The Priming of Religion in Political Attitudes:
The Role of Religious Programming
The United States is at the same time a secular and religious nation, yet the
numerous studies of religion and social life, particularly the mass media, do
little to provide a coherent picture. This research area suffers from
"scattered studies, simple and simplistic measures, ambiguous and apparently
contradictory findings, incomplete explanations, and a general lack of
thoughtful and coherent explanations" (Buddenbaum & Stout, 1996, p. 30).
Mass media and religion are often at odds. For example, there has long been a
stormy relationship between Christian fundamentalists and the mainstream press,
one illustrated by the 1993 Washington Post front page article that described
such persons as "largely poor, uneducated, and easy to command" (reported in
Barrett, 1993). A large number of media observers and religious leaders have
criticized the press for providing a distorted picture of religion and religious
beliefs when religion plays such an important role in the lives of many in
society (Willey, 1996; Silk, 1995). And yet, religion leaders point to the
media as an important way to get their message across. Many Christians reported
feeling the need to "cleanse" themselves after viewing network television news,
seeking out such programs as Pat Robertson's 700 Club, which includes a newscast
and special features that link religion and politics (Barrett, 1993).
My research investigates one aspect of religion and mass
communication--exposure to religious television and radio broadcasts. Using the
concept of priming as a theoretical perspective, I analyze national survey data
to demonstrate how such programming increases the likelihood of religion playing
a role in the formation of political and social attitudes, even when controlling
for other demographic, political, and religious factors.
Religion, Politics, and the Media
There is a significant segment of American society "for whom religious beliefs
have become politicized" (Miller & Wattenberg, 1984, p. 315). In many ways,
this movement resembles other special interest groups that use the mass media,
direct mail, and other tools in which to mobilize and persuade. The so-called
"electronic church" emerged largely through the efforts of televangelsts.
Persons scoring high in media religiosity were found to be older, female, less
educated, and blue-collar (Stacey & Shupe, 1982). Exposure to such programs did
not challenge the traditional church service and, indeed, Stacey and Shupe's
(1982) analysis revealed a strong positive correlation between the two,
suggesting that many such programs are preaching to the converted, a line of
reasoning similar to the selective exposure results discussed above. In
addition, many report using such programs as another way to associate with
religious others (Alexander, 1994; Hoover, 1988).
Most work on religion and traditional mass media use finds a positive
relationship between attendance to the former and exposure to the latter. Those
who regularly attend worship services also are more likely to read newspapers
(Sobal & Jackson-Beeck; Westley & Severin, 1964), although differences emerge
when considering religious orientations among Catholics and Protestants (Rigney
& Hoffman (1993). Buddenbaum's (1996) community integration perspective found
that such integration leads to newspaper reading among mainline Catholics and
Protestants, but that the findings were mixed for Christian fundamentalists. In
addition, there appears some form of selective exposure is at work for many
fundamentalists when it comes to content that challenges their religious
perspective (McFarland & Warren, 1992).
A survey by Buddenbaum (1981) found weak but significant associations between
viewing religious programs and number of demographic factors. Viewers were
older, more likely to be female, have a lower income, and less education than
those who did not view such programming. Viewing was most closely associated
with the need to avoid feeling lonely and to "know oneself." These results,
given their time frame and the state of religious programming at the time,
resemble that of talk radio. Early research found talk radio listeners to be
more socially isolated and alienated, although more recent studies have
challenged that view and presented a more politically efficacious and active
audience (Hollander, 1996).
Political content has come to be a major factor in religious programming,
particularly on such television shows as Pat Robertson's The 700 Club. Studies
by Abelman and his colleagues (Abelman, 1994; Abelman & Pettey, 1988; Abelman &
Neunendorf, 1985) demonstrate an increase in the political and social content of
such programs, often at a cost to the proportion devoted to strictly religious
topics. Topics covered on Robertson's The 700 Club are "more social than
religious, and more religious than political (Abelman, 1994, p. 890). This
analysis found that over one-third of the program's content in 1992 consisted of
social issues and just over 15% was devoted to political issues, a decrease in
political content from previous years. Abelman and Pettey (1988) found The 700
Club devoted 19% of content to political issues in 1983 and 29% in 1986. Other
programs during the same span greatly increased the political portions of their
content. In addition, viewers often report tuning in to deal with changes in
society or to increase the political influence of religious people (Alexander,
1994).
