Content-Type: text/html

TV and Religion 2
Running head: TV and Religion
Created in Whose Image? Examining Network TV's Treatment of Religion

Scott H. Clarke
423 Communication Arts Bldg.
Michigan State University
East Lansing MI 48824-1212
(517) 432-5915
<a href="/cgi-bin/wa?LOGON=A3%3Dind0309D%26L%3DAEJMC%26E%3D8bit%26P%3D6863574%26B%3D--%26T%3Dtext%252Fplain%3B%2520charset%3Diso-8859-1" >[log in to unmask]</a>

        TV and Religion 2
Abstract
39 fictional network television programs were analyzed for content relating
to religious characters. Results suggest that television characters'
religiosity is not an important consideration for television producers as a
whole. However, programs produced by Paxson Communications and Viacom; as
well as those airing on the PAX network, have significantly more religious
characters than average. This finding supports both gatekeeping theory and
Paxson's stated goal of increasing the amount of spirituality on network TV.

        TV and Religion 2
Created in Whose Image? Examining Network TV's Treatment of Religion
Research into commercial television's treatment of religion has been scant,
despite repeated charges that Hollywood is biased against the religious .
In fact, fewer than 15 published studies have treated religion as a content
variable, and some of those only as part of a larger project . The
remainder have been self-described "exploratory" studies of TV programming
, advertising  and music videos .
Furthermore, all prior research has examined religious content from an
audience effects perspective.  Several scholars have called for research
into the production of television content dealing with religion . Some have
suggested applying the "gatekeeping" model to content creators to determine
how their perceptions of religion and their audiences affect the programs
they produce . To date, however, no such studies have been conducted.
Neither has any comprehensive study of prime-time entertainment programming
– the most widely viewed – been conducted since the 1990-1991 television
season . At the time, only four broadcast networks existed. Today, seven
networks program a regular schedule nationwide, one with a stated goal of
increasing the amount of "spirituality" on television . Therefore, research
replicating and expanding earlier findings is also needed.
This study contributes to the existing literature, both by replicating and
extending earlier findings and by studying religious television content
from a structural perspective.

Gatekeeping in Television
Gatekeeping is a social theory originally developed to describe the process
by which family members select the food they eat . Communication scholar
David Manning White then borrowed the term to describe how an editor
decides which news stories to print . Since that time, hundreds of studies
on gatekeeping within news organizations have been conducted . Over time,
the focus of such research has also broadened to include "elements of the
process through which news items (for example) pass on their way from
discovery to transmission [to an audience] ."
Virts and Keeler  first suggested applying the gatekeeping metaphor to the
producers of fictional television content. They stated:
"The importance of studying television gatekeepers is rooted in the
hypothesis that the personal attitudes, beliefs, values and perceptions of
relatively few key people have direct bearing on the nature of television
programming.  It might be hypothesized further, then, that television
gatekeepers' perceptions of the significance of Judeo-Christian beliefs and
practices in society generally, their own religious beliefs and practices,
their own perceptions of Judeo-Christians and their beliefs and practices,
and their perceptions of professional and audience expectations regarding
the portrayal of religion greatly influence the ultimate portrayal of
religion in their programming ."

To date, however, this idea remains untested. This study will provide
empirical data to address this research gap.
Shoemaker argues that gatekeeping studies should be located within the
larger organizational structures of the system in which content is
produced, using the hierarchical model she developed with Reese .  The
latter model attempts to trace the various intra- and extra-organizational
influences on mass mediated content.
Taking the production team as our unit of analysis, then, an
extra-organizational influence that should be considered is the network to
which the program is sold . Relevant intra-organizational influences
include the production studio, executive producers, (program) producers and
individual writers . Each of these units has input into the content that
writers produce. Furthermore, each has its own set of routinized practices,
which "may act as surrogates or shortcuts for individual people's decisions ."

