Feldman PoliticalCultureAmbivalence 1992

The Political Culture of Ambivalence: Ideological Responses to the Welfare State
Author(s): Stanley Feldman and John Zaller

Source: American Journal of Political Science, Vol. 36, No. 1 (Feb., 1992), pp. 268-307
Published by: Midwest Political Science Association
Stable URL: https://www.jstor.org/stable/2111433
Accessed: 18-12-2024 03:52 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Midwest Political Science Association is collaborating with JSTOR to digitize, preserve and
extend access to American Journal of Political Science
This content downloaded from 172.96.140.89 on Wed, 18 Dec 2024 03:52:06 UTC
All use subject to https://about.jstor.org/terms
The Political Culture of Ambivalence: Ideological
Responses to the Welfare State*
Stanley Feldman, State University of New York at Stony Brook

John Zaller, University of California, Los Angeles
This paper explores the principles that people draw upon to justify their support for social
welfare policy in the United States. The data for this study were produced by open-ended questions
asked of a representative sample of the U.S. public. The results show that most people readily use
values and principles central to the political culture when discussing their policy preferences. The
wide diffusion of diverse values-individualism, humanitarianism, and opposition to big govern-
ment-leads to significant ambivalence in people's discussions of their issue positions. The impli-
cations of these patterns of belief for popular support of the welfare state are discussed.
Writing just after the crest of the postwar liberalism symbolized by the
Great Society, Free and Cantril (1968) observed an apparent contradiction in the
political beliefs of many Americans. A large proportion of the public enthusias-
tically supported the specific federal programs that constitute the modem Ameri-
can welfare state. However, a similarly large number of people also endorsed a
series of strong statements that condemned big government and praised eco-
nomic individualism. Free and Cantril described this as a "schizoid combination
of operational liberalism with ideological conservatism" (1968, 37). While in
practice people accept the role of the federal government in maintaining social
welfare, "the abstract ideas they tend to hold about the nature and functioning
of our socioeconomic system still seem to stem more from the underlying as-
sumptions of a laissez-faire philosophy than from the operating assumptions of
the New Deal, the Fair Deal, the New Frontier, or the Great Society" (1968, 30).
Other observers of U.S. society have detected similar sorts of conflicts but
*An earlier version of this paper was presented at the 1988 meeting of the Midwest Political
Science Association, Chicago. We would like to thank Michael Margolis and Mark Peffley for their
helpful comments. The data used in this paper were collected by the National Election Studies. The
analysis and interpretations reported here are the responsibility of the authors only. Collecting these
data made unusually heavy demands on the staff of the Institute for Social Research at the University
of Michigan, where the study was conducted. In this connection we are grateful to ZoAnne Black-
burn, Steve Pinney, and Santa Traugott, who supervised the study, and to the following members of
the interview and coding departments: Heather Andrews, Valerie Austin, Linda Bailiff, Sylvia Baker,
Lisa Bonet, Jim Buckmaster, Laurel Burke, Liese Busse, Sujean Choi, Anmarie Currier, Jileen
Deviries, Jane Fisher, Doris Guthrie, Stan Hasper, Lori Johnson, Kurt Harju, Ann Haven, Lynn
Hayes, Suzanne Hogan, Rachael Howe, Kelley Ladd, Yvonne Kaplan, Hedy Malki, Anna Moyer,
Ann Munster, Judith Peebles, Louie Pettit, Ramon Randolph, Karen Smith, Kathryn Terrazas, Jeff
Toliver, and Marielle Weindorf.
American Journal of Political Science, Vol. 36, No. 1, February 1992, Pp. 268-307
? 1992 by the University of Texas Press, P.O. Box 7819, Austin, TX 78713
RESPONSES TO THE WELFARE STATE 269
see them as an enduring feature of U.S. political ideology rather than a short-
term failure to adapt to changing circumstances. This conflict is usually seen as
a clash between two major elements of the political culture: achievement and
equality (Lipset 1979), capitalism and democracy (McClosky and Zaller 1984),
or freedom and equality (Rokeach 1973). Rather than making an ideologically
clean choice between these competing values, people tend to accept both of
them, emphasizing one over the other in specific situations, but never wholly
rejecting either.
These claims about American values are important because they purport to
describe a fundamental property of public opinion-one that, it is asserted,
greatly affects the course of public policy in the United States. For example,
Free and Cantril argue that Americans need to adapt their ideological convictions
to the changing role of government in order to "implement their political desires
in a more intelligent, direct, and consistent manner" (1968, 181). But such
claims are highly problematic in that they presume a level of ideological aware-
ness of the public that is, by many accounts, unrealistic. In fact, it has become
almost a commonplace in the political behavior literature that Americans are
ideologically innocent.' One can readily imagine, therefore, that many of the
people who endorse welfare state programs while proclaiming allegiance to lais-
sez-faire ideals are completely unaware of any tension between them. What ana-
lysts such as Free and Cantril call value "schizophrenia" may be just another
instance of ideology-free thought.
Our article is an attempt to resolve this difficulty. Contrary to the evidence
of "ideological innocence," we argue that Americans do, for the most part,
understand the philosophical underpinnings of the policies they endorse, and
that, much more often than the belief systems literature would lead one to expect,
Americans make use of cultural values and principles in explicating and justify-
ing their political preferences. Further, we show that, exactly as the Free and
Cantril analysis would suggest, it is liberals rather than conservatives who are
most beset by value conflict over social welfare because they are the ones who
must somehow reconcile activist government with traditional principles of eco-
nomic individualism and laissez-faire.
This paper thus has two separate but closely related aims. The first is to
confirm the oft-made claim that popular support for the welfare state in the
United States must continually struggle with the values of nineteenth-century
liberalism. What is novel about our confirmation is that it is the first to be based
on systematic coding of data obtained from open-ended probes of a large, na-
tionally representative sample. This gives us an evidentiary base for our claims
that is significantly stronger than that which has been available to previous re-
searchers in this area. The second aim of the paper is to clarify and propose a
'See Kinder and Sears (1985) for a recent review of the evidence.
270 Stanley Feldman and John Zaller
resolution to the Converse-Lane controversy over political ideology. We shall not

dispute Converse's claim that most people are relatively nonideological in that
they normally fail to organize tightly their symbolic concerns in accord with
liberalism or conservatism. In fact, our data will help show why this is true, at
least for social welfare issues. But in a more revisionist vein, we shall show that
Americans do understand and use cultural values and principles in evaluating
and articulating many of their political preferences. These two claims do not
contradict each other; rather they reflect a long-standing feature of the political
culture.
The Prevalence of Ideology in the United States
In Converse's (1964) analysis of mass belief systems, the central idea is that
of constraint, namely, the capacity of one political idea to control or "constrain"
another. Thus, if people both embrace a general principle, such as economic
individualism, and derive specific policies from this value, their more specific
ideas are constrained by ideological principle. If, on the other hand, people hold
general values that are manifestly inconsistent with most of their concrete policy
preferences, the preferences are said to be unconstrained by ideological prin-
ciple. Converse's conclusion, of course, was that most individuals have uncon-
strained belief systems in the sense that their attitudes are not organized by the
left-right continuum or any other sort of dimensional continuum. Inconsistency
in policy preferences and political beliefs is thus a characteristic of a lack of
ideological structure.
Several researchers have examined other types of evidence and have come
to an entirely different conclusion. In her in-depth interviews with 28 adults from
New Haven, Hochschild (1981) describes in detail the "ambivalence" that emer-
ges from her subjects' attempts to apply traditional norms of economic differen-
tiation to modern problems of social welfare. Many of her respondents seemed
at least somewhat aware of the underlying conflicts, although "they find it easier
to live with, and to try to ignore, even distressing normative tensions than to
undertake the enormous effort needed to resolve them" (1981, 258). Similarly,
Reinarman's (1987) interviews with six public sector and six private sector work-
ers are full of statements that alternately reflect belief in the free market and
individualism, recognition that the system is not completely open and that people
need assistance, and criticisms of the bureaucratization and performance of the
federal government. In both of these studies, conflict and ambivalence is inter-
preted not as confusion, inconsistency, or lack of sophistication but as a problem
of reconciling the multiple values, beliefs, and principles simultaneously present
in the political culture.
Value conflict and ambivalence of this sort are not easily identified with the
types of fixed choice questions typically used in mass survey instruments. And
worse, when they are identified, they may easily be taken as evidence of lack of
constraint-and hence lack of concern for abstract principles. Yet, as both
RESPONSES TO THE WELFARE STATE 27I
Hochschild and Reinarman show, it is dangerous to equate lack of consistency

with lack of understanding of the principles that underlie policy preferences (see
also Lane 1973).
The first step toward untangling these issues is to specify as clearly as pos-
sible the relationship between the content and structure of the political culture
and the individual-level organization of political preferences and beliefs.
By political culture we mean a set of values that are widely endorsed by
politicians, educators, and other opinion leaders and that animate the principal
political institutions of a society. In the United States, these values include free-
dom, equality, individualism, democracy, capitalism, and several others. This
distinctive constellation of values originated in radical British politics, diffused
to North America in the colonial period, inspired the organizers of the American
Revolution (Baylin 1967), and diffused widely enough by the 1830s to be clearly
observable at the mass level by the time of Tocqueville's visit to the United
States. More recent survey-based studies confirm that most Americans continue
to embrace the core values of this political culture (Devine 1972; Lipset 1979;
McClosky and Zaller 1984).
Yet many analysts contend that U. S. political culture embodies a substantial
amount of unresolved value conflict, especially between freedom and equality
(Lipset 1979; McClosky and Zaller 1984). How do people respond to this
conflict?
Even the most cursory examination of U.S. politics reveals that many of the
elite sources of political culture, especially partisan politicians, are more enthu-
siastic about some aspects of the culture than others. Thus, some tend to favor
freedom, especially individual economic freedom, over other values, especially
equality and popular sovereignty. These people are widely considered, and usu-
ally consider themselves, to be conservatives. Others manifest the reverse pref-
erence ordering and are considered liberals. These recognized liberal and con-
servative opinion leaders, by the policies they favor and the pronouncements they
make, effectively create a set of ideological conventions for organizing conflict
among the values of the political culture. Members of the public who attend
closely to politics become aware of these organizing conventions and tend to
internalize the one that is most congenial to their own value predispositions.
People who, on the other hand, devote little attention to politics remain relatively
innocent of liberal-conservative ideology.2
Ideology and culture are thus closely related. In fact, liberal-conservative
2There have been a number of discussions of the individual-level roots of ideology and ideo-
logical conflict. Space prevents us from discussing this literature in any depth. Hypotheses have
pointed variously to orientations toward change and group evaluations (Conover and Feldman 1981),
feelings of sympathy and humanitarianism (Conover 1988), priorities attached to the values of free-
dom and equality (Rokeach 1973), and orientations toward social order, social change, and social
benevolence (McClosky and Zaller 1984). These individual-level predispositions may then orient
people toward one or the other of the packages of cultural values: liberalism or conservatism.
ideology may be considered a response to conflicts within the culture. Faced with
tension between achievement and equality (Lipset 1979), between laissez-faire
principles and pressures for a welfare state (Free and Cantril 1968), between
freedom and equality (Rokeach 1973), or between capitalism and democracy
(McClosky and Zaller 1984), conservatives tend to opt for the former value set
and liberals for the latter. As McClosky and Zaller write, "Such ideological
conflict as exists in America is confined within a broad framework of almost
universal public support for the basic values of capitalism and democracy....
[But when] asked to decide between preserving a laissez-faire economy and en-
acting measures that promise greater social and economic equality, conservatives
emphasize capitalistic values while liberals emphasize democratic values. Al-
though both liberals and conservatives accept the basic values of the two tradi-
tions, each group emphasizes those parts of the [political culture] most compat-
ible with its own philosophical disposition" (1984, 233). In sum, nearly all
Americans have absorbed the principal elements of their political culture, and as
Hochschild in particular has shown, they are highly sensitive to its characteristic
fault lines. Yet they are relatively nonideological in that most do not reconcile
these tensions in ways that would lead to the develbpment of consistent liberal
or conservative ideologies. These two conclusions-which, we believe, help
clarify the Lane-Converse controversy-constitute our presuppositions as we en-
ter this study.
The Intersection of Political Culture, Ideology, and the Welfare State
Weak though ideological conflict may be in the United States, it obviously