Priming
Priming "describes the effects of prior context on the interpretation of new
information" (Fiske & Taylor, 1991, p. 237). Recent or frequently activitated
ideas come to mind more easily, making them more important in the processing of
stimuli. These cognitive shortcuts allow people to simplify the decision-making
process, to use a subset of all available information from an internal sample of
convenience. Priming has been used in a number of contexts, from stereotyping
(i.e., McKenzie-Mohr & Zanna, 1990) to the use of bumpers and teasers in
newscasts (Schleuder, White, & Cameron, 1993). News coverage, for example, can
influence the criteria by which political leaders are judged (Iyengar & Kinder,
1987). Iyengar and Simon (1993) found that extensive coverage of the Gulf War
resulted in foreign policy assessments overriding economic assessments in
evaluations of President George Bush. Involvement, knowledge, or expertise in a
subject domain can also mediate the impact of priming, at least in politics
(Krosnick & Brannon, 1993).
In a similar fashion, religion coverage along with strong religious ties could
result in an additive effect of the priming of religion in political contexts
and the nurturing of religion as a chronically accessible construct in decision
making and attitude formation (Bargh, Bond, Lombardi, & Tota, 1986; Higgins &
King, 1981). However, few if any studies have investigated the priming of
religion as a political construct. While discussing the limitations of their
content analysis, Abelman and Pettey (1988) note that "considering the success
of evangelical programs in gathering contributions from its viewership, concerns
regarding religious television's persuasiveness and possible impact on viewers'
political orientations are also worth considering" (p. 319).
Religious leaders have effectively translated morality and social values into
political capital, but have been less effective when it comes to connecting
religion to more abstract and less value-laden issues. Miller and Wattenberg
(1984), for example, found strong relationships between religious attitudes and
political attitudes on abortion, school prayer, and equal roles for women.
However, relationships between these same religious attitudes and such issues as
government services, jobs, and inflation versus unemployment concerns were
negligible. As Wald (1989) notes, religion is not "relevant to all
controversies" (p. 86).
My research extends this to the role of religious programming as a priming
device, either in an episodic fashion or for making religious attitudes
chronically accessible and applied in the political domain. Using a similar set
of issues as that of Miller and Wattenberg (1984) and controlling for a number
of demographic, political, and religious variables, the following two hypotheses
are generated:
H1: Exposure to religious programming will prime the importance of
religion in political attitudes.
H2: Priming of religion will be more likely to occur on issues where a
clear connection has been made on values and threat to religion than on
more abstract or unrelated issues.
A large proportion of recent scholarly research into religion and politics has
focused on the rise of Christian fundamentalists, either through such
personalities as Jerry Falwell's and Pat Robertson, or through such
organizations as the Moral Majority and Christian Coalition. Though sometimes
slowed by religious divisions and political blunders (Fowler & Hertzke, 1995),
evangelicals have been credited with extending the role of religious values to
public policy (Wuthnow, 1983). Therefore, it seems likely that Christian
fundamentalists would be more open to priming, generating the following
hypotheses:
H3: Exposure to religious programming will prime the political attitudes
of Christian fundamentalists more than that of mainline Catholics and
Protestants.
H4: The relationship in H3 will be more likely to occur on value-laden
issues than on issues less tied to religion or values.
Method
Data to test these hypotheses were drawn from a national telephone survey of
1,975 adults conducted May 31 to June 9, 1996 by the Pew Research Center for the
People and The Press.[1] The sample was a random digit sample of telephone
numbers selected from telephone exchanges in the continental United States in
proportion to their size. This survey includes a large number of items
measuring political and religious attitudes, making it useful to test the
hypotheses above.
Dependent Variables
Priming has to do with making temporarily accessible a way of looking at the
world and influencing how new information is processed or stored in memory.
Here, religion must be shown as a key factor in how one forms or maintains a
political attitude, even after other possible explanations are controlled for.
In the survey, respondents were asked their attitudes on seven public affairs
issues: The death penalty for convicted murderers, Clinton's sending of troops
to Bosnia, gay and lesbian weddings, abortion, government handling of the
environment, and women in the workplace.[2]
After each attitudinal item, respondents were asked:
Which one of the following has had the biggest influence on your
thinking on this issue?
A personal experience...
The views of your friends and family...
What you have seen or read in the media...
Your religious beliefs...
Your education...
Or something else?
