Media Consumption and Social Capital Patterns in Urban African Americans and
Professor, University of Missouri-Columbia
Postdoctoral Fellow, Stanford University
Doctoral Student, University of Missouri-Columbia
Correspondence concerning this article should be addressed to Esther Thorson,
Graduate Studies Center, 116 Walter Williams, School of Journalism, University
of Missouri, Columbia, MO 65211, 573-882-9590, e-mail:
[log in to unmask]
Paper submitted to the Mass Communication and Society Division of AEJMC
April 1, 1999
Media Consumption and Social Capital Patterns in Urban African Americans and
The survey research reported here was examined for links between exposure and
attention to newspapers, local television news, and entertainment television and
patterns of social capital exhibited by African Americans and Whites in a large
Midwest city. The news media of the city included a daily newspaper that has
been committed to public journalistic approaches for approximately three years.
Part of the public journalism effort has involved increased efforts to
communicate meaningfully with the large African American population in the city.
The three local television network news organizations had not been involved in
such efforts, and informal viewing of their nightly news programs showed typical
stereotyping of African Americans and no emphasis on involving the public in
their own communities. The main findings included high levels of community
participation in the form of voting and belonging to organizations by the
African American respondents; significant associations between newspaper use and
a variety of social capital measures for everyone, and significant interactions
between newspaper use and ethnicity on the social capital measures. Use of
local television did not show the same strength of relation with social capital
indices, and entertainment television was negatively associated with social
capital in both ethnicities. The findings are discussed in terms of both
theoretical considerations and implications for empowerment of citizens via news
Media Consumption and Social Capital Patterns in Urban African Americans and
Introduction and Overview
There has been a great deal interest in the concept of "social capital" in the
last ten years (e.g., Coleman, 1988; Putnam, 1995a; 1996). Only recently,
however, has research begun to explore the relationships between consumption of
news and social capital. Traditionally, American journalism has been
conceptualized as playing a significant role in democratic processes, providing
citizens with information, ideas, and opportunities for public discourse
(Siebert, Peterson, & Schramm, 1966). News that does this enables citizens to
participate intelligently in the functioning of their own communities. It makes
sense therefore to posit that news consumption will be associated with social
capital, particularly as the concept is operationalized in terms of community
participation. The work in this area, as is so often the case in communication
research, has focused largely on White, middle-class America (Portney & Berry,
1997). The present study embarks on new ground as it examines both the linkage
between use of the news and social capital patterns, and the possibility of
differential linkages for urban African Americans and Whites.
Underlying the research reported here is a macro-level hypothesis that when a
news medium makes an effort to inform citizens about their community in a way
aimed to encourage their input, understanding, and participation, greater
consumption of that news will enhance social capital. It is also hypothesized
that when a news medium pays special heed to a large minority population in its
city, and invites that population to become involved civically, the minority
group will show significant associated levels of social capital.
In the last few years, social capital has become a focus of research in many of
the social sciences, for example, sociology, communications, public health, and
economics. Coleman (1988) first defined social capital by contrasting it with
the concepts of financial and human capital. Financial or physical capital is
wealth in the form of money or property. Human capital (Schultz,1961) is
resources that result from individuals' education, skills, and capabilities.
Coleman (1988) defined social capital as of networks of social relationships and
norms directing behavior and attitudes within those networks. Networks are made
up of human relationships. Social capital within networks is a function of
norms for how these relationships are conducted. Norms associated with higher
social capital include (a) free and efficient transmission of information; (b)
mutual obligations and expectations for cooperation and compromise; (c) trust
among those in the network; and (d) sanctions for those who break the norms.
Networks with high social capital also have a norm defined by widespread
willingness to forgo self interest and act in the interest of others.
Social capital creates power and opportunities to get work done, just as do
physical and human capital. It should be noted that social capital, which was
originally defined at the level of individuals, has also been defined at an
organizational level (see, for example, Kreuter, 1998), where the relationships
and norms exist among organizations rather than individuals.
One of the reasons social capital is so important is because of its impact on
the successful functioning of "communities," that is, groups defined in as
diverse ways as schools, businesses, economic systems or public health in a
city. Coleman (1988) provided evidence that high school dropout rates were much
lower in Catholic than public schools, which he argued was due to the stronger
networking of Catholic parents themselves. Seipel, Hobbs, Beaulieu, and Elder
(1996) pointed out the importance of social capital in predicting successful
rural economic development. Kreuter (1998) showed that social capital in
relationships among community leaders was associated with more positive
indicators of public health on dimensions like the percent of live births,
percent of children being immunized, and percent of deaths due to traffic
accidents. Putnam (1993) argued that social capital is critical to a healthy
economy. In a recent case study, Gladwell (1999) examined the social capital of
an elderly woman living in Chicago. Lois Weisberg had developed a large and
highly diverse social network that provided the resources to accomplish such
civic ends as reviving Chicago parks, finding jobs for a variety of people, and
enhancing a variety of cultural activities in the city. The point of Gladwell's
research on Lois Weisberg is that people who create and then use social capital
get extensive amounts of civic work done.