remains an important influence. A principal purpose of this study is to show how
liberalism and conservatism relate to the values of the larger political culture.
Inasmuch as democratic and egalitarian traditions are present in the political
culture alongside individualism and limited government, both conservatives and
liberals are forced to deal with inherent tensions: conservatives cannot easily
dismiss equality and democracy as core values in U.S. society nor can liberals
ignore the values of individualism and limited government.3 Nonetheless, policy
disputes over the welfare state raise value conflicts that are far more troubling to
liberals than to conservatives. Conservatives can, after all, readily justify anti-
social welfare attitudes by appeals to the values of individualism and limited
I Individualism and egalitarianism are extremely broad concepts that are variously defined (and
ill-defined) in U.S. political culture and political ideology literatures. We focus on economic indi-
vidualism: the commitment to merit as the basis for the distribution of rewards in society and the
belief that people ought to work hard. Egalitarianism has a number of dimensions that must be
distinguished: equality of opportunity-each person should have the same initial chance of succeed-
ing; formal or legal equality-all people should be treated equally; and equality of rewards. In the
last case the desire need not be for complete equality but may involve a limited range of wealth, a
floor on income, or an income limit (Verba and Orren 1985; Hochschild 1981).
government. And if pressed that these values leave too little room for equality
and democracy, conservatives can reply that they are strong proponents of
equality-but equality of opportunity rather than of outcomes. Indeed, it has
been argued that Americans (especially conservatives) attach so much impor-
tance to equality of opportunity precisely beeause it appears to offer an unas-
sailable formula for reconciling economic individualism with egalitarianism
(Potter 1954; Verba and Orren 1985).
The solution for supporters of the welfare state is not so simple. Direct
appeals to equality in justification of a more active commitment to assisting the
poor run up against the values of limited government and a free market economy.
And to the extent that the welfare state requires some redistribution of income,
counterclaims will be generated that this infringes on individual liberty and
places limits on the extent to which individual effort is rewarded (see Verba and
Orren 1985). Perhaps even more critically, it may not be possible to develop a
more positive conception of equality in a political environment in which the
language of debate is dominated by liberal individualism (Hartz 1955; Wills
1971).
In view of these difficulties in finding ideological justification for their pref-
erences, supporters of the welfare state have historically tended to fall back on
pragmatic and ad hoc justifications. Clearly this was the case with the most
important American welfare reformer of the century, Franklin D. Roosevelt,
who, in seeking support for his New Deal policies, prided himself on the claim
that he was just looking for ideas that would work. Other major liberal figures of
the recent past, notably John Kennedy and Lyndon Johnson, were likewise re-
nowned for their pragmatic searches for down-to-earth solutions. We thus antici-
pate that, in attempting to reconcile their support for social welfare measures
with a national political tradition that is in many ways inhospitable to such mea-
sures, social welfare liberals will tend to point to particular programs that they
consider especially important and useful, rather than to invoke overarching ideo-
logical principles.
The claim that liberals are more susceptible than conservatives to value
conflict over social welfare policy receives support from two very different sets
of studies. Rokeach (1973) has argued that an individual's political ideology can
be recovered from the relative rankings of the values of freedom and equality. In
the U.S. case, conservatives rank freedom very highly but place somewhat less
emphasis on equality. Liberals, on the other hand, also rank freedom highly but
insist that equality is at least as important. They are confronted therefore with
the problem of balancing two equally important values. For this and other rea-
sons, Tetlock (1984, 1985, 1986) argues that liberals will be more prone to value
conflict than conservatives and that this will be reflected in the greater com-
plexity of liberal arguments-complexity defined as "guidelines or criteria for
coping with the tension between the desired and undesired effects of a policy"
or "rules that clarify why 'reasonable people' might take different stands on a
policy issue" (Tetlock 1986, 820). He finds support for this hypothesis across a
number of groups.
These arguments generate several additional hypotheses that we wish to
test. First, we anticipate that social welfare liberals will exhibit more value con-
flict over social welfare policy than will conservatives. Second, supporters of
social welfare should be more likely than opponents to use concrete references
to programs and social groups to defend their policy preferences. Conversely,
we expect social welfare conservatives to be more likely to justify their prefer-
ences by appeals to values, especially individualism. Finally, we also want to
investigate the substantive basis of support for social welfare policy. If arguments
about the classical liberal basis of the political culture and the belief in equality
of opportunity are correct, how is the social welfare policy defended by its
supporters?
Studying Ideological Conflict and Ambivalence
When questions such as these have been empirically studied, the method-
ological vehicle generally has been in-depth interviews with a relative handful of
purposefully selected respondents (see, e.g., Lane 1962; Sennett and Cobb 1972;
Lamb 1974; Hochschild 1981; Reinarman 1987). These studies have the advan-
tage of providing a wealth of information about the underlying beliefs (and rea-
soning processes) that people use in evaluating policy options. Their use of open-
ended methodology, in combination with the great amount of time devoted to
each respondent, produces a detailed picture of the respondent's belief system
not easily obtained from fixed-choice survey questions. And consensual aspects
of beliefs and values can be explored, since the purpose of these studies is typi-
cally not to analyze covariances across subjects.
Despite their valuable contributions, there are serious limitations to what
can be learned from in-depth interview methods. One derives from the difficulty
of generalizing from small, nonrepresentative samples. How typical are the opin-
ions of 15 people from "Eastport" or a similar number from the West Coast?
Although the authors of these studies resist overgeneralizing from their samples,
it remains highly desirable to know whether the results they report can be gen-
eralized to the public as a whole. Another shortcoming is that it is extremely
difficult in samples of 15 or 30 people to detect even large individual differences
in response patterns. Hochschild (1981), for example, emphasizes the absence
of ideological cleavage among her respondents. But with only 28 respondents
and no standard measures, how can she be certain of this? Questions also can be
raised about the degree to which the active participation of the investigator influ-
ences the respondents' answers to the questions and probes. Would these people
seem as "sophisticated" as they do without a probing researcher? A related prob-
lem arises in the selection of cases for analysis. In writing up their results, re-
searchers seem to gravitate toward individuals who have more interesting things
to say. The result is that the reader's impressions may be disproportionately in-
fluenced by the remarks of a handful of unusually articulate or colorful respon-
dents. Finally, the conclusions from these in-depth interviews have accumulated
with little connection to the bulk of the research in public opinion that uses mass
survey methodology. This has unfortunately created two almost independent lit-
eratures that are often seen as producing different conclusions about ideology
and political conceptualization.
In light of these problems, a secondary goal of this paper is to explore the
utility of new types of open-ended questions on mass survey samples. We want
to see whether focused, open-ended probes, in combination with systematic cod-
ing of all responses given, can be used to explore the sorts of questions typically
addressed with in-depth interviews. This would provide an additional method-
ological vehicle for studying questions related to political ideology and political
reasoning, one that would combine advantages of both survey research and depth
interviews.
Data and Methods
The data for this analysis are drawn from the 1987 National Election Studies
pilot study. The pilot study was based on reinterviews with a random sample of
450 people originally interviewed as part of the 1986 National Election Study.
The pilot study was a two-wave panel design with respondents first interviewed
in May and then again in June. For experimental purposes the survey was split
into two random half samples.
The responses used in this study were produced by open-ended probes
added to a pair of standard NES policy questions: (1) guaranteed jobs and stan-
dard of living and (2) increased government spending versus cuts in services.4
4The wordings of the two policy questions are as follows. Spendinglservices: Some people
think the government should provide fewer services, even in areas such as health and education, in
order to reduce spending. Other people feel it is important for the government to provide many more
services even if it means an increase in spending. Which is closer to the way you feel or haven't you
thought much about this? Jobs and standard of living: Some people feel the government in Washing-
ton should see to it that every person has a job and a good standard of living. Others think the
government should let each person get ahead on their own. Which is closer to the way you feel or
haven't you thought much about this? A third question-government assistance to blacks-was also
probed in the pilot study, but the responses to this question have not been used in this paper. Although
the patterns of responses to this question mirror those of the spending/services and jobs questions,
many of the comments on government assistance to blacks referred directly to blacks (i.e., discrimi-
nation against them or their opportunity to get ahead) or to attributes of blacks. One entirely new
category of responses is thus created (negative attitudes toward blacks), and many of the other re-
sponses that reflect the same general categories used for the two social welfare questions are inter-
twined with race. Using a common coding scheme runs the risk of confounding these general cate-
gories with race. The alternative strategy of creating a new set of categories for comments that
involve references to race would make the presentation and analysis exceedingly long and messy.
Two forms of the open-ended probes were used. For half the respondents (form
A) the issue questions were asked in the normal way but were immediately fol-
lowed by these open-ended questions:
Still thinking about the question you just answered, I'd like you to tell
me what ideas came to mind as you were answering that question. Exactly
what things went through your mind?
Are there any (other) reasons that you favor [the option just selected]?
Do you see any problems with [the option just selected]?
For the other half of the sample (form B), interviewers read the policy items in
the standard way, but without pausing for the respondent to answer, they asked
the respondent to "stop and think" about the question. For the jobs and standard
of living question, the exact probes were
Before telling me how you feel about this, could you tell me what kinds
of things come to mind when you think about government making sure that
every person has a good standard of living? (Any others?)
Now, what comes to mind when you think about letting each person get
ahead on their own? (Any others?)
These open-ended probes were designed to elicit the thoughts that respon-
dents had as they went about answering the two social welfare questions.5 Our
hope was that these probes would reveal how respondents frame issues, what
values or other considerations were especially important to them, and what de-
gree of ambivalence (if any) they experienced. In form A the intent was to have
people answer the survey question as they normally would and then immediately
find out what they were thinking about as they did SO.6 In the second form,
respondents were asked first to think about each component part of the question
before giving their position on the issue.
All responses to these probes were recorded by the interviewers and then
coded according to an elaborate coding scheme. The substance of each response
was coded along with the direction of the statement (i.e., whether it favored a
MThese two questions were chosen, in part, because they have been the primary NES social
welfare issue items for the last decade (the history of the guaranteed jobs item goes back to the
1950s). This, of course, does not ensure that they do an adequate job of representing Americans'
attitudes on social welfare policy. As a check on this, these two questions were factor analyzed along
with other questions on domestic spending priorities and general attitudes toward economic equality
and redistribution. A clear first factor emerged from the analysis with the two NES issue questions
loading strongly on that factor.
6This is based on work done by Ericsson and Simon (1984) that shows that people are able to
report the thoughts they had while completing a task if they are asked immediately after they have
completed the task.
given side of the issue), indications of affect, and degree of elaboration.7 The
coders were unaware of our intent to study the active use of values and principles
by the U.S. public. The substantive codes involved more than 150 discrete cate-
gories. For ease of presentation and interpretation, those categories have been
combined into two coding schemes. The first combines all the initial codes to
yield 15 master categories plus a residual code. The second focuses more closely
on just the major political values. The Appendix contains a detailed description
of the substantive comments that are included in each category.
To control for variations in "talkativeness," the coding schemes reflect
whether an individual made any comment that falls within each substantive cate-
gory. Thus, multiple comments in any one category or elaborations on a single
theme are not counted as separate comments. (Each category is coded as a 0-1
dummy variable.) To maximize the number of respondents we have to work
with, the two half samples have been combined. A detailed check of the distri-
butions of responses produced by the two different probes showed a very high
degree of similarity (see Table A. 1 in the Appendix).8 The substance of the
comments made seems not to have been affected by the form of the open-ended
probe. Finally, we have combined the two waves of responses to each policy
question. Since the two issue questions were asked in both waves, the results we
shall present come from each respondent having two distinct opportunities to talk
about these issues.
These data are rich in information about the considerations that people used
to respond to these two survey questions, but they are also difficult to manipulate
statistically. There is a trade-off between preserving the informational value of
the 150 codes and each respondent's opportunities for multiple responses, and
the presentation of simple summary statistics. Since this is an initial exploration
of the use of open-ended questions to explore popular understandings of public
policy preferences, we have chosen to present the data in relatively unprocessed
form. This will give the reader an opportunity to examine the patterns of re-
sponses in close to their natural state. A few summary statistics will be pre-
sented, but the bulk of the data presentation will involve analysis of frequency
distributions for the open-ended responses.
7The coding scheme was developed by the authors in consultation with Steve Pinney of the
coding section of the Institute for Social Research at the University of Michigan. The actual coding
was carried out by experienced coders of the ISR staff and subject to frequent cross-checking by
Pinney.
8Differences between the two forms are evident in the process by which people answered the
questions (see Zaller and Feldman 1988). For the present purposes, there appear to be few differences
in the substance of the dimensions people thought about. The most significant difference between the
two forms is that there are substantially more missing cases for the retrospective form than for the
stop and think form. Those people who offered no opinion on the issue question were dropped from
all of the following analysis. See Table A. 1 for a direct comparison of the two forms of the open-
ended probes.
The Nature of Beliefs about Social Welfare Policy
Table 1 shows the distribution of responses across the 15 master categories