To measure the importance of religious beliefs in attitudes, the results above
were coded as 1=religious beliefs, 0=otherwise, thus collapsing all other
responses except for missing data or refused responses into the "0" category.
Independent Variables
Religious Beliefs. A number of factors could lead to responding "religious
beliefs" above, from demographic factors to the importance religion plays in
one's life. Measures of "born-again" often disappear in multivariate analysis
when other religious measures are included (Dixon, Lowery, & Jones, 1992). Here
I look at three indicators of religiosity: (1) How often one attends church; (2)
How often one prays; and (3) how important religion is in daily life.[3]
Exposure to Religious Programming. Respondents were asked, "Which of the
following, if any, do you personally do...Watch religious TV programs or listen
to religious radio shows?" Responses could be yes (coded as 1) or no (coded as
0).
Political Ideology. A single standard question asks respondents to identify
themselves on a 1-to-5 scale from very conservative to very liberal.
Demographics. Six demographic variables were used here: Sex, age, income,
race, education and the type of community in which one lives. Sex and race are
dummy coded for regression purposes with sex as 1=female, 0=otherwise and race
as 1=white, 0=otherwise. Residence is a 1-to-4 scale with low scores
representing living in an urban area and high scores representing a residence in
a rural area.
Results
The national survey resulted in 1,975 completed interviews. The average age of
respondents was 43.0 (SD = 16.7). In terms of religious identification, the
majority of respondents said they were Christian (83.7%). Among Christians,
nearly all said they were Protestant (61.9%) or Roman Catholic (29.0%).
Slightly more then 40% of Christians identified themselves as "born-again."
Just over 57% of all respondents said that religion plays an important part in
their daily lives.
Influence of Religion
As anticipated, the influence respondents said religion plays on their
political attitudes varies according to the issue. Religion played a role in
attitudes on such issues as gay marriages (37.4%) and abortion (28.2%, see Table
1). However, only 2.9% of respondents reported that religion was the biggest
factor in their attitudes about the country's handling of the environment.
The differences among the issues presented above suggests that the value-laden
nature of an issue is a factor in how important religion is in attitude
formation.[4] To confirm this, a principal components factor analysis with a
varimax rotation was conducted on the issues above. The results (Table 2)
indeed show a two-factor solution along valence lines, with the first factor
being
made up of the questions on the needy, the environment, women in society, and
troops sent to
Table 1: Percent of all respondents who say religion had the biggest
influence on attitudes about political/social issues (N=1,975).
Political/Social Issues
% Say Religion
Biggest Influence
Abortion should be legal and available
28.2
Government's help for needy Americans
6.1
Country should do what can for environment
2.9
Society improved with women in workplace
4.1
Death penalty for convicted murderers
18.3
Clinton sending troops to Bosnia
5.6
Allowing gays and lesbians to marry
37.4
Note. Specific survey items in the Endnotes section. Missing or refused data
excluded from analysis.
Table 2: Factor Analysis of Individual Issues
Factor 1
Factor 2
Environment
.77
Women in Workplace
.69
Helping the Needy
.61
Troops to Bosnia
.59
Gay Marriages
.80
Abortion
.77
Death Penalty
.26
.65
Eigenvalue
2.5
1.2
% Variance Explained
35.1%
16.7%
Note. Factor coefficients under .25 excluded from table.
Bosnia, issues without largely indirect ties to religious attitudes.[5] The
second factor is comprised of questions on gay marriages, abortion, and the
death penalty.[6]
Due to this clear division, the two sets of items were used create a pair of
indices measuring overall likelihood to use religion as a factor in
political/social attitudes. In both instances, the issue responses were summed.
This results in the low-valence index ranging from 0 (religion never the biggest
influence) to 4 (religion named in all four issues) and the high-valence index
ranging from 0 to 3.[7]
Religiosity
Recall that the three items were used as measures of religiosity (how often
pray, how often attend worship services, and how important religion is in one's
life). To simplify further analysis, these items were subjected to a principal
components factor analysis to see if the items could be combined into a single
measure of religiosity. One factor emerged explaining 70.5% of the variance
with an eigenvalue of 2.1. Because one item is on a different metric than the
other two, all the items were standardized and these standardized variables were
then summed and averaged to create a religiosity index.[8] The index
demonstrated strong internal consistency (Cronbach's Alpha = .75).