Studies of Television's Religious Content
Previous studies of religious themes on television have generally taken one
of three forms: exploration of the types of content present across
genres  or over time , case studies of specific programs , or qualitative
analyses of audience responses to religious content .
While the paucity of research into religious content on television prevents
much generalization, several findings have been replicated across studies.
First, religion is practically absent from much of television. Head found
that "content referring to religion" appeared in only 10% of his sample of
64 TV dramas in 1952 . Nearly 40 years later, the picture hadn't changed
much: less than 6% of 1462 speaking characters appearing on one week of
prime-time entertainment television in 1990 could be linked to a specific
religious group .
Similar data have been generated with respect to television advertising. A
1996 study of TV commercials found religious symbolism present in only 2%
of the spots, several of which were actually repeats of the same
advertisement .
The one genre where religious themes are abundant is music videos, where
religious symbolism was found in 38% of rock clips in 1992  and in 49% of
rock and 64% of country music videos in 1995 . On the other hand, the
context of the music videos' treatment of religion should be noted (see below).
        On the whole, then, religious characters, institutions and symbols have
been found to be a rarity in commercial television programming.
Furthermore, previous research indicates that when religion does appear, it
is often presented as irrelevant to the plot  with negative or ambiguous
characters . In addition, religion tend to be church-centered, Catholic,
and treated in stereotypical ways (e.g., clergy wearing traditional
vestments, use of superficial clichés, religious rituals, etc.) . Religious
symbolism in music videos has been found in combination with sexual imagery
18% of the time overall, and as much as 28% in the rock format .
        A third finding of previous studies is the regular and irreverent use of
"religious" language. As much as 50% of expressions referring to God,
heaven, and hell are used outside a religious context . Instead, such
expressions are usually used to express shock, surprise or disbelief on the
part of a character, or in the midst of "crisis" situations .
        On the other hand, some scholars have suggested that prime time's
treatment of religion is mainly positive , but these assertions have been
made in program-specific studies. Whether these cases are anomalies is
unclear, given the limited amount of research overall.
        Missing from the above analyses is the role content producers play in
developing what appears onscreen.  Many cultural critics and religious
leaders have gone so far as to accuse Hollywood producers of injecting
their own apathy or hostility toward religion into their programs . This is
an easy criticism, but empirical research has yet to address the issue
directly.

Research Questions and Hypotheses
Having reviewed the extant research on the portrayals of religion on
commercial television and the relevance of gatekeeping theory to the study
of entertainment routines, the following research questions and hypotheses
are advanced:
RQ1: How does the percentage of religious characters on prime-time
commercial television compare to real-world data?

RQ2: How do prime-time religious characters' denominational affiliations
compare to real-world data?

These questions are based on previous findings that religious content and
characters are virtually absent from commercial television programming .
They are also informed by the findings of Rothman and his colleagues on the
mismatch in religious backgrounds and beliefs between Hollywood elites and
their audiences . Specifically, they will test whether this mismatch leads
to content that is unrepresentative of the actual viewing public.
H1: Because the PAX network was created with the stated goal of increasing
the amount of religious content on television, it will feature more
religious characters than do competing networks.

H2: Programs produced by Paxson Communications will feature more religious
characters than do competing studios' product.

These hypotheses are based on gatekeeping theory, especially as presented
by Virts and Keeler , on the intra-organizational influences on content .
They will also test whether Paxson's claim of increasing the amount of
"spirituality" on prime-time entertainment television has actually come to
pass .

H3: Religious characters will use "religious language" more reverently than
do other characters.

This hypothesis is based on previous research which indicates that most
television characters that use religious language do so casually . These
studies imply that religious characters tend to use religious language
mainly in the context of performing a religious ritual, such as attending
services or prayer. Furthermore, this hypothesis will test critics' claims
that content producers' assumed apathy and/or hostility toward religion
should lead to an increasingly casual use of religious language on
television; especially by the supposedly religious .
H4: Religious characters will be less central to programs' storylines than
are other characters.