The operating assumption in this work is that social capital, that is, these
kinds of networks and norms, is something to be fostered. Putnam's (1995b) work
has received a lot of attention because of his claim that social capital in the
U.S. is diminishing. The current health of social capital in America is
controversial, with some researchers arguing that Putnam is wrong (e.g., Ladd,
1996; Verba, Schlozman & Brady, 19xx), Nevertheless, one of the most
interesting aspects of the discussion is Putnam's (1995b) argument that the
culprit in the claimed loss of social capital is television. He argues that
television has produced a greater loss in social participation than have such
variables as entry of women into the workforce, pressures of time and money,
increase in divorce rate, values of baby boomers, and a variety of other
variables. It should be noted that Putnam (1995b) does not distinguish between
the effects of entertainment and news television For Putnam, what television
does is usurp time that previously went into community participation and other
related networking. Norris (1996) points out that it is important to consider
the diversity of television experience rather than assuming that all television
viewing is detrimental.
Measurement of social capital
Social capital is a complex concept, and has been measured in a variety of ways.
Sociologists have investigated people's networks of relationships (Granovetter,
1973; Burt, 1984 ; Marsden, 1990). Putnam (1995b) has indexed it with voting
turnout, membership in civic organizations, self-reported time spend in informal
socializing and visiting with others, and self-reported trust of others. Ladd
(1996) measured membership in political clubs, working with others on community
needs, and volunteering. Thus social capital is a concept involving both
attitudes and behaviors, and it is important to include both of these variables
when considering the full dimensionality of the concept.
News and Social Capital
In another approach where social capital plays an important role, a number of
communication researchers have become interested in whether news consumption,
attention and use is associated with or might even be said to influence social
capital. The intellectual foundations of this research can be traced back to
the Enlightenment philosophers, but a relatively recent articulation of the
concept came from Janowitz (1952), who defined the community press as an
instrument for integrating individuals into society. In social capital terms
this means linking individuals into networks that have strong civic norms.
Tichenor, Donohue & Olien (1980) based their work on this idea, examining how
different communities in Minnesota dealt with conflicts (for example, whether to
allow a nuclear power plant to be located in their town). Indeed, they found
that the effectiveness of these communities' behaviors was related to the news
media in their community. People whose community was served by a weekly rather
than a daily paper were not able to get the range of news they needed to deal
effectively with community issues. In our terms, without efficient transmission
of information via the media, social capital was diminished, and therefore
communities functioned less well.
In the same tradition, Stamm and Fortini-Campbell (1983) looked at correlations
between measures of how closely people were tied to their community and their
level of newspaper use. The correlations were high. For example, homeowners
were more likely to subscribe to the paper than non-homeowners. The longer
people had been subscribing to the paper, the more they perceived themselves as
involved in the community.
Most recently, Stamm and his colleagues (1997) looked at the linkage between
news media consumption, amount of interpersonal communication, and involvement
in community. They measured community involvement as paying attention to
community matters, having ideas for improving community, working for change and
getting people together to talk about issues. Both media use and interpersonal
communication were important for predicting the amount of social capital
Martinelli & Chaffee (1995) were interested in how media use would relate to the
way new immigrants became knowledgeable about and participant in politics.
Political knowledge of those who had just become citizens and were coming up on
their first election was well predicted by newspaper and television news
exposure and attention. In our terms, then, media use was associated with
political participation, a variable that is an index of social capital.
Civic Journalism News and Social Capital
Friedland, Sotirovic, & Daily (1998) suggested the kind of reporting referred to
as "civic journalism" encourages public participation. Chaffee and his
colleagues (1997) tested this hypothesis by comparing the level of social
capital among people in a city who were knowledgeable about civic journalism
projects to those who were not. He found that a vareity of participation
measures were associated with news exposure. Thorson and Leshner (1998) looked
at social participation and awareness of special ongoing news series (like civic
journalism projects on politics, urban transportation, or anti-smoking efforts)
in the news media in five cities. They found that greater news consumption,
both habitual and of the special civic coverage, was associated with belonging
to more civic organizations, being more likely to vote, and being more
knowledgeable about the news content itself. Here again was linkage between
news use and of social capital.
Of course, to date, all of the studies of news and civic participating have been
correlational, and therefore it cannot be claimed that news consumption causes
social capital to increase. But clearly the strong correlation between the two,
across variations in how the concepts are measured, suggests that some
interesting processes link the impact of news to norms and networks of
civically-based social relationships.