Exposure to Religious Programming
Among all respondents, 43% reported some exposure to religious programming
while 57% said they did not use such programming. As one would expect, persons
who use religious programming score higher in religiosity (M = .43) than those
who do not (M = -.38, t(1966) = -23.6, p<.001). Bivariate analysis shows that
persons who use religious programming are older, less educated, of lower income,
more likely to be male, politically conservative, and white.
Tests of the Hypotheses
Hypothesis 1 predicted that exposure to religious programming would prime the
use of religion as a factor in political attitudes, while Hypothesis 2 predicted
this relationship would be most powerful on high-valence issues. T-tests of the
exposure item proved to be significantly associated with religion being
identified as the most important reason for an attitude on the low-valence
issues, with those who do not use such programs (M = .09) less likely to name
religion as those who do use such programs (M = .31, t(1966) = -8.6, p<.001). A
stronger relationship in the same direction is found for the high-valence issues
(M = 0.5 to M = 1.2, t(1962) = -16.7, p<.001).
The next stage in analysis was to enter socio-demographic, political, religious
controls in order to test whether exposure to religious programming explains any
additional variance.[9] Table 3 provides the unstandardized regression
coefficients (standard errors in parentheses) of two hierarchical multiple
regression analyses, one on each of the two valence issues. As the table shows,
in a controlled analysis the socio-demographic variables play virtually no role
in the likelihood of religion playing a dominant role in forming or maintaining
a political attitude. The only exception is the weak but statistically
significant effect of education on the high-valence issues, with greater
education predicting less use of religion in political attitudes. Political
ideology plays no significant role in the low-valence issues but is related to
use of religion in the high-valence issues, with the more liberal the
respondent, the less likely the use of religion. As expected, religiosity is
strongly predictive of the use of religion in both instances, although it is
much more robust in the case of the high-valence issues.
Table 3: Hierarchical Multiple Regression on Priming
of Religion in Political Attitudes
Low-Valence
Issues
High-Valence
Issues
Sex
-.02 (.03)
-.06 (.04)
Age
.00 (.00)
.00 (.00)
Education
-.01 (.01)
-.03a (.01)
Race
-.01 (.03)
.07 (.06)
Income
-.01 (.01)
-.01 (.01)
Residence
.04 (.02)
.00 (.04)
R2
.02c
.04c
Political Ideology
-.03 (.01)
-.16c (.02)
Religiosity
.12c (.02)
.46c (.03)
R2 Change
.05c
.24c
Religious Programs
.08b (.03)
.28c (.05)
R2 Change
.00b
.01c
TOTAL R2
.07c
.29c
a p<.05 b p<.01 c p<.001
Note. Entries are unstandardized regression coefficients (standard errors in
parentheses). Analysis entered in three blocks: demographics, political
ideology and religiosity, and then exposure to religious programming. Residence
is measured on a 1-to-4 scale with low being urban and high being rural.
Political Ideology is scored as high=liberal. Race is a dummy variable with
1=white, 0=otherwise.
Even after these controls, exposure to religious programming remains a
significant factor in priming the use of religion in political attitudes,
providing support for Hypothesis 1. On low-valence issues, the result is
statistically significant but relatively weak (_ = .08, p<.01), but on
high-valence issues the effect is more robust (_ = .28, p<.001). Examination of
the R2 for each model shows that religiosity, religious programming and
political ideology play a much more important role on the high-valence issues as
compared to the low-valence issues, supporting Hypothesis 2. Hypothesis 3
predicts that the above relationships will be more likely to occur for Christian
fundamentalists as compared to others, while Hypothesis 4 predicts that among
fundamentalists, exposure to religious programming will be more likely to prime
religion among high-valence issues.[10] Overall, Christian fundamentalists were
much more likely to name religion as an important factor in their political
attitudes on both the low- and high-valence issues than were mainline
Protestants and Catholics.[11] T-tests on the two priming variables reveal that
exposure to religious programming is significantly associated with the
likelihood to use religion in political attitudes among Christian
fundamentalists, mainline Protestants, and Catholics (see Table 4). The table
also dramatically demonstrates the greater likelihood to use religion among the
high-valence issues.