This hypothesis is based on previous research which indicates that
religious characters are generally portrayed as irrelevant to a program's
plot . It is designed to test whether this relationship has changed with
the addition of 3 broadcast networks and to determine the relative
influence of PAX on the centrality of religious characters .
Measurement. For all hypotheses and research questions, the unit of
analysis was a speaking character. The number of religious characters per
episode was measured at a ratio level. A religious character was defined as
one clearly identified as such by storyline, clothing or work environment
as religious (e.g., priest) or who identified themselves (or are identified
by others) as such in the dialog. Other cues included religious activity
(prayer, crossing oneself, Bible reading, etc.).
For H3, religious language was defined as expressions referring to God,
heaven, hell or religion in general; and was measured at a ratio
level.  Whenever the above expressions were used outside a religious
context or in a casual manner, they were coded as "casual use ." If these
expressions were used as a statement of genuine faith, they were coded as
"religious" in context.
For H4, religiosity was measured on a ratio scale, from "Not religious" to
"devout." Characters were coded "not religious" only if they were
identified as such by a program's dialog. A character was judged to be
"nominally religious" if he/she made some statement of belief (or was
identified as a believer by another character or by wearing religious
jewelry or clothing), but engaged in no religious behavior. Those
characters judged to be "devout" were identified in the same manner, but
had to also engage in some kind of religious behavior.
        A character's centrality to the episode plot was determined by counting
the number of program segments in which he/she appeared and dividing by the
total number of segments. Those characters appearing in 50%-100% of the
program segments were coded as "main" characters, those in 25%-49% were
coded as "secondary" characters, and those in 1-24% of segments were coded
as "background" characters.
All other variables were measured at a nominal level. These included
network, production studio, religious denomination and program type.
Sample. The population of content for this study was all fictional
television programs airing on ABC, CBS, FOX, NBC, PAX, UPN and the WB
during their respective prime time programming periods. These networks were
selected because they are available for free and over-the-air to a national
audience. Therefore, they have a larger audience share than do even the
largest cable networks.
"Prime time" on ABC, CBS, NBC and PAX runs from 7-11 p.m. Eastern on
Sundays and 8-11 p.m. Eastern on weekdays and Saturdays. FOX programs 7-10
p.m. Sundays and 8-10 p.m. the remainder of the week; UPN, 8-10 weekdays
only; and the WB, 5-10 p.m. Sundays and 8-10 p.m. weekdays. PBS was
eliminated from analysis because, as a programming service (not a network),
it only supplies only one content stream (of many) that individual stations
may air, making national comparisons of programming impractical.
Sports, news magazines and other "reality" programs, movies and non-fiction
specials were not included because their coding scheme would necessarily be
different and therefore, not comparable to the one used for fictional
shows. Also, "Diagnosis, Murder," which aired weeknights at 10 p.m.,
Mondays at 8 p.m. and Thursdays at 9 p.m. on PAX, was excluded from
analysis. This exception was made because the program is not a PAX original
production; rather, repeats from past CBS seasons. Therefore, if included,
H1 and H2 could not be fairly tested, as PAX's original programming would
be obscured by content originally produced for another network.
Programs were recorded over a one-week period, from September 23-29, 2002.
This period was chosen because September 23 was the official start of the
2002-2003 programming season and most of the content airing was original.
It was also chosen to avoid a "sweeps" rating period, when network content
tends to be atypical. This recording process yielded 78 programs with 62.5
total hours of content eligible for inclusion in the study.
Due to the unequal number of programming hours across networks, a purposive
sample was taken by network to make each one's content more representative.
Nielsen ratings data for each program aired during the sample week was used
to select the top 3 comedies and the top 3 dramas for each of the seven
networks . Since PAX does not air any situation comedies, the final sample
consisted of 18 comedies and 21 dramas. Each was rank ordered from first to
third place by network and assigned to one of three independent coders.
Each coder received one comedy and one drama per network, to preclude
systematic bias against any particular network. The programs were also
assigned to coders on a rotating basis from first- to third-ranked by
network, to preclude systematic bias by program popularity.
Data collection. Three independent coders received two training sessions,
after which they coded 10% percent of the programs (2 comedies and 2
dramas) selected for analysis. A Scott's pi was calculated, demonstrating
intercoder agreement ranging from .80-1.00. Therefore, the remainder of the
programming was randomly divided among the coders for viewing.

Findings
        This analysis of 39 episodes of prime-time commercial television
identified a total of 548 speaking characters. How these characters compare
to their real-world counterparts is the focus of the first two research
questions.
RQ1: How does the percentage of religious characters on prime-time
commercial television compare to real-world data?