African Americans and News Consumption
There is little research on how minorities, including African Americans, use
mainstream news (Gandy, 1998; Mares, 1996; Messaris,1994). But from the
research does exist, it is clear that news does not carry the same information
for African Americans as it does for Whites. There is mounting evidence, for
example, that African Americans are often invisible in local and national news
(Entman, 1994a, 1994b; Rodgers, Thorson & Antecol, 1998), or if they are
present, it is in negative stereotypes such as being the victims or perpetrators
of crime, having lower education or income levels, or holding low-status jobs
(e.g. Gandy, 1996). Goshorn and Gandy (1995) showed, for example, many
newspapers that covered a wire service story about discrimination in mortgage
markets used headlines that failed to note the fact that Whites were more likely
to be granted mortgages than African Americans. Thus even when the story was
about a minority group, the headline "downgraded the central issue to greater
invisibility." The reporting of the Rodney King riots in Los Angeles tended to
be presented as African American race riots, rather than events involving
Whites, Asians, and Latinos (Abelmann & Lie, 1995). Likewise, Smith (1994)
showed that television news during the riots focused on African American-White
relations, rather than African American-Korean relations or involvement of
Latinos. Street violence was most likely to feature African Americans
prominently. This was in contrast to arrest records which indicated most of the
people arrested for rioting were Latinos, (Miles, 1992). Gilens (1996) recently
showed that in the three leading news magazines, there was overrepresentation of
African Americans in stories about poverty, about twice as many as would have
been predicted from objective statistics about the racial character of poverty.
Indeed, Gilens (1996) noted that public opinion surveys have revealed that
Americans tend to substantially overestimate the proportion of African Americans
among the nation's poor. The representation of African Americans as unempowered
in these ways may then lead those who consume more of this kind of news to feel
themselves less capable or motivated to participate in their community.
Indeed, Gandy and Matabane (1989) showed that African Americans were less
trusting, and more dissatisfied or critical of mass media, even though they were
more active consumers of them. Tan and Tan (1979) argued that programs that
depict African Americans in low status social roles causes low self-esteem in
black audiences. A study at about the same time (Research Analysts, 1981)
showed that African Americans and Hispanics watched more television than Whites.
Allen and Hatchett (1986) concluded that exposure to negative portrayals of
African Americans common in 1970's situation comedies was associated with more
agreement with negative stereotypes about African Americans.
African Americans and Social Capital
Just as there is little research about news use by African Americans, there is
little study of social capital patterns in African Americans. Verba,
Schlossman, and Brady (1995) indicated that African Americans belonged to more
associations on average than Whites, mostly because they were more likely than
Whites at the same SES level to belong to religious and ethnic organizations.
Putnam (1996) pointed out that from 1972 to 1994, after education was controlled
for, 17% of African Americans agreed that "most people can be trusted," as
compared to 45% of Whites and about 27% of other ethnicities. Putnam (1996)
also argued, based on the General Social Survey, that the decrease in trust and
number of organizations belonged to dropped more for African Americans than for
The generally higher levels of social capital exhibited by African Americans
were further substantiated by Portney and Berry (1997). They showed in cities
that had either large (56% African American) populations or small ones (5%
African Americans), participation in face to face political activities was
higher for African Americans than for Whites. Participation of low
socioeconomic status residents of predominantly African American neighborhoods
was nearly twice that of low SES residents of low minority population
neighborhoods. Neighborhood organizations turned out to be a big factor in
African American participation levels.
Thus, although there are not many studies of African Americans and social
capital, they clearly show higher participation levels, and they show
significantly lower generalized trust in others as another index of social
Ethnicity, news, and social capital
This take on the literature about ethnicity, news, and social capital leads
naturally to the question of whether news plays a role in social capital levels
of African Americans and Whites. The focus of the present research, therefore,
was to first look at whether news consumption in a Midwestern city was
associated in a patterned way with not one but several important measures of
social capital. Some of the indices of social capital employed here are the
typical political or civic behaviors like voting, doing volunteer work, and
belonging to community organizations. Other measures were derived from the
health communication literature. This literature suggests that the health of
citizens depends on some specific behaviors that could be said to represent
social capital (Kreuter, 1998). One such measure is whether people perceive
their neighbors as willing to intervene when children and youth engage in
antisocial behavior like creating graffiti or fighting. Another is how
neighborly people feel toward those who live near them. Thus, we look at social
capital both in terms of attitudes, which likely reflect norms operating, and in
terms of behaviors that involve networking with others for the betterment of
We examined news and social capital in a Midwest city that had three important
features. First, its single mainstream newspaper had a significant history of
civic journalism. Second, although we could locate no formal studies of local
television content in the city, daily watching of the three networks indicated
no efforts at civic journalism, and a consistent stereotyping of African
Americans into roles of crime perpetrators and victims, or as sports stars.
Third, the city had a significant population of African Americans, 50% in the
city itself, although less in the donut of suburbs surrounding the city.