Table 4: T-Tests of Priming Scores on Low-Valence and High-Valence Issues Among
Religious Affiliations by Those Exposed or Not Exposed to Religious Programming
Christian Fundamentalists
Low High
Valence Valence
Mainline
Protestants
Low High
Valence Valence
Catholics
Low High
Valence Valence
No Religious
Program Exposure
.13
(.36)
1.0
(1.0)
.09
(.37)
.41
(.68)
.06
(.23)
.59
(.83)
Religious
Program Exposure
.40
(.80)
1.7
(1.1)
.26
(.64)
.98
(1.0)
.25
(.62)
1.2
(1.0)
Mean Difference
of Column Entries
.27
.71
.17
.56
.19
.60
Note. Entries are mean scores with standard deviations in parentheses. All
column entries (differences between no religious programming exposure and
exposure to such programs) are significantly different at the .001 level by
one-tailed t-test. The low-valence index ranged from 0-to-4 and the
high-valence index ranged from 0-to-3.
Table 5: Hierarchical Multiple Regression on the Low- and High-Valence Issues by
Religious Affiliation
Christian
Fundamentalists
Low High
Valence Valence
Mainline
Protestants
Low High
Valence Valence
Catholics
Low High
Valence Valence
Sex
.04
-.08
-.05
.00
-.02
-.12
Age
.00
.00
-.00
.00
.00
.00
Education
-.04
-.07
-.03
-.04
-.01
.03
Race
-.10
-.04
-.01
.00
.02
.19
Income
-.01
-.01
-.02
.01
.00
-.02
Residence
.06
.20
-.01
.06
.02
.01
Political Ideology
-.06
-.25c
.02
-.16c
-.03
-.11a
Religiosity
.21b
.61c
.13b
.43c
.06b
.42c
Religious Programs
.12
.29a
.04
.11
.07
.29b
R2 Change
.00
.01a
.00
.00
.01
.02b
TOTAL R2
.08b
.30c
.07c
.23c
.06b
.22c
a p<.05 b p<.01 c p<.001
Note. Entries are unstandardized regression coefficients (standard errors in
parentheses). Analysis entered in three blocks: demographics, political
ideology and religiosity, and then exposure to religious programming. Residence
is measured on a 1-to-4 scale with low being urban and high being rural.
Political Ideology is scored as high=liberal. Race is a dummy variable with
1=white, 0=otherwise.
Again, multivariate analysis is conducted to control for the explanatory power
of demographic factors, political ideology, and religiosity. As Table 5 shows,
there is mixed support for Hypotheses 3 and 4. Exposure to religious
programming primes religion for Christian fundamentalists only on high-valence
issues (_ = .29, p<.05). Exposure is not a factor for mainline Protestants,
while Catholics are also primed to consider religion, but again only for
high-valence issues (_ = .28, p<.01).
The similarities between Catholics and Christian fundamentalists raises the
likelihood that abortion is the key issue drawing both together. To test this
possibility, logistic regressions similar to those discussed in an earlier
endnote were again conducted on the individual high-valence issues (abortion,
death penalty, and gay marriages) among the three religious groups. Table 6
shows the results. This table was simplified for reporting purposes, giving
only the unstandardized logistic coefficients for the effect of religious
programming exposure on likelihood to name religion on the individual issues,
although all the control variables entered in analyses above were also used
here. As the table shows, among Christian fundamentalists, exposure to
religious programming primes religion only in the case of abortion. However,
among Catholics, religious programming primes religious attitudes on gay
marriages and the death penalty (near the
Table 6: Logistic Regression of Exposure to Religious Programming
on the Individual High-Valence Issues by Religious Affiliation.
Christian
Fundamentalists
Mainline
Protestants
Catholics
Abortion
.68b
.34
.09
Death Penalty
.52
.15
.58a
Gay Marriages
.11
.20
.76b
a p<.10 b p<.05 c p<.01
Note. Entries are unstandardized logistic coefficients. For each religious
group, the complete model was entered first (demographic factors,
ideology, and religiosity). The entries represent the remaining explanatory
power of the religious programming exposure item.
traditional level of statistical significance). The results here demonstrate
that the priming effects of religious programming differ depending on religious
affiliation and the specific political issues in question.