for the two policy questions. For each question the distributions are shown sepa-
rately for those who gave two prosocial welfare, two antisocial welfare, or mixed
responses to the policy questions across the two waves of the pilot study.9 For
convenience we will often refer to the prosocial welfare position as "liberal" and
the antisocial welfare position as "conservative." The ideological labels do not
necessarily imply any broader understanding of an ideological continuum or
extension to ideological views outside the social welfare domain. Before exam-
ining the differences among these three groups, it is useful to get a general feel
for the comments our respondents gave to these series of open-ended probes.
The first thing to note is the widespread and active use of abstract termi-
nology. Looking first at the services and spending item, we find that 53% of
respondents made at least one remark that either invoked a value such as indi-
vidualism, humanitarianism, or limited government or made an argument at a
comparable level of abstraction.'0 Overall, one out of every 10 of the individual
comments from the probes of the spending/services question referred to an ab-
stract value or principle. The remainder referred mainly to specific programs or
problems that were especially important to the respondent. These results are all
the more impressive in that they derive from a question that is fairly concrete
and does not explicitly raise these value concerns.
Turning now to the item on job guarantees and living standards, we again
find that a high proportion of respondents invoked some value or principle. The
largest number of these responses refer to individualism. Altogether, about three
out of four people invoke this value in some way. Some references were little
more than repetitions of the question ("Well, I think people should get ahead on
9As can be seen in Table 1, a large number of people gave different responses to the social
welfare questions across the two waves of the pilot study. This is consistent with all other studies
that have looked at response stability in issue preferences. The major explanations for this temporal
instability are that it either reflects nonattitudes or high levels of random measurement error. A
third interpretation is that this response instability is a result of conflicting or ambivalent considera-
tions (Zaller and Feldman 1988). The evidence to be presented here will offer some support for the
latter view.
'?We used a very conservative definition of abstract remarks. This includes responses in cate-
gories 9 through 14 as well as those comments in category 15 that refer to some value or principle
(not just references that people need help). If the definition of abstract remarks were expanded to
include categories 5, 7, and 8 as well, the proportion of respondents making at least one abstract
comment rises to 75%. We are using abstract and concrete here to refer to the nature of the terms
used in the responses. We think it is important to distinguish between use of a general principle (like
all people should get ahead on their own) and reference to a specific condition or government pro-
gram. This distinction does not necessarily imply that the reasoning processes underlying these
responses parallel the same abstract and concrete patterns. Indeed, it is possible that concrete re-
sponses can be given by someone reasoning abstractly while abstract comments may be given by
someone who is thinking about the world in very simple manner.
Table 1. Distribution of Responses by Policy Preferences
(In Percentages)
Spending/Services Jobs/Standard of Living

Anti- . Pro- Anti- . Pro-
Mixed
welfare welfarexe
welfare welfare
comments 5.9 9.5 13.5 7.3 7.7 9.7

1. Personal
2. Politics 13.7 4.7 2.3** 1.5 1.8 3.2

3. Specific group
references 21.6 31.4 46.6** 8.8 14.8 19.4
4. Specific program
comments 86.3 91.7 96.2 40.9 40.2 69.4**
5. Sophisticated
program comments 21.6 17.8 16.5 .7 2.4 6.5 *
6. National
conditions 7.8 8.3 8.3 8.8 19.5 24.2**
7. Anti-American 9.8 0.0 0.0** 16.8 6.5 1.6**
8. Taxes/budget 56.9 38.5 35.3** 9.5 7.7 14.5
9. Positive
government role 2.0 2.4 7.5 16.8 33.7 43.5**
10. Antigovernment
orientation 58.8 32.0 22.6** 58.4 39.1 37.1**
11. Individualism 35.3 21.3 13.5** 95.6 72.8 50.0**
treatment 0.0 0.0 0.0 8.0 5.9 3.2
12. Equality of
13. Qualified
assistance 9.8 7.7 10.5 21.2 24.9 33.9*
14. Lack of
opportunity 0.0 1.2 2.3 10.2 7.1 22.6**
15. Fairness/equality/
assistance 7.8 16.9 25.6** 22.6 24.9 33.9*
comments 1.9 8.1 1.5 .7 10.1 9.7**
16. Other abstract
N= 51 169 133 137 169 62

Note: Frequencies total to more than 100% due to multiple responses to each question.
* = differences across groups significant at .05 level; ** = differences across groups significant at .01
level.
their own if they can"), but usually people provided their own renditions of the
concept of individualism. NES policy precludes us from providing verbatim tran-
scriptions of any individual's remarks. However, we can convey the flavor of
these remarks by quoting from the code book to the pilot study some of the most
commonly invoked codes referring to individualism: "I
Code 140. Individuals should make it on their own; people must make use
of the opportunities they have; people should be responsible for themselves;
people should just work harder; people have the right to work as much or
little as they want; they control their own fate.
Code 141. Dependency; living off handouts is bad; welfare makes people
dependent; "if it's too easy to get welfare, no one would work anymore";
people become lazy or lose self-respect if they are on welfare; "the more
you give, the more they want."
Code 142. People who don't/won't work don't deserve help; people who are
poor deserve to be poor; "if you can't make it in America, you have only
yourself to blame"; anyone who really tries cap make it.
There are certain common themes across these two social welfare questions:
specific mentions of federal programs and groups affected; taxes, budget prob-
lems, and general antigovemment feelings; and values of individualism and com-
passion for the disadvantaged. However, the job guarantee item produces many
more comments that invoke values and principles than does the services item. In
fact, 95% of respondents made use of at least one value or general concept in
discussing their thoughts on this issue. Moreover, 61 % of all of the responses to
the open-ended questions invoked a value or principle. This is all the more im-
pressive since the responses in most of these categories are unaffected by the
respondents' levels of political information (see Table A.2 in the Appendix).'2
'The italicized phrase is the central theme; the supplemental language was loosely adapted
from the transcripts (to avoid things that particular respondents might have said) in order to indicate
to the coders the kinds of remarks that would justify use of the code.
'2These two policy questions clearly differ in the extent to which they appeal to "prime" values
and principles. The spending/services question is very concrete. On the other hand, the guaranteed
jobs question explicitly mentions letting people get ahead on their own. It is perhaps not surprising
to find more references to values and principles in responses to the second question. There are several
things to keep in mind, however. First, despite the lack of priming by the question, there are a large
number of responses to the spending/services question that refer to values and principles. If a con-
crete question like this defines the lower limit to the number of such responses, it is still substantial.
Second, we are much more interested in how those on opposing sides of each issue defend their
positions than on comparisons of numbers of responses across issues. Although the levels may vary,
our analysis shows many common patterns in the responses to each question. Finally, some may be
concerned that responses to the guaranteed jobs question are biased because the conservative position
more clearly evokes value concerns than the liberal position. While we cannot completely dismiss
These findings should remove whatever doubt may exist that the results
obtained from Lane-style in-depth interviews depend on either the particular
samples chosen or the probing interview technique used by the investigators.
When the subject of discussion is the substance of government welfare policy
rather than presidential politics, Americans are, by our survey-based evidence,
quite able to make active use of values and principles in articulating their views.
The values they invoke are not the partisan values of liberalism and conservatism
but the common cultural values of America's liberal tradition.
Prosocial and Antisocial Welfare Arguments
We have hypothesized that social welfare liberals would experience more