Given the overt and long-term efforts of the city newspaper to tell people
about their community and invite them to participate in that community, we
Hypothesis 1: Greater use of the local newspaper would be associated with
higher measures of civic social capital attitudes including trust in government,
satisfaction with living in the city, feelings of neighborliness, perception
that others will intervene when things go wrong, and behaviors such as voting,
belonging to civic organizations, and doing volunteer work.
Given the lack of civic efforts in the city's local television news, we expected
Hypothesis 2: Greater use of local television news would be associated with
fewer positive measures of civic social capital attitudes and behaviors than
seen with newspaper use.
Based on the findings of Putnam (1995b, 1996), but distinguishing entertainment
from news television, we hypothesized that watching entertainment television
would be associated with less social capital, particularly for those measures
like belonging to organizations and volunteering where a significant amount of
time was involved.
H3: Those with higher levels of entertainment television viewing will show
lower levels of voting, number of organizations belonged to, and number of hours
The literature on social capital levels of African Americans led us to expect
that they would show higher social capital levels on all measures except trust
H4: Higher social capital levels will be exhibited by African Americans than
Whites except for trust in government.
Given the self-reported efforts of the city's newspaper to increase
representation of African Americans, and to reduce the stereotyping of those
representations, along with the major thrust of civic journalism efforts that
consistently invite community participation, we expected that African Americans
would show positive changes in social capital with higher levels of newspaper
H5: African Americans will show greater improvement in social capital with
increased newspaper consumption as compared to White Americans.
Based on the research of Entman (1994a, 1994b) we expected that greater
consumption of local television news would not have salutary effects on social
capital levels of African Americans, who would be likely to experience reduced
trust, reduced self-esteem, and reduced emotion as a result of the
representations they saw in that news (Gandy, 1998).
H6: African Americans will show less improvement in social capital with
increased television news consumption as compared to White Americans.
H7: Both African Americans and Whites should show damage to social capital
under higher levels oaf viewing entertainment television.
Sample. The hypotheses were tested with a random-dialed phone survey in a
large Midwest city administered to 733 adults in October, 1998. The survey was
conducted by the staff of a professional survey center at a Midwestern
university. Stratified random digit dialing and the Troldahl-Carter-Bryant
method for the selection of respondents (Lavrakas, 1993) were employed to ensure
a sample representative of the metropolitan area. Computer Assisted Telephone
Interviewing (CATI) software was used in the selection of respondents from a
given household and for entering their responses. The response rate was 62%
after business and disconnected phones were removed.
Survey research has been an externally valid method for studying media effects.
Nevertheless, it is clear that cross-sectional surveys, such as the one used
here, cannot completely address the question of causal ordering since the main
data analysis techniques are correlational and correlations are only one element
in a cause-and-effect relationship. This fact has been the cause of criticism
(e.g., Lastovicka, 1995, pp. 70-71). Still, if a null or contrary correlation
is obtained, it is a powerful indication that a particular media effect
inference should be abandoned or at least reconsidered in terms of other
variables that might be affecting it. In the absence of correlation, issues of
time ordering and causality are moot (Chaffee, McDevitt & Thorson, 1997). For
this reason, a survey approach at the initial stages of research, is an
efficient manner to test a variety of possible effects at the same time
(Chaffee, McDevitt & Thorson, 1997). Further, this study also goes beyond
simple correlations by examining the impact of demographic variables on the
various dependent variables, by removing that impact via hierarchical
regressions, and then by examining the impact of the variables of interest,
namely, ethnicity and media-use.
Dependent variables. Four attitudinal items related to social capital were
measured. Government Trust was made up of the average score of four 4-point
items. Respondents were asked to indicate their agreement with: "I trust
government to respond to community needs," "I trust government to be free from
corruption," "I trust government to try to be fair," and "I trust government to
keep citizens informed" (M = 2.47, sd. = 0.53, alpha = 0.80). City satisfaction
was made up of the average score of five 4-point items.
Respondents were asked to indicate their agreement with: "Local businesses give
back to community," "People in this community are free to express a dissenting
opinion without reprisal," "Business leaders are here to make money and that's
it," "There are times I feel unwelcome in my community," and "Religious
organizations contribute substantially to the community" (M = 2.78, sd. = 0.39,
alpha = 0.63). Neighbors was made up of the average score of six 4-point items.
Respondents were asked "How often do you borrow or exchange things with your
neighbors," "How often do you visit your neighbors," "How often have you helped
your neighbors," "How likely are you to call your neighbors," "How likely are
you to ask for your neighbors help," and "How much do you have in common with
your neighbors" (M = 2.12, sd. = 0.67, alpha = 0.79). Neighbor Intervention was
made up of the average score of three 4-point items. Respondents were asked to
indicate whether they thought their neighbors would intervene if they saw kids
"skipping school," "spray painting," and "fighting" (M = 1.85, sd. = 0.83, alpha
One interpersonal behavior related to social capital was measured. Talk to
Others was made up of the average score of two 4-point items. Respondents were
asked to indicate their agreement with: "It's important to talk to others," and
"It's important to be open to what others have to say" (M = 1.88, sd. = 0.42.
corr = 0.20, p<.0001).
