.74 0.00 -0.02
Group 0.22 0.10 2.16 0.03 0.08
Iraq War -0.90 0.02 -52.83 0.00 -0.34
Interaction -0.10 0.05 -1.93 0.05 0.04
cut1 -1.93 0.12
cut2 -1.04 0.12
cut3 0.72 0.12
Pseudo-[R.sup.2] .49
N 26060
American Legion
b SE z p Impact *
Party ID 0.20 0.01 24.88 0.00 -0.07
Nat. Economy -0.48 0.01 -37.96 0.00 0.18
Black 0.05 0.04 1.2 0.23 -0.02
Hispanic 0.10 0.04 2.54 0.01 -0.04
Gender 0.27 0.02 13.07 0.00 -0.10
Educ. -0.08 0.01 -10.94 0.00 0.03
Age 0.01 0.004 2.44 0.01 0.004
Age2 -0.0001 0.00004 -2.37 0.02 0.00
Ideology 0.01 0.001 15.28 0.00 -.004
Unemployed 0.13 0.07 1.78 0.08 -0.05
Church Attend. 0.06 0.01 7.73 0.00 -0.02
Group 0.17 0.09 1.76 0.08 -0.06
Iraq War -0.90 0.02 -52.86 0.00 0.34
Interaction -0.05 0.05 -1.09 0.27 0.02
cut1 -1.93 0.12
cut2 -1.05 0.12
cut3 0.72 0.12
Pseudo-[R.sup.2] .49
N 26060
Source: 2006 CCES.
* Probability impact on strong disapproval of Bush.
TABLE 7
Impact of AARP Membership, Attitudes toward Social Security
Privatization, and Their Interaction on Presidential Approval
Impact on
Robust Strong
b SE z P Disapproval
Party ID 0.26 0.01 33.53 0.000 -.10
Nat. Economy -0.55 0.01 -45.32 0.000 .22
Black -0.09 0.04 -2.18 0.03 .03
Hispanic 0.08 0.04 2.14 0.032 -.03
Gender 0.33 0.02 16.17 0.000 -.13
Education -0.12 0.01 -16.57 0.000 .05
Age 0.02 0.004 4.96 0.000 -.01
Age2 -0.0002 0.00004 -4.11 0.000 .00
Ideology 0.01 0.0006 19.31 0.000 -.005
Unemployed 0.18 0.07 2.66 0.008 -.07
Church Attend. 0.07 0.01 9.05 0.000 -.03
AARP 0.10 0.06 1.66 0.096 -.04
Social Security
Privatization -0.24 0.01 -25.48 0.000 .09
Interaction -0.07 0.02 -3.55 0.000 .03
cur1 -0.13 0.11
cut2 0.58 0.11
cut3 2.09 0.11
Source: 2006 CCES.
n = 26063, Pseudo-[R.sup.2] = .43.
TABLE 8
Impact of Sierra Club Membership, Attitudes on Environment-Jobs
Trade-off, and Their Interaction on Presidential Approval
Impact on
Robust Strong
b SE z p Disapproval *
Party ID 0.27 0.01 34.71 0.000 -0.11
Nat. Economy -0.62 0.01 -51.42 0.000 0.24
Black -0.12 0.04 -2.84 0.005 0.05
Hispanic 0.09 0.04 2.35 0.02 -0.04
Gender 0.32 0.02 15.75 0.000 -0.13
Education -0.11 0.01 -14.5 0.000 0.04
Age 0.02 0.004 4.76 0.000 -0.007
Age2 -0.0002 3.76E-05 -5.28 0.000 0.0001
Ideology 0.01 0.001 20.66 0.000 -0.005
Unemployed 0.21 0.07 2.94 0.003 -0.08
Church Attend. 0.06 0.01 8.23 0.000 -0.02
Sierra Club 0.49 0.28 1.73 0.084 -0.18
Environment/Jobs -0.14 0.01 -14.41 0.000 0.05
Interaction -0.28 0.09 -2.99 0.003 0.11
cur1 0.05 0.11
cut2 0.73 0.11
cut3 2.23 0.11
Source: 2006 CCES.