The Consequences of Priming
Finally, I explore the consequences of this priming in terms of attitudes about
political issues. I must emphasize that this work is in progress, and below I
present a simplified causal model that excludes the demographic and political
factors above, most of which were not-significant, and includes only
religiosity, exposure to religious programming, priming of religion on an issue,
and attitude on the abortion issue.[12] Figure 1 shows the model for abortion
among all respondents and the standardized path coefficients (high on the
abortion question represents opposition to abortion). Religiosity is a
significant predictor of all three of the other variables in the model. The
standardized regression coefficient (beta) of .30 between religiosity and
priming of religion on attitudes about abortion represents the chronic
accessibility notion, that religiosity itself makes religion more important in
forming and maintaining political attitudes. The beta of .14 between exposure
to religious programming and priming represents the ability of such programs to
prime the importance of religion in attitudes about abortion. All coefficients
are significant at the .05 level except for the predicted relationship between
exposure to religious programming and attitudes about abortion. This simplified
model suggests that, at least for abortion, both the priming by religious
programming and the chronically accessible concept of religion can both
influence attitudes.
Figure 1 Here
Discussion
That religion and politics are intertwined is nothing new. Over 150 years ago
Alexis de Tocqueville, in his classic work on America, noted that "By the side
of every religion is to be found a political opinion, which is connected with it
by affinity" (Tocqueville, 1945, p. 310). Religious programming has also been
around for a long time, although it is only relatively recently has it emerged
as a viable political factor (Wald, 1992).
The analysis reported here finds that exposure to such broadcasts can prime
the importance of religion in the formation and maintenance of political
attitudes, even when controlling for a host of demographic and political factors
and the religiosity of respondents in the national survey. Religious
broadcasts, which have become laden with social and political content, are most
likely to prime religion as a factor on issues more closely linked in the first
place to religion or core values--abortion, the death penalty, and gay rights.
While some priming occurs on other issues, such as sending troops to Bosnia, the
linkage is less stable. Clearly religious broadcasts are more successful in
priming the importance of religion, of making it more a factor in how and what
people think about an issue, when that issue is tightly bound up in basic
religious beliefs and values.
Priming appears to work most successfully on Christian fundamentalists on the
issue of abortion, while Catholics are most susceptible to priming on the issues
of the death penalty and gay rights. Mainline Protestants are the least
susceptible to priming by religious broadcasts, demonstrating once again the
schism that exists between fundamentalists Protestants and mainstream
congregations. The consequences of this priming remain to be explored, but an
early look here at the abortion issue suggests that religious broadcasts can
lead to priming, which in turn can result in greater opposition to abortion.
However, this priming is not as powerful a predictor as is religiosity, which
tends to make religion a chronically accessible construct, also influencing
attitudes about abortion. The dual nature of chronic accessibility and priming
deserves greater attention and more careful causal modeling among the various
religious affiliations. For example, the priming ability of religious
broadcasts may be overwhelmed by chronic accessibility among Christian
fundamentalists on the issue of abortion, but such programming may indeed prime
the importance of religion, and shape political attitudes, among other groups.
The research here, relying on the social psychological constructs of priming
and construct accessibility, also fall into the broad "values" theory of
explaining the growth of the religious right in America. Wuthnow (1983) argues
that the Christian right has been stimulated in part by fears of challenges to
their core beliefs and a need to tie religious beliefs to public policy.
Indeed, viewers of religious broadcasts list the need to get a Biblical position
on political and moral issues as the most important reason for viewing or
listening (Alexander, 1994). While much of this may be "preaching to the
choir," for some it appears to be priming the importance of religion in
political attitude formation and maintenance.
References
Abelman, Robert (1994). News on The 700 Club: The cycle of religious activism.
Journalism Quarterly, 77, 887-892.
Abelmen, Robert & Neunendorf, Kimberly (1985). How religious is religious
television programming? Journal of Communication, 35, 98-110.
Abelman, Robert and Pettey, Gary (1988). How political is religious television?
Journalism Quarterly, 65, 313-319, 359.
Alexander, Bobby C. Televangelism reconsidered. Atlanta, GA: Scholars Press.
Bargh, J. A., Bond, R. N., Lombardi, W. L., & Tota, M. E. (1986). The additive
nature of chronic and tempory sources of construct accessibility. Journal of
Personality and Social Psychology, 50, 869-879.
Barrett, Laurence I. (1993). The "religious right" and the pagan press.
Columbia Journalism Review, 32, 33-34.
Buddenbaum, Judith M. (1996). The role of religion in newspaper trust,
subscribing, and use for political information. In Religion and Mass Media
(Daniel A. Stout and Judith M. Buddenbaum, Eds.), pp. 123-134. Thousand Oaks,
CA: Sage.
BuddenBaum, Judith M., and Stout, Daniel A. (1996). Religion and Mass Media.