Finally, three civic behaviors related to social capital were measured. Voting
was made up of the sum of four yes/no items: "Are you registered to vote," "Did
you vote in the last local government election," "Did you vote in the last state
election," and "Did you vote in the last national election" (M = 3.06, sd. =
1.34, alpha = 0.82). Number of Organizations was created by asking respondents
whether they belonged to a variety of civic organizations (yes/no) such as
business civic groups, religious organizations, charity groups, and social
clubs. The total yes answers represented the respondent's score (M = 2.12, sd.
= 0.67). Number of Volunteering Hours was made up of one item: "How many hours
a week to you do volunteer service" (M = 6.98, sd. = 18.76).
It should be noted that where necessary, items were reverse coded such that the
higher scores always represented more positive results from a social capital
point of view.
Independent variables. Ethnicity. All respondents were asked to indicate
their ethnic makeup. Because the interest here was with respect to
White/African American similarities and/or differences, only those two ethnic
groups were used in the analyses. Whites made up 52.5% of the sample and
African Americans made up 47.5% of the sample.
The media-use variables were operationalized as follows. Respondents were
asked how days a week they read a newspaper and watched local television news.
They were also asked, for each day of the week, how many hours of entertainment
television they watched. Although Price and Zaller (1993, p.159) suggested that
such simple exposure measures were "sufficient to produce various attitudinal
effects that interest communication researchers," it is clear that attention
also plays an important role (e.g., Chaffee & Schleuder, 1986). However, from
our perspective, attention should not be considered a unique variable since it
is dependent upon exposure. To include attention in our analyses, we weighted
exposure (i.e., multiplying) by the amount of attention paid during those
exposures. Such a procedure can realistically only be done with attention to
domain-specific vehicles such as news broadcasts and newspapers rather than to
domain-general exposures such as exposure to entertainment television.
Accordingly, respondents were also asked to indicate their level of attention
during their exposures to newspapers and local television news which were then
used to weight the newspaper and local television news exposure variables.
Specifically, then, Newspapers exposure weighted by attention was made up
newspaper exposure weighted by newspaper attention (M = 18.63, sd. = 15.80).
Local television news weighted by attention was made up of local television news
exposure weighted by attention (M = 27.22, sd. = 16.16). Entertainment
television hours/day was created by averaging the respondents daily
entertainment television exposure responses (M = 2.38, sd. = 1.95).
Controls. There are several factors that were thought important to control in
our analyses. These included gender (1 = Male), age, income (1 = Less than
$5,000, 2 = More than $5,000 but less than $10,000, 3 = More than $10,000 but
less than $15,000, 4 = More than $15,000 but less than $20,000, 5 = More than
$20,000 but less than $25,000, 6 = More than $25,000 but less than $30,000, 7 =
More than $30,000 but less than $35,000, 8 = More than $35,000 but less than
$40,000, 9 = More than $40,000 but less than $45,000, 10 = More than $45,000 but
less than $50,000, 11 = More than $50,000), education level (1 = Less than High
School, 2 = Some High School, 3 = High School Degree, 4 = Some College, 5 =
College Graduate, 6 = Some Graduate School, and 7 = Graduate Degree), and years
of residence in the community.
Statistical Analysis. For each dependent variable, the control variables were
entered as Step 1 and the two main effects as well as their interaction term
were entered in Step 2. The significance of each step was assessed through
incremental F-tests (Gujarati, 1995). Specifically, for each dependent
variable, after the variance accounted for by the control variables was removed,
various blocks were entered and removed. First, Ethnicity, Newspaper exposure
weighted by attention, and their interaction were entered to ascertain their
contribution to explained variance and to determine which of them, if any, were
predictors. These variables were then removed and the process was repeated for
Ethnicity, Local television news exposure weighted by attention and their
interaction, and, finally, for Ethnicity, Entertainment television hours/day and
The significance of the variables in the control block, as well as Ethnicity,
was assessed by two-tailed t-tests. Because of the directional nature of the
hypotheses (i.e., we were only interested in increased media exposure weighted
by attention, or increased exposure in the case of entertainment television),
the significance of the media-use variables and the interaction term were
assessed by one-tailed t-tests. It should be noted that this provides an
extremely stringent test. Undertaking regression analysis in this fashion means
that if one block of variables is significant beyond the controlled effects of
the previous blocks, even if the increments to R2 are not large, the effect is
Of the analyzed respondents, 44.9% were male and 55.1% were female (1 = Male);
and as mentioned, 385 were Whites and 348 were African-Americans (White = 1).
The mean age of respondents was 46.34 (sd. 18.23) and they had lived in the
community, on average, about 32.13 years (sd. 21.87). They had a mean income of
6.27 (sd. 2.98) which was between $25,000 and $30,000. Most respondents had
completed at least some college (M = 4.06, sd. 1.53).