conflict in discussing welfare policy because they are the ones who must recon-
cile their policy preferences with the pronounced individualistic and antigovern-
ment emphasis of the political culture. A closer reexamination of Table 1 pro-
vides initial confirmation of that expectation.
Looking first at the services/spending issue, we find that those opposed to
increasing government programs draw upon three highly consistent concerns:
taxes and budgets, opposition to big government (laissez-faire and bureaucracy),
and individualism. In contrast, the comments of social welfare liberals are char-
acterized by positive responses to particular government programs and feelings
of sympathy for affected groups but also by concern about taxes and the size of
government. In fact, 35% of social welfare liberals mentioned tax and budget
issues, and 22% made some antigovernment statement. It is also notable that
scarcely any advocates of more government services advocated a positive role
for government or a more egalitarian society. Thus, few supporters of more gov-
ernment services were able to offer a consistent ideological justification for their
positions, and many indicated awareness of negative consequences of govern-
ment spending. Antisocial welfare people are much more consistent in their re-
jection of greater services-complaining about tax and budget problems, spe-
cific programs, and big government, on the one hand, and invoking the values
of economic individualism, on the other. Though, at the same time, many do
cite one or more government programs of which they approve and show some
sympathy for needy groups.
Responses to the jobs and living standards question provide further support
for our argument. Those on the conservative side of this issue are virtually unani-
mous in their use of individualistic arguments, citing such things as the value of
hard work and the sufficiency of equal opportunity, and problems of fairness and
this concern, it is important to note that those opposed to greater government spending are also much
more likely to raise individualistic arguments than those in favor despite the lack of any explicit
primes in that question.
dependency. Almost 60% of them also complain about the general role of gov-
ernment. At the same time, only a small handful of those opposed to government
guarantees admit that some people are disadvantaged and need assistance and
that the government has some role in providing social welfare services.
Supporters of guaranteed jobs and living standards show substantially more
evidence of sympathy for the disadvantaged, but much of it is qualified: their
basic posture is that people should get ahead on their own but those who cannot
do so should receive assistance. They are also likely to make references to par-
ticular programs or to the poor state of the national economy. A little over 40%
of them made a favorable reference to the positive role of the federal govern-
ment, but more made either a negative reference to the role of government or a
positive reference to individualism. Thus, while those opposed to government
guarantees indicated some ambivalence in their qualified admission that some
people are disadvantaged, supporters of the welfare state were deeply conflicted.
When they spoke in purely abstract or principled terms, they were more unfa-
vorable than favorable toward social welfare, even though they had endorsed the
social welfare position on both waves of the survey.
People who support guaranteed jobs and living standards defend their posi-
tions with two main types of arguments. The first type is an endorsement of
specific programs or indication that certain people need assistance. The second
type of argument is more abstract. Although less widespread than the conserva-
tives' use of individualism and antigovernment arguments, this argument centers
on the contention that there is not full equality of opportunity along with support
for government activity to intervene when people need assistance. Thus, while
opposition to social welfare stems in part from a commitment to equal opportu-
nity, support for social welfare is often justified by beliefs that significant barriers
to equal opportunity exist (see Kluegel and Smith 1986).
The Extent of Value Conflict
Table 1 shows that supporters of social welfare are more likely than oppo-
nents to report values and beliefs contrary to their expressed policy position.
However, it is difficult to determine the magnitude of the value conflict or the
difference between social welfare liberals and conservatives from the data in this
form. We therefore went back to the responses to the open-ended probes and
coded each response as prosocial welfare, antisocial welfare, or ambivalent. In
addition, we classified each response as abstract (reflecting a value or principle)
or concrete (typically referring to a specific program, group, or condition of the
country).'3 The results are shown in Table 2.
The most striking finding in Table 2 is that values and general principles
"3The codes used to construct the abstract and concrete categories are the same as reported in
note 10.
Table 2. Levels of Conflict for Abstract and Concrete Responses

(In Percentages)
Spending/Services Jobs/Standard of Living

Response Anti- Mixed Pro- All Anti- Mixed Pro-
Type welfare welfare welfare welfare All
Antiwelfare 74 47 23 45 76 56 29 58
Abstract:
Ambivalent 20 16 26 20 11 13 19 13
Prowelfare 06 37 51 35 13 31 52 29
100 100 100 100 100 100 100 100
Antiwelfare 53 28 18 27 42 14 10 22
Concrete:
Ambivalent 29 29 30 29 37 49 35 44
Prowelfare 19 43 52 42 21 37 55 34
100 100 100 100 100 100 100 100
Abstract 19 11 09 11 70 58 49 61
N = 51 169 133 353 137 169 62 368
predominate in the discussion of the job guarantees issue (61% of all comments),
while more concrete, group- and program-related comments are predominant in
the discussion of the government services item (89% of all comments). This
makes a point that, although it is perhaps unsurprising, has often been over-
looked in discussions of the public's proclivity to use abstract terms, namely,
that this proclivity depends heavily on the particular issue or problem about
which people are questioned and the ways in which those issues are framed.
The second point to notice in Table 2 is that, in the population as a whole,
the balance of abstract remarks favors the antisocial welfare position on both
issues, while the balance of concrete remarks favors the prowelfare position on
the same issues. This tendency is most pronounced in responses to the job guar-
antees item. Abstract comments run heavily against the social welfare position
(58% to 29%), but concrete comments run in favor of it (34% to 22%). This
finding supports Free and Cantril's observation that Americans tend to be am-
bivalent welfare statists: philosophical conservatives but operational liberals. 14
'4It is, of course, possible that there is some general principle or value underlying references
to programs or groups that our methodology was unable to discover. We do think this is unlikely,
however. The data show that these questions and probes were generally successful in illiciting refer-
ences to principles and values. The coders were also instructed to note any evidence of higher-order
concepts in the responses. Despite this, the concrete comments appeared in sizable numbers primar-
ily where prior studies suggested they would.
The value conflict that fuels this ambivalence takes different forms for social
welfare liberals and conservatives. The frequencies for abstract comments are
consistent with our hypothesis that social welfare liberals should be more suscep-
tible to value conflict than conservatives. For both issues, barely half of all ab-
stract comments by social welfare supporters are favorable to welfare. While
prowelfare abstract comments are relatively few among those who twice opposed
social welfare (6% for spending/services and 13% for jobs), the corresponding
antiwelfare comments among social welfare liberals are significantly greater
(23% and 29%). In addition, social welfare liberals are more likely than conserv-
atives to make statements that show ambivalent feelings about these two issues.
The picture is less clear for the concrete responses. In the case of spending
and services, supporters and opponents of social welfare show virtually identical
levels of conflict. The results for the guaranteed jobs question are more in line
with our expectations. Social welfare conservatives are twice as likely as liberals
to give concrete comments inconsistent with their expressed issue position.
The Underlying Values: A Closer Look
Our analysis of Tables 1 and 2 has suggested the existence of considerable

ambivalence among social welfare liberals and conservatives, but the nature of
the coding scheme used in these tables obscures differences that come through
clearly once the comments have been broken down more finely. Our category of
individualism, for example, encompasses a variety of work and opportunity val-
ues on which social welfare liberals and conservatives differ more sharply than
is indicated in Table 1. Simultaneously, commitment to certain general values is
sometimes hidden by the way the categories have been constructed. For example,
liberals often said that people should work hard but may still need help in over-
coming the effects of the weak national economy. Although such comments in-
dicate that the respondent is working within an individualistic framework, Table
1 does not count such remarks in the individualism category.
To deal with these limitations of our initial breakdown of the pilot study
codes, we have devised a second and more detailed classification scheme that
pays particular attention to three sets of values and principles: individualism,
humanitarianism, and opposition to big government. This coding scheme both
combines certain categories from Table 1 into more global ones and simultane-
ously subdivides those categories into more finely differentiated sets of re-
sponses. Note that two subcategories, "qualified help" and "lack of opportu-
nity," are simultaneously counted under individualism and humanitarianism. On
the one hand, both sets of responses indicate an awareness that some people do
need assistance. At the same time, this need for assistance does not challenge
the underlying assumption of individualism. A description of this new coding
structure is presented in the Appendix. The results are shown in Table 3.
When all individualism-related comments are combined, we see that the
differences between those who are consistently prosocial or antisocial welfare on