Thousand Oaks, CA: Sage.
Buddenbaum, Judith M. (1981). Characteristics and media-related needs of the
audience for religious TV. Journalism Quarterly, 58, 266-272.
Dixon, Richard D., Lowery, Roger C., & Jones, Lloyd P. (1992). The fact and
form of born-again religious conversions and sociopolitical conservatism.
Review of Religious Research, 34, 117-131.
Fiske, Susan T. & Taylor, Shelley E. (1991). Social cognition. New York:
McGraw-Hill.
Fowler, Robert Booth, and Hertzke, Allen D. (1995). Religion and politics in
America. New York: Westview Press.
Hadden, Jeffrey K., and Shupe, Anson (1988). Televangelism: Power and politics
on God's frontier. New York: Henry Holt and Company.
Higgins, E. T., & King, G. A. (1981). Accessibility of social constructs:
Information-processing consequences of individual and contextual variability.
In N. Cantor & J. F. Kihlstrom (Eds.) Personality, cognition, and social
interaction (pp. 69-122). Hillsdale, NJ: Erlbaum.
Hollander, Barry A. (1996). Talk radio: Predictors of Use and Effects on
Attitudes about Government. Journalism and Mass Communication Quarterly, 73,
102-113.
Hoover, Stewart M. (1988). Mass media religion. Newbury Park, NJ: Sage.
Iyengar, Shanto, & Simon, Adam (1993). News coverage of the Gulf Crisis and
public opinion: A study of agenda-setting, priming, and framing. Communication
Research, 20, 365-383.
Iyengar, Shanto, and Kinder, David R. (1987). News that matters. Chicago:
University of Chicago Press.
Krosnick, Jon A., & Brannon, Laura A. (1993). The media and the foundations of
presidential support: George Bush and the Persian Gulf Conflict. Journal of
Social Issues, 49, 167-182.
McFarland, S. G., and Warren, J. C. Jr. (1992). Religious orientations and
selective exposure among fundamentalist Christians. Journal for the Scientific
Study of Religion, 31, 163-174.
McKenzie-Mohr, D., & Zanna, M. P. (1990). Treating women as sexual objects:
Look to the (gender schematic) male who has viewed pornography. Personality and
Social Psychology Bulletin, 16, 296-308.
Miller, Arthur H., and Wattenberg, Martin P. (1984). Politics from the pulpit:
Religiosity and the 1980 elections. Public Opinion Quarterly, 48, 301-317.
Rigney, D., & Hoffman, T. J. (1993). Is American Catholicism
anti-intellectual? Journal for the Scientific Study of Religion, 32, 211-222.
Schleuder, J. D., White, A. V., & Cameron, G. T. (1993). Priming effects of
television-news bumpers and teasers on attention and memory. Journal of
Broadcasting and Electronic Media, 37, 437-452.
Silk, Mark (1995). How religion makes the news: Conventions of reporting.
Christian Century (November 1), 1012-1016.
Sobol, J. and Jackson-Beeck, M. (1981). Newspaper non-readers: A national
profile. Journalism Quarterly, 58, 9-13.
Stacey, William, and Shupe, Anson (1982). Correlates of support for the
electronic church. Journal for the Scientific Study of Religion, 21, 291-303.
Stout, Daniel A., and Buddenbaum, Judith M. (1996). Religion and mass media.
Thousand Oaks, CA: Sage Publications.
Tocqueville, Alexis de (1945). Democracy in America. Volume 1. Edited by
Phillips Bradley. New York: Vintage Books.
Wald, Kenneth D. (1992). Religion and politics in the United States.
Washington, DC: Congressional Quarterly Press.
Westley, B. H., and Severin, W. J. (1964). A profile of the daily newspaper
non-reader. Journalism Quarterly, 41, 45-50.
Willey, Susan (1996). Journalism and religion. Quill (January/Feburary),
28-30, 35.
Wuthnow, Robert (1983). The political rebirth of American evangalicals. In The
New Christian Right (Robert C. Liebman and Robert Wunthnow, Eds., pp. 168-187).
New York: Aldine.
Endnotes
[1] My thanks to the Pew Center for allowing me access to these data. The
analysis
and interpretations are my own.