Pattern of impact of the demographic control variables
The importance of considering and controlling for demographic variables was
evident in the pattern of results on the eight dependent variables (see Tables 1
to 3 for F-tests; betas not shown in the tables). On the four attitudinal items
related to social capital, age was a significant predictor on Government Trust
and City Satisfaction while income and education were significant predictors on
City Satisfaction and Neighbors. There were no significant predictors on
Neighbor Intervention. On Talk to Others, our interpersonal behavior measure,
only education was a significant predictor although gender and years of
residence approached significance. Finally, on the three civic behaviors
related to social capital, namely voting, number of organizations and
volunteering hours, gender and years of residence were predictive on the first
two, income was predictive of the last two, and education was predictive on all
three. Age was predictive only of voting.
Tests of the hypotheses
The hypotheses were tested with a series of regressions, the results of which
are shown in Tables 1-3. In each analysis, the demographic variables were
entered first. These variables included: gender, income, education, time lived
in the city, and age. For each subsequent block, race was entered, then the
media consumption variable of interest, and then the interaction term of
ethnicity and media consumption. Whenever an interaction term was statistically
significant, we looked at the results graphically to determine whether the
impact of more media use increased or decreased the difference between African
Americans and Whites.
We evaluate first Hypothesis 1, which suggested that greater exposure and
attention to newspapers would be positively associated with the measures of
social capital. Indeed, newspaper exposure and attention was positively related
to thinking your city was safe, voting, the number of organizations people
belonged to and number of hours of volunteering. Inconsistent with idea that
newspaper use is a consistent boon to social capital, more newspaper reading was
negatively associated with trusting government, and was not related to three
measures of interpersonal interaction, talking to others, being neighborly, or
perceiving others to be willing to intervene when youth are acting out. Thus it
would seem that newspaper use is highly associated with the standard civic
behaviors, but not related to interpersonal interactions.
Hypothesis 2 suggested that more exposure and attention to local television news
would be somewhat negatively associated with the measures of social capital.
Local TV news had a negative impact on government trust, but a positive impact
on belonging to organizations, and city satisfaction. There was no significant
main effects of local television news on the other five indices of social
Hypothesis 3 suggested that entertainment TV use would have a negative impact on
all social capital measures. More use of entertainment TV had significant
negative impact on the number of organizations people belonged to, and the
number of hours they spent volunteering per week. In one case, more
entertainment TV meant a higher social capital score, that is for intervening.
Entertainment television had no impact on any of the other five dependent
variables. Thus the negative impact of entertainment TV was not homogenous
across all measures of social capital. Instead, it worked most of its influence
on the two major measures of getting up and doing things, i.e., belonging to
organizations, and hours of volunteering.
In terms of main effects of the media, then, entertainment TV was the most
consistently negative influence, although it did have a positive impact on the
intervening variable. Newspaper use had the most consistently positive effect,
with local TV news having some important positive effects, but not as
consistently across the different measures of social capital.
Hypothesis 4 articulated the expectation that African Americans would show
higher levels than Whites on most of the social capital measures. The picture,
however, was different and more complicated than predicted. The impact of
ethnicity was examined only after removing effects due to typical class
indicators like age, income, and education, Whites were more likely to think
their city was safe, more likely to be neighborly, more likely to see others as
intervening when youth acted out, and to spend more hours volunteering. African
Americans, however, showed higher government trust, and belonged to more civic
organizations than Whites.
In general we expected that newspaper exposure and attention would decrease
ethnicity gaps wherever they occurred (Hypothesis 5). We expected television
news to accomplish less gap-closing than newspapers (Hypothesis 6), and we
expected no gap-closing with the use of entertainment television (Hypothesis 7).
The results, however, were considerably more complex.
We look first at newspaper interactions with ethnicity. Although neither
ethnicity nor newspaper exposure had main effects on perceived intervention of
others, there was a significant interaction, which is depicted in Figure 1. At
higher levels of newspaper use, whites actually showed lower levels of perceived
likelihood of intervention. African Americans showed the same level of the
dependent variable regardless of newspaper use. More importantly, however,
there were significant interactions of ethnicity with newspaper use on the
number of organizations belonged to and number of volunteer hours. For both
measures, newspaper use increased social capital, being associated with more
organizations and more volunteer hours. For both organizational membership and
volunteer hours, there was an opening gap with higher newspaper use. African
Americans showed higher values on both dependent variables with higher newspaper
use, as can be seen in Figures 2 and 3. There were no other significant
interactions between ethnicity and newspaper use. The hypothesis of greater
impact of newspaper on African Americans and Whites was, however, supported for
two very important measures of social capital.
We look next at local television use. As we saw earlier, African Americans
showed higher government trust than Whites. Figure 4 shows that that effect
disappeared for high exposure to local television news. Trust for African
Americans diminished with the higher levels of television news.