each question are somewhat different in nature then they first appeared. For the
jobs and living standards question, opponents of social welfare show virtually
unanimous concern for individualism, but now fully three out of four social
welfare liberals do so as well. What varies is the way in which the two types of
people express their concerns. Those opposed to job guarantees are much more
likely to invoke rigorous norms of individual effort-the value and efficacy of
the hard work, the fairness of the economic system, the existence of equal op-
portunity, and problems of dependency. Although these elements of individual-
ism are raised by welfare liberals as well, the latter are much more likely to
contend that some people do work hard and still need help, possibly because true
equality of opportunity does not exist. Still, over 40% of those supporting jobs
and living standards guarantees express the basic idea that people should simply
work hard and take care of themselves.
On both the spending/services and jobs and living standards questions, wel-
fare liberals are more likely than their conservative counterparts to express sym-
pathy for the disadvantaged. What is interesting is the qualified way in which
these concerns are expressed. The most common single theme is that, even
though people should get ahead on their own, they may still need help. Others
argue more directly that people should try to get ahead on their own but that
opportunities are not currently equal and should be made more so. However,
ideas that would provide a clear egalitarian basis for support for the welfare
state-for example, a clear appeal for equal outcomes, or claims of individual
entitlement-are rarely mentioned. In fact, supporters of social welfare are
hardly more likely to invoke these values than welfare opponents.
Table 3 also shows that, for both questions, more social welfare conserva-
tives than liberals express antigovernment sentiments, although fully half of wel-
fare state supporters make at least one such comment. Again, however, there are
sharp differences in the form these feelings take. Among opponents of social
welfare programs, there are a large number of laissez-faire/limited government
comments; supporters make relatively few principled statements favoring overall
government restraint. Opponents are also frequently critical of government bu-
reaucracy, worried about the limits of what government can do, and upset over
taxes and the budget. But we find a surprisingly high number of welfare liberals
who express similar concerns. On the spending/services question, more than half
of welfare liberals complain about taxes, the budget deficit, or problems of bu-
reaucracy. For jobs and living standards, one liberal in four talks about the limits
of government action in solving these problems and others mention tax/budget
problems and bureaucracy. Thus, where opponents of social welfare often op-
pose big government per se, supporters attack what many see as the undesirable
side effects of big government: bureaucracy, taxes, and the difficulties of solving
large-scale social problems.
****
Q) * * * * * * *
a 00 1 t - ? oo ^^
(n SC--en 3l 4
os cn o ^ oo oo oo o. (Z (ON en
Cu 0 o 0o oo el; _ tl 0o
|~~~~~~r -t c) c_ o r
Cu~~~~~~~~~~~~~~~~~~~~1
CNC
_O .c' t t __ t) cO tt cO c
C)4
- - os c) _ ro _ r- c) _
c- c) oo mc c) r- c o o c)
C,,
C) ~ ~ ~ C
t~~~~~~~~ O ._ e 6 o tob o
. ~ |. 4. N o0
Cu ~~~~~~~~~~~>
-~
****
eeoo ?c - OO 4 o en 0 W
00? ~ - ~el C -6 -4 t 4
eN~~t~ ~~~~s1C' Rt}Nm
-M C1 -O ^'o m
-
cn 4 06 el o o o6 r-o b
00 r % - '
oo N vE - so N > ? It C1 c v
* * ** *
* **
00 00 00 00 ~c O)00O IC ~ c We
(1 - -ei- ', eN ^
qt O 00 OO r-riN 4 t en ^0
'I O oo o-oo ta
O O 0 00 00o O t O 4 O O
F0 6 6 o Fo o6 --c -
00 (1~ 'I~ 0
- - -N00t~ ~m
~
m
| E * - E_
0YDQnQWRV
These results confirm and clarify our initial expectations that social welfare
conservatives would be able to draw upon a consistent set of values and beliefs
to support their opposition to social welfare policy. Their individualism state-
ments elaborate several themes concerning the efficacy of hard work and the
value of opportunity, and their antigovernment comments are strongly rooted in
the laissez-faire ideology. There is some evidence of ambivalence in their beliefs,
most clearly in occasional remarks that some people really do need assistance.
For the most part, this sympathy is expressed in strongly individualistic terms.
Many social welfare liberals, on the other hand, support social welfare policies
while also endorsing individualism and criticizing the size of the federal govern-
ment. Their ambivalence is more extensive and deep seated. Many do qualify
their endorsements of individualism and criticism of government, but few wel-
fare liberals evoke egalitarian principles of the type that might provide a strong
ideological foundation for the welfare state. Their abstract justification for social
welfare policies involves a conviction that some people may need help despite
their efforts and a belief that the government is obligated to provide that
assistance.
Egalitarianism, Sophistication, and Support for Social Welfare
Our analysis so far has focused on how people who favor or oppose particu-
lar welfare state policies describe and justify their views. We shift now to a
related question: How do people who differ in their professed commitment to
egalitarian values describe and justify their social welfare attitudes? We want to
see whether people who are, in a general sense, strongly egalitarian can mobi-
lize egalitarian arguments on behalf of specific social welfare attitudes. That is,
we want to determine how, if at all, egalitarian values are translated into argu-
ments for social welfare policies. Many have argued that egalitarianism is the
value dimension that is most relevant to policy debate over social welfare (Feld-
man 1988; Huntington 1981; Kinder 1983; Lipset 1979; Rokeach 1973; Verba
and Orren 1985). If so, how can we account for the apparent absence of egali-
tarian arguments among supporters of social welfare policy?
To measure egalitarianism we use a series of NES items designed to mea-
sure support for greater formal and economic equality in the United States (see
the Appendix for the exact wordings of these questions). An additive scale based
on these items has a roughly normal distribution. Those in the upper third of the
scale repeatedly endorsed statements calling for greater equality. Studies have
shown that these equality items are powerfully related to a number of attitudes
toward economic and racial policies (Feldman 1988; Kinder and Sanders 1987).
How is this reflected in the open-ended responses?
When we examined the effect of egalitarianism on justifications for social
welfare attitudes, we found patterns that closely resembled those in Table 2.'1 In
"5These data are not shown but are available from the authors upon request.
particular, persons scoring high on egalitarianism showed, as expected, rela-

tively high levels of sympathy for the disadvantaged, but again this concern often
took the form of qualified statements of assistance. A little over 40% of those
high in equality also made statements endorsing a positive government role in
dealing with social welfare issues, but here as elsewhere, this support for gov-
ernment action was unaccompanied by statements that justified it in terms of the
value of equality. This is a very notable fact. Our coding scheme was highly
sensitive to the possibility that respondents would justify their preferences by
appeals to egalitarian values and included many ways in which they might do so.
Yet very few made such statements. We infer from this that the American politi-
cal culture provides few explicitly egalitarian (as against pragmatic or humani-
tarian) arguments that are useful for justifying welfare state policies.
These new results are at once consistent with what we have found before
and are puzzling. In the closed-ended egalitarianism items, many people do en-
dorse egalitarian statements that go well beyond simple equality of opportunity.
And variation in support for equality is strongly related to policy preferences in
the social welfare domain. What is conspicuously missing, as we have stressed,
is the verbal justification of these preferences in terms of egalitarian principles.
Is it not possible, however, that the sort of ideological superstructure we expect
exists only among those who are politically sophisticated? If so, many of the
conclusions we have reached to this point about the extent of ambivalence and
structuring of beliefs may be masking large differences in the use of these di-
mensions between the more and less politically aware segments of the public.
To investigate this we constructed a measure of political information from
items in the pilot study data and those previously asked in the 1986 survey (see
Zaller 1986 for a discussion of this measure). To retain enough cases for com-
parison purposes, the information measure and the equality measure were both
dichotomized at their median. 16 These results are shown in Table 4.
As can be seen, there is virtually no support for the hypothesis that higher
levels of sophistication are required for the expression of abstract values and
principles. Among those low in egalitarianism, there is some tendency for in-
formation to be related to the extent and elaboration of the individualistic state-
ments made. More sophisticated nonegalitarians make somewhat greater use of
several individualistic arguments than their less sophisticated counterparts and
are slightly less likely to advocate assistance for the disadvantaged. There are
even more modest effects of sophistication for those scoring high on the egali-
tarianism scale. But there is no consistent tendency for those lower in informa-
'6We also estimated probit equations for each of the response categories that included the
continuous versions of the equality and information scales as well as a multiplicative interaction
termn. Those results virtually replicate the data shown in Table 4. We have presented the simple
frequencies in order to show the overall use of each type of response as well as differences across
issue preferences and information.
= = cN I) m t oc, ' -- r 'b
"0
M IC "I ?t "I r- W) 00 00
"0 * * *
- s o - <;? Cs t-oo C - o
C ~ ~ ~~~0 1
2eo o o
Cu~~~~~~~~~1 00 WC W - o - t o t o oo
ob. o o) c) c) o bo
O O 3 O o O 00 O o o
? ~~~~~~*
Cu "0 =C (ON o00 oQ O en
o~ ~ 3- - ooo
a- o too-_
Cu ~ Sc
lqt 00 O~ * c c
C~ C- cli 6 r * *- - c 6 r
00 O (ON lC -C r en -e 0C
- ; ttr 6 ~64o
0o~ o~ o~c'r-~ 0o ooo-i C'i4
000 ON"C C O ~00 00 (ON Lf)

e4 el4 IC Clf 'qt 0
el - - ~~~~~~~~~~Cf 00 u
IC O 0t 00 O C t- e400-
e~C;Ci6l 0e~6 6OC LrO' el r~ -
el O Olq f - 4 ON (ON ') 00 0
16.~~~~~~.
40 -
oo~~~~i
2t
tion to make less use of individualistic arguments and only a barely observ-
able increase in the proportion of statements on assistance that are not highly
qualified.
Apparently, then, the failure to invoke explicit egalitarian values has little
to do with levels of political sophistication. The values of individualism and
sympathy for the disadvantaged are widely diffused in the public independently
of political sophistication. Such is not the case, however, for antigovernment
arguments. Increases in political sophistication are consistently related to greater
expression of all types of antigovernment statements: laissez-faire, bureaucracy,
limits, and taxes/budgets. Moreover, this relationship holds for those low and
high in support for equality. '7 This produces two very different results, however.
For those low in support for equality, greater sophistication is associated with
even higher levels of opposition to an active federal government, making these
people even more consistently conservative in their arguments against providing
social welfare benefits. Among those high in egalitarianism, concern over big
government among the more sophisticated leads people to become even more
conflicted in their expressions of support for social\welfare. Seeing a need for
the federal government to aid those who have not received equal opportunity,
they are then faced with not only their commitment to individualism but often
with fears about the negative consequences of big government. Thus, increasing
sophistication among those high on the equality scale is associated with, not
greater commitment to egalitarian principles, but rather neoliberalism: "a 'tough-
minded' and 'nonideological' approach that recognizes fiscal and other limits
while still seeking to promote social justice and equity" (Medcalf and Dolbeare
1985, 56).
Conclusions and Implications
Ideological Responses to the Welfare State
The data we have examined show clearly that most Americans can draw
with apparent ease upon several elements of the U.S. political tradition in justi-
fying their social welfare preferences. The elements of the tradition most com-
monly invoked were suspicion of big government, humanitarianism, the Protes-
tant ethic, and above all, economic individualism. The fact that references to
these values emerged spontaneously while discussing policy questions is evi-
dence that they have real meaning to the people who invoked them. This conclu-
sion corroborates the position of Lane, Hochschild, and others who contend that
ordinary people view the world through the prism of a distinctive cultural
'7We obviously cannot interpret this relationship to mean that political sophistication causes
increased antigovernment sentiments. Sophistication is related to such factors as education and in-
come that could be responsible for this relationship. For our purposes, it is sufficient to observe how
response patterns differ across levels of sophistication.
bias-the ideology of classical liberalism. This set of ideas, according to many