Only 32 characters in the sample (5.8%) could be positively identified as
religious, replicating Skill & Robinson's  finding of 5.6%. The few
religious characters were divided nearly evenly between the "nominally"
religious (3.3%) and the "devout" (2.5%).
Census data on reported attendance at religious services  was used as a
proxy for public religiosity. This makes sense, because this measure (like
the television categorization scheme) is based on behavioral evidence of
faith. Applying the categories developed for the prime time characters,
respondents who attended services "weekly" were labeled "devout;" those who
attended "less than weekly" were termed "nominally religious;" and those
who reported "rarely to never" attending services were considered
"non-religious."
Table 1 compares the religiosity of prime time characters to the general
public – revealing a striking difference (p < .05). Over two-thirds of the
populace is "nominally religious" (24.0%) or "devout" (44.0%), a number
nearly 12 times that evidence on television. Furthermore, even this measure
is conservative, since survey respondents who attended services
occasionally were considered "non-religious."
Table 1: TV Characters' Religiosity vs. U.S. Census Data:
Religiosity
Television Characters
N=548
U.S. Population
N=281,421,908
Not Religious/Can't Tell*
   94.2
   32.0
Nominally Religious
     3.3
   24.0
Devout
     2.5
   44.0
Total:
100.0
100.0
p < .05                                      *Includes "Can't Tell" data
for TV characters and "No Answer" data for Census figures.
Source: United States Census Bureau (2001)

RQ2: How do prime-time religious characters' denominational affiliations
compare to real-world data?

Just 29 characters in the television sample could be identified with any
particular religious denomination, and two-thirds of those (19) could only
be defined as "Protestant." Table 2, below, compares these statistics to
real-world data.
Only 3.5% of prime-time characters were identified as "Protestant,"
compared to 56% of the general public  (p < .05). Interestingly, not a
single character from the present sample could be identified as Catholic, a
finding at odds with earlier studies .

Table 2: TV Characters' Religious Denomination vs. U.S. Census Data:
Religiosity
Television Characters
N=548
U.S. Population
N=281,421,908
Not Religious/Can't Tell*
   94.7
     8.0
Protestant (all)
    3.5
   56.0
All Others
    1.8
   36.0
Total:
100.0
100.0
p <
.05                                                              *Includes
"Can't Tell" data for TV characters and "No Answer" data for Census figures.
Source: United States Census Bureau

Hypothesis 1 tested how religious characters were distributed by network.
H1: Because the PAX network was created with the stated goal of increasing
the amount of religious content on television, it will feature more
religious characters than do competing networks.

Even with the small number of religious characters overall, the data
presented in Tables 3 and 4, below, demonstrate convincingly that PAX has
increased the presence of religious characters through its programs.
Network data in Table 3 indicate that over a third of all religious
characters in the sample appeared on PAX, even though the network accounts
for only 8.4% of all characters in prime time. The WB also provided
significantly more than its respective share of religious characters (21.9%
religious vs. 13.5% of all characters), while UPN had no religious
characters at all. The remaining four networks' percentage of religious
characters is significantly less than their percentage of characters
overall (p <.001). Therefore, Hypothesis 1 is supported by the data.

Table 3: Type of Characters by Network:
Network
% Religious Characters
N=32
% All Characters
N=548
PAX
    34.4
    8.4
WB
    21.9
  13.5
NBC
   15.6
  21.5
CBS
   12.5
  16.2
ABC
     9.4
  15.5
FOX
    6.3
  12.0
UPN
  --
  12.8
Total:
100.0
100.0
?2 = 35.93, df = 6, p < .001



As Table 4 indicates, 24 of the 39 programs analyzed (61.5%) featured no
religious characters whatsoever. However, the three dramas selected for
analysis from the PAX network (PAX airs no comedy programs) rank first,
third and fifth overall in the percentage of religious characters within
them (p <.001).
Table 4: Type of Characters by Program:
Program
% Religious Characters
N=32
% All Characters
N=548
Body & Soul (PAX)
  15.6
    2.6
Frasier (NBC)
  12.5
    1.8
Doc (PAX)
  12.5
    3.3
7th Heaven (WB)
   9.4
    2.6
Reba (WB)
   9.4
    1.8
Everybody Loves Ray (CBS)
   6.3
    1.5
Just Cause (PAX)
   6.3
    2.6
My Wife & Kids (ABC)
   6.3
    2.9
CSI (CBS)
   3.1
    4.9
Firefly (FOX)
   3.1
    2.7
Gilmore Girls (WB)
   3.1
    2.9
JAG (CBS)
   3.1
    4.6
John Doe (FOX)
   3.1
    3.5
Law & Order: CI (NBC)
   3.1
    3.8
The Practice (ABC)
   3.1
    2.9
All Others
--
  55.6
Total:
100.0
100.0
?2 = 97.14, df = 38, p < .001


Hypothesis 2 tested how religious characters were distributed by production
studio.
H2: Programs produced by Paxson Communications will feature more religious
characters than do competing studios' product.