Also as we saw earlier, Whites showed significantly higher city satisfaction
than African Americans. TV news use had little effect on African Americans
(Figure 5), but there was less city satisfaction by Whites with higher levels of
local TV news use. This effect will be discussed in terms of the amount of
negative city news that has been demonstrated to make up local television
broadcasts (ref). Overall, however, there was not a strong pattern of
interactivity of local news use and ethnicity, and therefore Hypothesis 6 of
lesser impact of TV news than newspapers was supported.
As we saw, entertainment television showed significant negative impact on
social capital. It was not clear whether we should expect this variable to
interact with ethnicity. Figures 6, 7 and 8 show, however, that interactions
did occur. In Figure 6 we see greater damage to perceptions of neighbors being
willing to intervene for African Americans than for Whites. In contrast, Whites
suffered more damage from higher levels of entertainment television on voting
and belonging to organizations than did African Americans, although there was a
consistent pattern of lowering social capital for both groups. This then
provided support for Hypothesis 7.
These results lend support to the hypothesis that if news is reported in a way
that encourages civic participation, knowledge and attitudes, it will be
associated with higher levels of social capital, even as measured in different
ways. The strong pattern of impact of newspaper use on the social capital
measures shows does not demonstrate causality, but it sets the stage for testing
In addition, newspaper use showed two very strong interactive effects with
ethnicity, one on belonging to organizations, and the other on volunteering.
Again, because the study is correlational and because we only looked at the
impact of a single city newspaper, we certainly cannot conclude that the civic
journalist approach used at that newspaper played a role in this positive
outcome or not. It may be that simply reading more newspaper news of any times
has a particularly salutary impact on minority social capital. But the result
is intriguing, and demonstrates the importance of looking further into this
question. Because so many social manipulations show increasing gaps (ref), it
is critically important to understand ones that actually put minorities out
further ahead of the majority group.
Use of local television news did not show a particularly positive pattern of
relationship to social capital. Higher local TV news use was associated with
trusting government less. Overall, it was associated with being more satisfied
with ones own city, but interestingly, the enhancement was significantly less at
higher levels of use. Neither of these outcomes presents a very encouraging
view of local television news for the fostering of social capital. This is
particularly dismaying given the preponderance of people who report that
television is their main source of news.
The relationship of entertainment television to social capital was supportive
of Putnam's contention that television is subtracting from people's
participation in their own community. It should be noted, however, that it is
now television news, but entertainment television that is the culprit. This
makes sense in terms of the large number of hours that people report spending
with entertainment television. It would be difficult to watch television four
or five hours each night and still coach little league baseball, lead a Brownie
troop, and sing in the church choir. Indeed, the damage to organizational
membership and volunteering is the strongest effect observed here. There was
little impact of entertainment television on attitudes. It seems to be the
behaviors that are lessened.
African Americans in the city tested here showed higher levels of social
capital, particularly in the behavioral area: voting and belonging to civic
organizations. It is important to observe the differential media effects on
this impressive participation. Newspaper use enhanced it. Local television did
not seem to have much impact, but entertainment television use was a large
negative for both voting and organizational membership. For those interested in
further harnessing the power of social capital in minorities, newspapers are
clearly an important tool-and entertainment television is a factor to be
It should also be noted that attitudinal measures related to social capital
were more of a mixed bag for African Americans. They showed higher trust in
government, but lower satisfaction with their city, and less connectedness with
their neighbors. They were also less likely to perceive others as willing to
intervene when things went wrong. Interestingly, none of the media variables
measured here, newspaper and local television news use and entertainment
television had much impact on these attitudinal variables.
Whites in the city tested here showed comparatively positive levels of social
capital in terms of being satisfied with how their city worked, connectedness
with neighbors, and high levels of volunteering. Newspaper use enhanced their
belief that their neighbors would intervene when problem behaviors were seen,
and the number of organizations they belonged to. Local TV news use showed
little effect on them, and for voting and belonging to organizations,
entertainment television was a negative factor. Thus the overall picture of the
impact of these media experiences is not all that different from African
Americans and Whites. Both groups benefit clearly from newspaper use and both
groups are harmed by entertainment television.
These are clearly exploratory findings and face the many critiques of a single
first study. We don't know enough about specific content of the newspaper and
television news and the entertainment television to be able to ascertain what
was content effects and what were time use effects. We don't know whether the
fact that the newspaper in the city uses public journalistic approaches is
relevant. We need to know more about the variations in the the main local
television news programs in town. But the results are intriguing. They do
suggest a perhaps greater impact of newspaper on at least this minority group
than on Whites, although the generally salutary impact of newspaper was
impressive. It will be important to look more closely at these issues and more
broadly. What are news and entertainment television impacts in more diverse
cities like Los Angeles or Miami? Are African American responses to these media
different from other minority groups like Asians and Latinos? These and many
other such questions will help us to eventually understand better how media
relate to creation and fostering of social capital, and perhaps to increase it
levels to better aid the social problems of our cities and towns.