scholars, virtually defines the political culture and has, by our data, diffused
thoroughly in the population. At the same time, however, our findings also con-
firm Converse's claim that most Americans are nonideological, where the term
ideological refers to the use of the liberal-conservative continuum to organize
one's political attitudes. We found some evidence of consistent left-right ideo-
logical structuring of justifications for welfare state policy preferences and some
tendency for more politically aware persons (particularly among conservatives)
to become more consistent. But these tendencies are rather modest.
Our results offer strong support to studies, especially that of Hochschild,
that have identified ambivalence as a fundamental feature of political belief sys-
tems. Most people are internally conflicted about exactly what kind of welfare
system they want. Even those who take consistently pro- or consistently antiwel-
fare positions often cite reasons for the opposite point of view. The reason for
this, we believe, is that ordinary people are both regularly exposed to arguments
that extol individualism and decry big government and also to political rhetoric
that urges sympathy for the poor and state action to ameliorate existing social
ills. Sensitivity to such arguments obviously does not make ideological consis-
tency, in the liberal-conservative sense, more likely. Indeed, the more people
attend to these contradictory messages, the more difficult it may be to maintain
a consistent ideological liberal-conservative position. Seen this way, ideological
consistency requires not just attention to politics and political debate but a rejec-
tion of some elements of the political culture in favor of others.
Ambivalence and inconsistency are not found with equal frequency in all
segments of the population. Social welfare conservatives exhibit relatively less
value conflict. Many of them do, to be sure, admit that people sometimes need
government help but the ability of conservatives to appeal to a wide range of
individualistic and antigovernment values keeps their sympathy for the needy
within definite and comfortable bounds. Conflict does emerge for opponents of
social welfare when they think about specific government programs and their
beneficiaries. Thus, the more welfare conservatives think about welfare policy
in abstract terms, the better they can feel.
Ambivalence with respect to social welfare policy is more pronounced
among welfare liberals. They must reconcile their humanitarian impulses with
the conservative principles of individualism and limited government. Many find
this difficult to do. They end up acknowledging the values of economic individu-
alism even as they try to justify their liberal preferences.
That the U.S. political tradition is inhospitable to the welfare state is
scarcely news at this point. Many prominent analysts, as noted earlier, have
viewed this as a fundamental feature of this nation's political culture. What is
surprising is the extent to which the elements of this inhospitality are internalized
even in the minds of people who most strongly support welfare state policies.
There are alternative ways to interpret our findings on liberal ambivalence

that ought to be noted. One possibility is that our analysis is time bound. Our
results showing that social welfare liberals have higher levels of value conflict
than welfare conservatives may reflect the current defensive position of liberal-
ism in the United States rather than any inherent differences between liberal and
conservative reactions to the welfare state. More generally, since our data come
from a single snapshot of attitudes in the summer of 1987, analysis of these data
cannot, by itself, establish that liberals are always more conflicted over social
welfare policy than conservatives. However, there is a great deal of other evi-
dence that our data have captured a long-standing and largely constant feature of
opinion in the United States. Hartz (1955), writing in the mid-1950s, discussed
the lack of an ideological justification for the welfare state in its formative years.
Free and Cantril (1968), reporting on survey data collected in 1965 (the height
of enthusiasm for the welfare state in this country), vividly showed the contrast
between opposition to the philosophical underpinnings of the welfare state and
high levels of support for specific programs. Hochschild's (1981) in-depth inter-
views from the mid-1970s explored in detail the ambivalence that emerges when
Americans are asked to think about social welfare. (See also Rokeach 1973;
Reinarman 1987.)
Our results are thus consistent with observations of Americans' attitudes
over the past several decades. Our hypotheses-derived from an examination of
the literature on U.S. political culture-also led us to expect the patterns of
conflict and ambivalence found in these data. Our conclusions are thus supported
by a substantial body of empirical analyses and theoretical discussions of U.S.
politics and society in addition to the data presented here.
A second alternative interpretation focuses on our theoretical framework.
Aaron Wildavsky (1987a, 1987b) has recently proposed a typology of political
cultures, one which he designates as egalitarian culture. Egalitarians, he argues,
are levelers who tend to like social welfare measures because they make citizens
more equal, but who dislike the government that undertakes these measures be-
cause it is hierarchically structured. Hence, they end up praising and attacking
the welfare state simultaneously.
Thus, whereas we attribute liberal ambivalence to conflict within the politi-
cal culture, Wildavsky attributes it to conflicts within an ideology. Which expla-
nation is correct cannot be determined with the data presently available.'8 But
either way the main conclusion remains that social welfare liberals are more
deeply conflicted over this issue and that they can express their ambivalence in
relatively abstract and principled terms.
It is important to recognize that liberals may not be more internally con-
" It should be noted, however, that there is little evidence in our data that supporters of social
welfare programs are strong egalitarians who wish to impose a more equal distribution of resources
in U.S. society.
flicted than conservatives on all issues. In other issue domains, it may turn out
that conservatives are the ones who must reconcile competing cultural values.
(See Tetlock 1986 for evidence on this point.) Civil liberties may be a case in
point. Efforts to justify limiting freedom of expression come into conflict with
the libertarian tradition in this country (see McClosky and Zaller 1984). Here,
civil libertarians may be better able to bring their position into line with the basic
principles of the political culture, and those who wish to limit freedom of ex-
pression may be the ones who experience ambivalence most deeply and who
must offer pragmatic or nonprincipled reasons to justify their preferences.
There are clearly some ways around the ambivalence and conflict we have
described. One partial solution evident in our data is the tendency of supporters
of social welfare to draw upon more concrete justifications for their positions
while opponents of social welfare depend much more on abstract values and
principles. This tends to diminish the total impact of the most serious conflicts
for each group. A second solution is the ability of people to shape general values
to particular political positions. So, while opponents of social welfare programs
extol the value of hard work, welfare state supporters argue for greater govern-
ment activism so that all individuals will have a chance to work. Similarly, while
the antigovernment attitudes of social welfare conservatives often take the form
of principled commitment to limited government and the ideal of laissez-faire,
liberals emphasize instead the problems of bureaucracy and the practical limits
to what government can accomplish-emphases that leave open the possibility
that government can sometimes play a positive role in promoting social welfare.
It is perhaps an indication of the power of cultural values that they can be simul-
taneously accepted overwhelmingly by the public and interpreted in various ways
congenial to peoples' partisan political beliefs.
The Measurement of Complex Attitudes
As we have seen, the pilot study data suggest that many people do not
possess a single attitude toward the welfare state. Rather, they seem to possess a
range of only partially consistent reactions to it. Complex attitudes of this type
are, we believe, best captured by a combination of open-ended and closed-ended
interview techniques, as we have attempted in this paper. Studies that rely exclu-
sively on closed-ended questions can, as Free and Cantril (1968), in particular,
have shown, reach substantive conclusions similar to our own, but they cannot
tap the underlying structure of these attitudes as well as studies that employ a
more diverse measurement strategy. Since, moreover, public attitudes toward the
welfare state are probably not uniquely complex, it seems likely that studies of
other types of attitudes can also profit from a mixed measurement strategy.
Although the open-ended probes used in this study have not produced the
wealth of information of a Robert Lane-type interview, they have, with the aid
of a suitably elaborate coding scheme, generated quite a lot of informa-
tion-much more, certainly, than can standard closed-ended questions alone. A

major advantage of our approach is that such information is available for a rep-
resentative cross-section of the public, thereby eliminating the concerns about
generalizability that arise in depth studies. The large increase in the number of
cases available in our approach also makes it possible to detect and analyze
relationships among types of beliefs. Finally, our use of strictly standardized
probes to elicit responses and professional coders to tally the results help to allay
the concern, always present with data from depth interviews, that investigators
may have inadvertently prejudiced their findings either by injecting their own
attitudes into the interview process or by giving disproportionate weight to some
kinds of individuals or some kinds of comments.
This is not to say that open-ended survey techniques should replace depth
interviews. Depth interviews have proven immensely useful in generating in-
sights about the nature of public opinion and will surely continue to do so. But
the techniques employed here should prove a useful supplement to the older
method.
Although we have focused largely on common patterns across the two issue
questions, some significant differences should not be ignored. One important
lesson to emerge from this study is that people may appear substantially more or
less "ideological" depending on what they are asked to talk about. In particular,
our respondents were far more inclined toward the use of abstract and ideological
terms when discussing job guarantees than discussing government services.
Thus, the "level of conceptualization" (Campbell et al. 1960) that citizens bring
to the evaluation of politics may vary as much across types of issues and situa-
tions as it does across types of individuals.
The public's more abstract and principled response to the job guarantees
item raises another key measurement issue: the jobs item measures what Free
and Cantril have called the "philosophical" dimension of the welfare state, while
the spending/services item measures the "operational" dimension of this issue.
These concerns are not simply alternative aspects of the same attitude; despite a
high intercorrelation, they also appear to represent somewhat independent res-
ervoirs of opinion. This can be seen most clearly in the differing patterns of over-
time change in the marginal distributions of the two items. From 1984 to 1986,
the percentage of people who felt that individuals should get ahead on their own
increased by 9% at the same time that the number of people favoring greater
spending on government programs increased by 7%. We can only reconcile these
apparently contradictory trends by noting that the guaranteed jobs question leads
people to think much more about abstract values and principles while the spend-
ing/services question generates thoughts about concrete programs and groups.
These two sources of attitudes toward the welfare state do not necessarily re-
spond identically to political events. Thus, over this two-year period, the public
seems to have become more philosophically conservative and more operationally

liberal at the same time, thereby intensifying the conflict that, as we have seen,
pervades the public's attitudes toward social welfare issues.
If, as we have suggested, the job guarantees item captures the public's en-
during philosophical response to the general problems of the welfare state, re-
sponses to this item ought to exhibit less over-time variation than responses to
the services item, which would more closely mirror vicissitudes in spending
priorities, the popularity of particular programs, and other short-term forces.
Data from the five election studies between 1980 and 1988 (the only period in
which both questions were asked) are clearly consistent with this expectation.
The mean absolute change (between adjacent election studies) in the percentage
of those opposing guaranteed jobs is 3.3%. The mean change in those supporting
greater services is 10.3%. In several respects, the behavior of these two issue
questions reflects the pattern of considerations we have identified in these data.
Justifying the Welfare State
Writing at a time when the progressive accomplishments of the New Deal