* Probability on strongly disapprove, n = 25277,
Pseudo-[R.sup.2] = .42.
TABLE 9
Impact of Abortion Group Membership, Attitudes on Stem Cell
Research, and Their Interaction on Presidential Approval
NARAL
b SE z p Impact
Party ID 0.27 0.01 31.75 0.00 -.11
Nat. Economy -0.61 0.01 -45.93 0.00 .25
Black -0.06 0.05 -1.24 0.22 .03
Hispanic 0.06 0.04 1.4 0.16 -.02
Gender 0.33 0.02 14.83 0.00 -.14
Education -0.1 0.01 -11.71 0.00 .04
Age 0.02 0 4.97 0.00 -.01
Age2 0 0 -5.19 0.00 .00
Ideology 0.01 0.001 15.35 0.00 -.005
Unemployed 0.2 0.08 2.62 0.01 -.06
Church Attend. 0.014 0.01 1.53 0.13 -.01
Group 0.7 0.79 0.89 0.38 -.25
Abortion -0.14 0.01 -9.61 0.00 .05
Stem Cell -0.27 0.03 -9.25 0.00 .11
Interaction * -1.37 0.81 -1.69 0.09 .43
cut1 -0.26 0.13
cut2 0.4 0.13
cut3 1.88 0.13
Pseudo-[R.sup.2] 0.43
N 22265
Christian Coalition
b SE z p Impact
Party ID 0.28 0.01 33.43 0.00 -.11
Nat. Economy -0.61 0.01 -48.35 0.00 .25
Black -0.06 0.04 -1.26 0.21 .03
Hispanic 0.08 0.04 1.91 0.06 -.02
Gender 0.32 0.02 14.72 0.00 -.14
Education -0.11 0.01 -13.8 0.00 .04
Age 0.02 0.004 4.63 0.00 -.01
Age2 0.00 0.00 -5.01 0.00 .00
Ideology 0.01 0.001 18.53 0.00 -.005
Unemployed 0.24 0.07 3.2 0.00 -.06
Church Attend. 0.04 0.01 4.52 0.00 -.01
Group 0.71 0.21 3.29 0.00 -.25
Abortion --
Stem Cell 0.38 0.03 14.05 0.00 -.14
Interaction * -0.68 0.23 -2.93 0.00 .26
cut1 0.5 0.11
cut2 1.16 0.11
cut3 2.64 0.11
Pseudo-[R.sup.2] 0.43
N 23829
National Right to Life
b SE z p Impact
Party ID 0.28 0.01 33.45 0.00 -.11
Nat. Economy -0.61 0.01 -48.38 0.00 .25
Black -0.05 0.04 -1.21 0.23 .03
Hispanic 0.08 0.04 1.93 0.05 -.02
Gender 0.32 0.02 14.68 0.00 -.14
Education -0.11 0.01 -13.82 0.00 .04
Age 0.02 0.004 4.65 0.00 -.01
Age2 0.00 0.00 -5.02 0.00 .00
Ideology 0.01 0.001 18.53 0.00 -.005
Unemployed 0.24 0.07 3.19 0.00 -.06
Church Attend. 0.04 0.01 4.68 0.00 -.01
Group 0.21 0.18 1.15 0.25 -.08
Abortion --
Stem Cell 0.38 0.03 13.82 0.00 -.14
Interaction * -0.2 0.19 -1.01 0.31 .08
cut1 0.51 0.11
cut2 1.16 0.11
cut3 2.64 0.11
Pseudo-[R.sup.2] 0.43
N 23829
Source: 2006 CCES.
* For the NARAL-Stem Cell Interaction, I code stem cell support 1,
stem cell opposition 0.
For the National Right to Life and Christian Coalition-Stem Cell
Interactions, I code stem cell support 0, stem cell opposition 1.