[2] Three of the items were generated by questions asking respondents to
say whether
they strongly favor, favor, oppose, or strongly oppose the following
statements: (1) The
death penalty for persons convicted of murder; (2) President Clinton's
decision to send
20,000 U.S. troops to Bosnia as part of an international peacekeeping
force; and (3)
Allowing gays and lesbians to marry legally. Four of the items had
individual response
structures. For the abortion question: "Which comes closest to your view?
Abortion
should be generally available to those who want it...Abortion should be
available but
under stricter limits than it is now...Abortion should be against the law
except in cases
of rape, incest, and to save the woman's life...Abortion should not be
permitted at all."
The other items had the following introduction: "I am going to read you
some pairs of
statements that will help us understand how you feel about a number of
things. As I read
each pair, tell me whether the first statement or the second statement
comes closer to
your own views--even if neither is exactly right." For the Needy variable,
"The
government should do more to help needy Americans, even if it means going
deeper into
dept, OR...The government today can't afford to do much more than help the
needy." For
the Women variable: "Society has been improved because women are now
represented in the
work place, OR...Society made a mistake in encouraging so many women with
families to
work." For the Environment issue: "This country should do whatever it
takes to protect
the environment, OR..This country has gone too far in its efforts to
protect the
environment."
[3] The exact wording on these religiosity items and their response
alternatives are:
[4] This is not to suggest that values do not play a part in attitudes
about the
needy or women in society. However, religious attitudes are more often
linked to issues
such as abortion in the political arena.
[5] Obviously two of the issues can have religious overtones--helping the
needy and
the role of women in society. However, rarely do religious organizations
take positions
against the needy and even most Christian conservative organizations
recognize the role
of women in society and the workplace while supporting the idea of women
remaining at
home.
[6] These results differ significantly from Miller (1996), who found five
factors
among issues: civil liberties, morality, women's rights, economic--quality
of life, and
economic--assistance programs. Differences are no doubt due to the larger
number of issue
s used by Miller, whose aim was to explore the structural differences among
attitudes.
[7] The average score for respondents on the low-value index was 0.2
(SD=0.6), with
87% of the respondents saying religion was not a big factor in any of the
issues. On the
high-value index, the range was from 0 to 3, with an average of 0.8
(SD=1.0), with just
51% saying religion did not play a role in any of the issues making up the
index. Both
indices demonstrated adequate internal consistency, with the high-value
index having a
Cronbach's Alpha = .65 and the low-value index having a Cronbach's Alpha =
.60. The
Pearson correlation between the two indices is r=.36 (p<.001).
[8] Standardizing the variables gives them a mean of 1 and a standard
deviation of
1, equalizing their effect on the dependent variable.
[9] In addition to the analysis of the two dependent variables, I analyzed
each
issue individually through logistic regression. As the Appendix shows, the
logistic
regression provides results similar to that found by collapsing the
individual items into
two factors. Exposure to religious television and radio programming is a
significant
predictor in all issues except for government's handling of the
environment. The
strongest priming effects are seen for abortion and gay marriages, although
all
coefficients are relatively close in strength, ranging from .06 to .13 even
after the
large set of control variables).
[10] The survey first asked respondents to identify their religious
affiliation.
Among those who responded Christian and, in a follow-up question,
Protestant, respondents
were asked whether they consider themselves a "fundamentalist Christian," a
"Pentecostal
or charismatic Christian," or a "progressive Christian." These were coded
as Christian
fundamentalists while other Protestants were coded as mainline Protestants.
Catholics
were coded as such from the second question asking for affiliation beyond
being i
dentified as Christian. The results of the coding placed 15.7% of
respondents as
Christian fundamentalists, 23.3% as mainline Protestants, and 24.3% as
Catholics.
[11] Oneway analysis of variance with the Duncan multiple range test
reveals that
Christian fundamentalists are significantly more likely to name religion as
an important
factor on the low-valence issues (M = .31 compared to .18 for mainline
Protestants and .10
for Catholics) and on the high-valence issues (M = 1.5 to .71 for mainline
Protestants
and .75 for Catholics). As the mean scores suggest, mainline Protestants
were more
likely to use religion than were Catholics on the low-valence issues but
the two groups
were virtually identical on the high-valence issues.
[12] The analysis here was conducted using Amos structural equation
modeling
software (Arbuckle, 1997). Because the model was saturated (an equal
number of distinct
sample moments and distinct parameters to be estimated), it is impossible
to report a Chi
Square test of the fit of the overall model due to 0 degrees of freedom.
|