still seemed fragile, Hartz (1955) argued that the justification for the welfare
state in the United States rested mainly on pragmatic concerns: its proponents
emphasized the need for government action to solve a wide range of practical
problems, but shied away from linking their programs to a European-style ide-
ology of positive government and egalitarian democracy.
More than 30 years after Hartz, we find that supporters of the welfare state
still lack a clear ideological justification for their positions. Many welfare liber-
als expressed serious misgivings about the welfare state. Moreover, even the
people most friendly toward welfare state policies exhibited very little support
for greater equality of outcomes, social responsibility for guaranteeing everyone
the material necessities for a decent life, or individual entitlement to such neces-
sities. Although the effect of many social welfare policies is to move society a
bit closer to equality, supporters of those policies do not seem able to mobilize
egalitarian arguments in their behalf. In fact, our analysis was unable to solve a
key puzzle: even people who scored high on a closed-ended egalitarianism scale
made few or no references to egalitarianism in the open-ended remarks. How are
we to account for this?
It is of course possible that such egalitarian justifications do exist but that
our questions could not elicit them. Even so, given the ease with which people
could draw upon other cultural values, the inability to articulate egalitarian prin-
ciples is significant. We can suggest two alternative explanations. It is possible
that the people who do have egalitarian values (as indicated by their responses to
the closed-ended questions) are unable to articulate those principles because of
the strength of economic individualism in the United States and the absence of
clear egalitarian rhetoric.'9 It is also possible that it is the egalitarianism scale

that is at fault. Positive responses to these questions may reflect such other sen-
timents as humanitarianism or social benevolence. In this case the open-ended
data are telling the correct story. Humanitarian sentiments are widespread; egali-
tarian principles are rare.
Our data suggest, therefore, that support for welfare state policies derive
from sympathy and humanitarianism, not egalitarian principles.20 That sympathy
exists in the context of support for an individualistic system. Advocates of social
welfare programs thus acknowledge that people should work hard and govern-
ment should have limited powers but when people fail or have their opportunities
limited they should receive some measure of help. Programs to assist the disad-
vantaged are defended by a combination of the perception of the lack of equal
opportunity, sympathy for the disadvantaged, and support for specific programs.
Our analysis is thus in accord with a major conclusion of the Verba and Orren
study of elites: "American leaders, even the most egalitarian ones, [similarly]
opt for measures to increase equality that are consistent with the norm of oppor-
tunity" (1985, 257).
The power of this largely pragmatic justification for social welfare should
not be underestimated. As Ladd and Lipset (1980) have argued, despite the in-
creasingly conservative tone of recent political debate, public support for most
government welfare programs remains high. And after two full terms in office,
Ronald Reagan had remarkably little success in dismantling the welfare state.
However, as our data have shown, the values of the political culture severely
limit the development of a more encompassing ideological justification for the
welfare state. Under these circumstances it is perhaps not surprising that liber-
alism was seen as losing support among the public as the economic pie began to
expand less rapidly in the 1970s (Medcalf and Dolbeare 1985). Forced to defend
their positions in the aftermath of high inflation and economic stagnation, many
liberals have been accused of giving up on the principles of a liberal social wel-
fare state. And while it would be naive to assert any simple connection between
'9It is important to distinguish here between equality of opportunity and formal equality, on
the one hand, and egalitarianism as a principle for evaluating the distribution of resources in society.
The first two are, according to virtually all political analysts, part of the political culture; the third
has lacked articulate supporters (see Verba and Orren 1985).
20Since respondents were not asked to explain their open-ended comments, our methodology
does not permit us to examine the origins of the principles that people use to justify their positions
on social welfare policy. We therefore cannot investigate the basis for the expressions of sympathy
and humanitarianism that appear most frequently among those supportive of social welfare. Discus-
sions of the origins of altruistic sentiments have suggested that such factors as personality, socializa-
tion, and social background may explain individual differences (for a summary, see Conover 1988).
Alternatively, sympathy and humanitarianism may derive from Judeo-Christian ethics (Bellah et al.
1985; Reichley 1985; Wald 1987). If so, Judeo-Christian ethics could provide a principled basis of
support for social welfare more comparable to the role of individualism and antigovernment orien-
tations in opposition to welfare.
public opinion and government policy, it is interesting to note that in comparison

to other Western democracies the United States ranks relatively low in the extent
of the welfare state and the distribution of wealth in society (Verba and Orren
1985; Smith 1987) as well as in popular support for the principles of the welfare
state (Smith 1987).
Opponents of the welfare state are better served by the political culture.
Conservatives stand ready to cite traditional beliefs in individualism and limited
government to provide a consistent basis for their critiques of the growth of the
welfare state. These critiques are not easily dismissed by a majority of the public
even as they stand behind many of the specific programs that continue to form
the basis of domestic policy.
Manuscript submitted 21 May 1990

Final manuscript received 20 June 1991
APPENDIX
I. Construction of the 15 master codes from the open-ended responses. The original coding
scheme for the open-ended probes was reduced to the smallest number of categories that did not
require combining substantively inconsistent comments. Most of the description provided for each
category is drawn directly from the instructions used by the coders. The full set of codes included in
each category (from the pilot study code book) is available from the authors upon request.
1. Personal comments. All statements with an explicit self-frame of reference. For example:
R's taxes are too high; R is not prejudiced; R makes it on own; R once got help; R's personal
experience with a specific program or friend or relative's experience.
2. Politics. Statements that reflect concern with political context. For example: political con-
flict would block implementation of R's preferences; politicians would never give the people what
they want on this; the lobbyists wouldn't like it; it would really shake up a lot of people.
3. Specific group references. Comments that reflect concern with the problems or needs of
particular groups. For example: elderly (senior citizens); women as group; working/single mothers;
poor people; working people; business people, industrialists.
4. Specific program comments. All simple statements (positive or negative) about particular
government programs. For example: child care; defense; jobs or job training; small business assis-
tance; social security; unemployment benefits; welfare.
5. Sophisticated program references. Comments that discuss programs at a fairly general or
abstract level. For example: references to "New Deal"-type programs; focus on how programs work;
complaints about waste or inefficiency in specific programs or areas.
6. National conditions. Comments about the state of the country or consequences for national
conditions if policies are changed. For example: economic conditions of country (high unemploy-
ment, poor business conditions, competition from abroad); natural resource problems; program
would help the whole country by spurring the economy; crime would increase/decrease.
7. Anti-American. Reference to foreign or un-American ideology; idea is like socialism or
communism; goes against the American way; it's against capitalism.
8. Taxeslbudgets. Government is spending too much; problem of unbalanced budget; com-
plaints about overall tax burden; society can't afford what it will cost; high taxes hurt the economy.
9. Positive government role. All statements that assert some sort of role for the government
in providing social welfare benefits. For example: the government is responsible for making sure that
citizens have a chance to get ahead; government responsibility for taking care of its citizens; govern-
ment responsibility for handling a problem; optimism or conviction that we could do it if we really
wanted to.
10. Antigovernment orientation. All comments critical of the role of government in social
welfare policy or of the performance of the government. For example: concept of limited govern-
ment; government is not responsible for citizens' welfare; government has gotten too big, government
bureaucracy or red tape; wastefulness of government; inherent limits to what anyone can accomplish;
problem should be handled by private sector or by market forces.
11. Individualism. All positive statements referring to the work ethic, equal opportunity, or
fairness. For example: individuals should make it on their own; people should be responsible for
themselves; welfare makes people dependent; people who are poor deserve to be poor; some people
are lazy; references to pride and self-esteem from work; equality of opportunity is important and
does exist; unfair to those who work if some people don't have to work.
12. Equal treatment. Equal rights for individuals; no one should have special privileges; all
groups should be treated the same.
13. Qualified assistance. People should make it on their own, but some may need help; any
comment indicating that R feels it is important both that individuals help themselves and that govern-
ment provide aid; helping the people who need it without giving too much help to those who don't.
14. Lack of opportunity. Equality of opportunity is important but does not exist; common
people don't have a fair chance; government should create more opportunity; some people work hard
and still don't get ahead.
15. Fairnesslequalitylassistance. All statements supporting assistance to the needy regardless
of the justification given. For example: people need the help; some people can't make it on their own;
we should help our own before spending money on foreign aid, foreign wars; social responsibility,
duty, or obligation to care for the needy; individuals are entitled to have enough to live decently;
economic equality; wealth should be spread around more.
II. Construction of the elaborated categories for individualism, humanitarianism, and antigov-
emnment orientations.
1. Total individualism. Sum of categories 11, 12, 13, and 14.
a. Equality of opportunity. Equality of opportunity does exist; everyone gets the same chance
at the start of the race, but you expect some to finish ahead of others.
b. Fairness. Unfair to those who work if some people don't have to work; welfare cheats;
many people are on welfare who don't deserve to be.
c. Workleffort. Individuals should make it on their own; people who don't work don't deserve
help; some people are lazy.
d. Value of work. Inherent positive value of work and achievement; it is good for people
to work.
e. Dependency. Living off handouts is bad; welfare makes people dependent; people on wel-
fare should be required to do useful work in exchange for support.
f. Competitiveness. It would erode competitiveness of the economic system; survival of the
fittest.
g. Qualified help. People should make it on their own, but some may need help; people should
try to get ahead on their own, but government should help when necessary; problem is
helping the people who need it without giving too much help to those who don't.
h. Equal treatment. Equal rights for individuals; no one should have special privileges; all
groups should be treated the same.
i. Lack of equal opportunity. Equality of opportunity is important but does not exist; common
people don't have a fair chance; government should create more opportunity; some people
work hard and still don't get ahead.
2. Total humanitarianism. Sum of categories 13, 14, and 15.

a. Equal outcomes. Economic equality; wealth should be spread around more.
b. Social responsibility. Social responsibility, duty, or obligation to care for the needy; society,
or individuals in society, have obligation to take care of those who need help.
c. Individual entitlement. Individuals are entitled to have enough to live decently; people have
a right to have enough to eat.
d. People need help. People need the help; some people can't make it on their own; we should
help our own.
e. Qualified help. People should make it on their own but some may need help; people should
try to get ahead on their own, but government should help when necessary; problem is
helping the people who need it without giving too much help to those who don't.
f. Lack of equal opportunity. Equality of opportunity is important but does not exist; common
people don't have a fair chance; government should create more opportunity; some people
work hard and still don't get ahead.
3. Total antigovernment. Sum of categories 8 and 10.
a. Limited government/laissez-faire. Concept of limited government; government is not re-
sponsible for citizens' welfare; government trying to do too much; government has gotten
too big; private charities, churches should take care of the needy rather than the government.
b. Bureaucracy. Government bureaucracy or red tape; government is inefficient; wastefulness
of government.
c. Limits. Inherent limits to what anyone can accomplish; impossibility of planning large-
scale social change.
d. Taxes/budget. Government is spending too much; problems of unbalanced budget; com-
plaints about overall tax burden; society can't afford what it would cost; taxes are too high;
high taxes hurt the economy.
III. Egalitarianism questions:

