[ Pobierz całość w formacie PDF ]ranking member of the staff in order to explain my project and what
I was requesting from each member. In many cases I was able to meet
with the member s chief of staff, and in most cases was able to meet
with someone ranking at least as high as legislative director.
Questionnaires were successfully delivered in person to the offices
of every member of the House between July 1 and July 8, 2004. The
final response rate was just under 10 percent.3 Clearly, with just 435
members in the House, this response rate makes statistical analysis
challenging. Because of the small number of respondents, the results
discussed below should only be viewed as suggestive, and no solid
conclusions can be reached. But as becomes clear below, interesting
and suggestive patterns nevertheless emerge.
The primary purpose of this survey is to compare the attitudes of
Republicans to the attitudes of Democrats. In order to make such com-
parisons, it is necessary to confirm that Republicans and Democrats
who responded to the survey do not differ systematically from their
co-partisans who did not respond. For example, if only very liberal
Democrats and very conservative Republicans responded to the sur-
vey, then partisan differences on distributional attitudes would likely
be magnified in the survey data. To assess my ability to make reason-
able cross-party comparisons with data from this survey, I examine
the correspondence of characteristics between those who responded
to the questionnaire and those who did not. Specifically, I compare
the ideological positions of respondents to nonrespondents, measured
with roll-call voting behavior.
These data, which are presented in Table 3.3, lead to the con-
clusion that potential response bias does not interfere substantially
3
The overall response rate appears to be very low but is not substantially lower than
some mail surveys of the general public. As an even better point of comparison, only
40 percent of members responded to a survey conducted in 1973 by a subcommittee in
the House itself the Intergovernmental Relations Subcommittee of the House Com-
mittee on Government Operations (Stenberg and Walker 1977). Democrats responded
at a higher rate than Republicans, which is likely due to the fact that the letter of intro-
duction was provided by a Democratic member. My hope was to provide a letter of
introduction from a member of both parties, but despite my best efforts could not
secure help from any Republican members.
Political Conflict over Who Gets What? 69
table 3.3. Comparison of Respondents and Full House
Characteristics
Characteristic Respondents Full 106th House
Percent Republican 32% 51%
Chamber DW-NOMINATE Average -0.15 0.05
Republican DW-NOMINATE Average 0.46 0.46
Democrat DW-NOMINATE Average -0.43 -0.39
with comparisons across party. Given that the goal of the analysis
discussed below is to observe correlations between distributional atti-
tudes and ideology and to compare partisan preferences on these issues,
the most important findings here are in the third and fourth rows
of the table. The ideology of Republican (and Democratic) respon-
dents closely mirrors the ideology of all Republican (and Democratic)
members when analyzed by party. Thus, conclusions about differences
between Republican and Democratic House members can likely be
made without fear of problems created by nonresponse bias.4 While
this is reassuring, the small sample size still makes statistically signif-
icant differences difficult to find, and all the results from this survey
should be viewed as illustrative, rather than definitive. In addition,
it is clearly not appropriate to generalize from this sample to the
entire House, since Democrats were much more likely to respond than
were Republicans. The only inferences I seek to make, however, are
comparisons across party.
Results of the House Survey
In discussing the results generated by the House survey data, I
focus on four potential sources of partisan disagreement regarding
distributional and redistributional policymaking. The first potential
4
It is unlikely with a response rate of less than 10 percent that the data analyzed here
have the properties of a random sample. Members of the House self-selected into
the sample. Democrats were more likely to respond than Republicans. However, it
appears that participation within party was random with respect to ideology. Moder-
ate Republicans were no more likely than conservative Republicans to participate in
the survey. The same is true of Democrats. Still, because of the low response rate, the
results discussed in this section should only be viewed as suggestive. However, what
they suggest turns out to be interesting, as the discussion below demonstrates.
70 The Politics of Income Inequality in the United States
source of disagreement is substantive preferences regarding how much
inequality should exist in society. It is possible that Democrats and
Republicans fundamentally disagree about how much income inequal-
ity should exist in the ideal society. The second potential source of
disagreement is over the level of priority that income inequality should
be given vis-á-vis other economic outcomes. The time of policymakers
and the resources of government are scarce, so even if Democrats and
Republicans agree about the level of inequality that should exist, pol-
icy differences will be more likely to occur if Democrats place a higher
priority on reducing income inequality than do Republicans. The third
potential source of disagreement is the theoretical role of government.
Policy disagreement regarding income inequality could simply be a
part of broader ideological disagreements about the appropriateness of
government intervention in the economy. The fourth potential source
of disagreement that I address is group-based. That is, Republicans
and Democrats might produce policies with divergent distributional
consequences because of the types of groups that each party views as
deserving of government aid.
Substantive Preferences Regarding Income Inequality
In order to gain traction on the question of whether there is partisan
disagreement about how much inequality should exist, in one section
[ Pobierz całość w formacie PDF ]zanotowane.pldoc.pisz.plpdf.pisz.plkajaszek.htw.pl