Table 5, below, is a comparison of production studios' output. Programs
produced by Viacom (Paramount, Spelling, CBS Productions, and others)
contained the highest percentage of religious characters (28.1%), followed
by Paxson Communications (21.9%) and Fox (Regency, Twentieth Television,
and others) (18.8%). However, it should be noted that Viacom programs
contain 31.2% of all characters, 6 times as many as Paxson's and nearly
twice as many as Fox's. Therefore, Paxson Communications does produce a
greater percentage of religious characters overall than does any other
studio, supporting Hypothesis 2. The remaining data further demonstrate
that the number of religious characters vary significantly by production
studio
(p <.001).
Table 5: Type of Characters by Studio:
Studio
% Religious Characters
N=32
% All Characters
N=548
Viacom
     28.1
   31.2
PAX
    21.9
     5.1
Fox
    18.8
   17.9
Pebblehut
    12.5
    3.3
AOL/Time-Warner
    9.4
   24.5
Disney
    6.3
    8.8
Universal
     3.1
    3.8
All Others
--
     5.5
Total:
100.0
100.0
?2 = 32.86, df = 9, p < .001



Hypothesis 3 tested whether the use of "religious language" varied by the
religiosity of the character using it.
H3: Religious characters will use "religious language" more reverently than
do other characters.

Table 6, below, presents data collected on the language habits of speaking
characters. This analysis indicates that 80.8% of all characters don't use
any religious language at all. Furthermore, 68.8% of religious characters
used religious language to express genuine belief
(p < .001). On the other hand, there was only a slight observed difference
in the casual use of religious language overall (13.7%) and its use by
religious characters (9.4%). Therefore, Hypotheses 3 is only partially
supported by the data.

Table 6: Type of Characters by Religious Language Use:
Use of Religious Language
% Religious Characters
N=32
% All Characters
N=548
Religious
   68.8
      5.5
None
   21.9
    80.8
Casual/Non-Religious
    9.4
   13.7
Total:
100.0
100.0
?2 = 263.12, df = 2, p < .001


Hypothesis 4 tested the relative importance of religious characters in
prime time entertainment programs, as evidenced by their centrality to the
plot.
H4: Religious characters will be less central to programs' storylines than
are other characters.

Analysis of individual characters' centrality to program plots indicates
that, contrary to previous research, religious characters tended to have
major roles. As Table 7, below, demonstrates, nearly half (48.4%) of all
characters were main characters, but three-quarters (75.0%) of religious
characters were main characters (p < .01). Furthermore, only 6.3% of
religious characters were background characters, compared to 21.9% overall.
Therefore, Hypothesis 4 was not supported by the data.


Table 7: Type of Characters by Plot Centrality:
Character Type
% Religious Characters
N=32
% All Characters
N=548
Main
   75.0
    48.4
Secondary
   18.8
    29.7
Background
     6.3
    21.9
Total:
100.0
100.0
?2 = 10.24, df = 2, p < .01


Discussion
The general results of this study suggest that religious characters are not
very important considerations in the production of prime-time television
programming, replicating earlier findings .
Further, the data presented here support gatekeeping theory , inasmuch as
both the studio producing a program and the network that airs it were both
significantly correlated with the number of religious characters appearing
in that program. Data also indicate that PAX is meeting its goal of
increasing the amount of spirituality on television , both in the programs
it produces and in those the network airs.
Finally, some data do not square with earlier findings: most religious
characters were Protestant rather than Catholic, and were generally main
characters versus background players. Perhaps this is due to the profound
influence of PAX programming on the overall count of religious characters,
or perhaps it signals a trend toward featuring religious characters more
prominently .
A limitation of the study is the low number of religious characters
identified. Often, there were too few characters to generate reliable
statistical analyses, requiring that some data be combined (and thus,
obscured). A follow-up study – already in progress – should include all the
programs currently airing in prime-time, not just the top 3 comedies and
dramas by network. This would permit more rigorous statistical testing and
improved generalizability of results.
Future studies should also examine the role of individual producers and
writers of prime time fiction, to determine how their role both shapes and
is shaped by the structure and operation of the studio and network system
of television production . Further research into this under-studied area
should help clarify many of the issues raised by this project.
        TV and Religion 2
References