1. One of the big problems in this country is that we don't give everyone an equal chance.
2. If people were treated more equally in this country we would have many fewer problems.
3. It would be better for everyone if the distribution of wealth in this country were more equal.
4. We have gone too far in pushing equal rights in this country.
5. This country would be better off if we worried less about how equal people are.
6. All in all, I think economic differences in this country are justified.
00 Z -
0 * e *- . O
.'
*X~~~~~~~
0 X- t ^ ? o;0
R < o ^ ?~ o N > 0o
0~~~~~~~~() 0
co5 -\0' *_ t**o ^ *'

? _ to
&~~~~~~~
: 00 -
* 00~~~~~~0
F < o oo X F * -2
0 2 0 0 cu.0
;e- El E -E e 0
0 0 0 cn M. 0 Q Z CQ <E
r-
00 (ttO
1t C1 N-4 r-
~6 O 0; C1 1- oco
v~ ,6 C
00 C)
*c *O
r.2c -No N
*
00
C)
NN C N ( cl
C)
cq C) ON ci 0 0O -
CQ
ci oN ci c ixb
00
O, 06 _ - en Q 3
00 O - 0 ' Xc *
(1N N *
_' NN 0 O\ o O O~ N C) oI
C) )
ci 00m 0 o 00 - _ ~ 00 Q
ci%O 0 O~~~~~ \O . c E0
E
0 0~~~~~~~~~~~~~0
C
o O _ N ,,e, t . z~~~~~~~~~~~~~~~~C
*0~~~~~~ E 0gXEE'yE'
t c 00 00 cli ; ;
00
0 rr 0 mo. - 0 ( to
IS,
MO
01o.00 J 0 c -
, - 0 - oo - - m o
t~~ ~~~ 0 tf X 6 q IC ON C1 F _
C~~~~~~~~~~~~Nc
00 o6~~o - 0 -C
~~~~~~~~~
.0 .E 00
Cu 0X t'I=c1 3) aI-0'
mE~~~~~~~~~~~~
>~~~~~~~~~~~t o00Z00-o-F a0
ta~~~~~~ M
0 0 tf0 0 m 0X
,l W 00- Cl f 0 r4
,~ r6 ~ 0oc ~ V4 O4 (00
f 00 - Clr
W) CN C14 W0' N Cl C.
0' -0 ci Cl 00 0' 4 N';
010
Cl - 0 N 0 NMA \0-u C
C3Cl -wi 11
C)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.
REFERENCES
Baylin, Bernard. 1967. The Ideological Origins of the American Revolution. Cambridge: Harvard
University Press.
Bellah, Robert N., Richard Madsen, William M. Sullivan, Ann Swindler, and Steven M. Tipton.
1985. Habits of the Heart. Berkeley: University of California Press.
Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The American
Voter. New York: Wiley.
Conover, Pamela Johnston. 1988. "So Who Cares? Sympathy and Politics." Presented at the annual
meeting of the Midwest Political Science Association, Chicago.
Converse, Philip E. 1964. "The Nature of Belief Systems in Mass Publics." In Ideology and Dis-
content, ed. David Apter. New York: Free Press.
Devine, Donald J. 1972. The Political Culture of the United States. Boston: Little, Brown.
Ericsson, K. Anders, and Herbert A. Simon. 1984. Protocol Analysis. Cambridge: MIT Press.
Feldman, Stanley. 1988. "Structure and Consistency in Public Opinion." American Journal of Po-
litical Science 32:416-40.
Free, Lloyd A., and Hadley Cantril. 1968. The Political Beliefs of Americans: A Study of Public
Opinion. New York: Simon and Shuster.
Hartz, Louis. 1955. The Liberal Tradition in America. New York: Harcourt Brace Jovanovich.
Hochschild, Jennifer L. 1981. What's Fair? American Beliefs about Distributive Justice. Cambridge:
Harvard University Press.
Huntington, Samuel P. 1981. American Politics: The Promise of Disharmony. Cambridge: Harvard
University Press.
Kinder, Donald R. 1983. "Diversity and Complexity in American Public Opinion." In Political
Science: The State of the Discipline, ed. Ada Finifter. Washington, DC: American Political
Science Association.
Kinder, Donald R., and Lynn M. Sanders. 1987. "Pluralistic Foundations of American Opinion on
Race." Presented at the annual meeting of the American Political Science Association,
Chicago.
Kinder, Donald R., and David Sears. 1985. "Political Behavior." In The Handbook of Social Psy-
chology, 3d ed., ed. Gardner Lindzey and Eliot Aronson. Reading, MA: Addison-Wesley.
Kluegel, James R., and Eliot R. Smith. 1986. Beliefs about Inequality. New York: Aldine De
Gruyter.
Ladd, Everett Carll, and Seymour Martin Lipset. 1980. "Public Opinion and Public Policy." In The
United States in the 1980's, ed. Peter Duignan and Alvin Rabushka. Stanford: Hoover
Institution.
Lamb, Karl A. 1974. As Orange Goes. New York: Norton.
Lane, Robert E. 1962. Political Ideology. New York: Free Press.
. 1973. "Patterns of Political Belief." In Handbook of Political Psychology, ed. Jeanne Knut-
son. San Francisco: Jossey-Bass.
Lipset, Seymour Martin. 1979. The First New Nation. New York: Norton.
McClosky, Herbert, and John Zaller. 1984. The American Ethos. Cambridge: Harvard University
Press.
Medcalf, Linda J., and Kenneth M. Dolbeare. 1985. Neopolitics: American Political Ideas in the
19.80's. New York: Random House.
Potter, David M. 1954. People of Plenty. Chicago: University of Chicago Press.
Reichley, A. James. 1985. Religion in American Public Life. Washington, DC: Brookings
Institution.
Reinarman, Craig. 1987. American States of Mind. New Haven: Yale University Press.
Rokeach, Milton. 1973. The Nature of Human Values. New York: Free Press.
Sennett, Richard, and Jonathan Cobb. 1972. The Hidden Injuries of Class. New York: Random
House.
Smith, Tom W. 1987. "The Welfare State in Cross-National Perspective." Public Opinion Quarterly.
51:404-21.
Tetlock, Philip E. 1984. "Cognitive Style and Political Belief Systems in the British House of Com-
mons." Journal of Personality and Social Psychology 46: 365-75.
. 1986. "A Value Pluralism Model of Ideological Reasoning." Journal of Personality and
Social Psychology 50:819-27.
Tetlock, Philip E., Jane Bemzweig, and Jack L. Gallant. 1985. "Supreme Court Decision Making:
Cognitive Style as a Predictor of Ideological Consistency of Voting." Journal of Personality
and Social Psychology 48:1227-39.
Verba, Sidney, and Gary R. Orren. 1985. Equality in America. Cambridge: Harvard University
Press.
Wald, Kenneth D. 1987. Religion and Politics in the United States. New York: St. Martin's Press.
Wildavsky, Aaron. 1987a. "Choosing Preferences by Constructing Institutions: A Cultural Theory
of Preference Formation." American Political Science Review 81: 3-22.
. 1987b. "The Media's 'American Equalitarians.'" Public Interest 88:94-104.
Wills, Gary. 1971. Nixon Agonistes. New York: Knopf.
Zaller, John. 1986. "Pretesting Information Items on the 1985 NES Pilot Study." Report to the Board
of Overseers for the National Election Studies.
Zaller, John, and Stanley Feldman. 1987. "Frames of Reference and the Survey Response." Pre-
sented at the annual meeting of the American Political Science Association, Chicago.

Feldman PoliticalCultureAmbivalence 1992

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Feldman PoliticalCultureAmbivalence 1992

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Feldman PoliticalCultureAmbivalence 1992

Uploaded by

Copyright:

Available Formats

The Political Culture of Ambivalence: Ideological Responses to the Welfare State

Author(s): Stanley Feldman and John Zaller

Stanley Feldman, State University of New York at Stony Brook

resolution to the Converse-Lane controversy over political ideology. We shall not

The Prevalence of Ideology in the United States

Hochschild and Reinarman show, it is dangerous to equate lack of consistency

The Intersection of Political Culture, Ideology, and the Welfare State

Weak though ideological conflict may be in the United States, it obviously

Studying Ideological Conflict and Ambivalence

Data and Methods

The Nature of Beliefs about Social Welfare Policy

Table 1 shows the distribution of responses across the 15 master categories

Spending/Services Jobs/Standard of Living

comments 5.9 9.5 13.5 7.3 7.7 9.7

2. Politics 13.7 4.7 2.3** 1.5 1.8 3.2

N= 51 169 133 137 169 62

Prosocial and Antisocial Welfare Arguments

We have hypothesized that social welfare liberals would experience more

The Extent of Value Conflict

Table 2. Levels of Conflict for Abstract and Concrete Responses

Spending/Services Jobs/Standard of Living

The Underlying Values: A Closer Look

Our analysis of Tables 1 and 2 has suggested the existence of considerable

differences between those who are consistently prosocial or antisocial welfare on

Egalitarianism, Sophistication, and Support for Social Welfare

particular, persons scoring high on egalitarianism showed, as expected, rela-

000 ON"C C O ~00 00 (ON Lf)

Conclusions and Implications

Ideological Responses to the Welfare State

bias-the ideology of classical liberalism. This set of ideas, according to many

There are alternative ways to interpret our findings on liberal ambivalence

The Measurement of Complex Attitudes

tion-much more, certainly, than can standard closed-ended questions alone. A

seems to have become more philosophically conservative and more operationally

Justifying the Welfare State

Writing at a time when the progressive accomplishments of the New Deal

clear egalitarian rhetoric.'9 It is also possible that it is the egalitarianism scale

public opinion and government policy, it is interesting to note that in comparison

Manuscript submitted 21 May 1990

2. Total humanitarianism. Sum of categories 13, 14, and 15.

III. Egalitarianism questions:

co5 -\0' *_ t**o ^ *'

You might also like