References Abelmann, N., & Lie, J. (1995). Blue dreams: Korean Americans
and the Los Angeles riots of 1992. Cambridge, MA: Harvard University Press.
Allen, R. L., & Hatchett, S., (1986). The media and social reality effects.
Self and system orientations of blacks. Communication Research, 13(1),
97-123. Burt, Ronald S. (1984). Network items and the general social survey.
Social Networks, 6, 293-339. Chafe, S. H., McDevitt, M., & Thorson, E. (1997,
Aug.). Citizen response to civic journalism: Four case studies. Paper presented
to the Mass Communication & Society Division at the 80th Annual AEJMC
Conference, Chicago, IL. Chaffee, S. H., & Schleuder, J. (1986). Measurement
and effects of attention to media news. Human Communication Research, 13,
76-107. Coleman, James S. (1988). Social capital in the creation of human
capital. American Journal of Sociology 94 Supplement, S95-S120. Entman, Robert
(1994a). African Americans according to TV news. Media Studies Journal,
8(3-4), 28-38. Entman, Robert M (1994b). Representation and reality in the
portrayal of Blacks on network television news. Journalism Quarterly, 71(3),
35-62. Friedland, Lewis Al, Sotirovic, Mira, & Daily, Katie (1998). Public
journalism and social capital: The case of Madison, Wisconsin. In (Eds.), E.
Lambeth, P. Meyer, & E. Thorson, Assessing Public Journalism, Columbia, MO:
University of Missouri Press. Gandy, Oscar H., Jr. (1996). If it werenUt for
bad luck. Framing stories of racially comparative risk. In V. Berry and C.
Manning-Miller (Eds.), Mediated messages and African-American culture:
Contemporary issues. Thousand Oaks, CA: Sage. Gandy, Oscar H., Jr (1998)
Communication and Race: A Structural Perspective. New York: Oxford University
Press. Gandy, Oscar H., & Matabane, P. (1989). Television and social
perceptions among African Americans and Hispanics. In M. Asante & W. Gudykunst
(Eds.), Handbook of international and intercultural communication. Newbury
Park, CA: Sage, 318-348. Gilens, Martin (1996). Race and poverty in America:
Public misperceptions and the American news media. Public Opinion Quarterly,
60, 515-541. Gladwell, Malcolm (1999). Six degrees of Lois Weisberg. The New
Yorker, January 11, 52-63. Goshorn, K. & Gandy, Oscar H (1995) Race, risk
and responsibility: Editorial constraint in the framing of inequality. Journal
of Communication, 45(2), 133-151. Granovetter, Mark S. (1973). The strength of
weak ties. American Journal of Sociology, 78(6), 1360-1380. Gujarati, D.
(1995). Basic Econometrics (3rd ed.). New York: McGraw-Hill. Janowitz. Morris
(1952). The community press in an urban setting. Glencoe, IL: Free Press.
Kreuter, Marshall W., Young, Laura, & Lezin, Nicole A. (1998). Measuring
social capital in small communities. Report from Health 2000, 2900
Chamblee-Tucker Rd., Atlanta, GA. Ladd, Everett C. (1996). The data just
donUt show erosion of AmericaUs Rsocial capitalS. The Public Perspective,
June/July, 1-22. Lavrakas, P. J. (1993). Telephone survey methods: Sampling,
selection, and supervision (2nd ed.). Newbury Park, CA: Sage. Mares, 1996. The
role of source confusions in televisionUs cultivation of social reality
judgments. Human Communication Research, 23, 2. Marsden, Peter V (1990).
Network diversity, substructures, and opportunities for contact. In (Eds.), C.
Calhoun, M. W. Meyer, & W. R. Scott, Structures of power and constraint.
Cambridge: Cambridge University Press. Martinelli, Kathleen A., & Chaffee,
Steven H. (1995). Measuring new-voter learning via three channels of political
information. Journalism & Mass Communication Quarterly, 72(1),
18-32. Matabane, 1988. Television and the black audience: Cultivating
moderate perspectives on racial integration. Journal of Communication , 38,(4),
21-31. Messaris, 1994. Television and social perceptions among African
Americans and Hispanics. In (Eds.), M. Asante, & W. Gudykunst, Handbook of
international and intercultural communication. Newbury Park, CA: Sage,
318-348. Miles, J. (1992). Blacks vs browns. The struggle for the bottom of
the rung. The Atlantic Monthly, October, 41-68. Norris, Pippa (1966). Does
television erode social capital? A reply to Putnam. PS: Political Science &
Politics, September, 474-480. Portney, Kent E., & Berry, Jeffrey M. (1997).
Mobilizing minority communities. American Behavioral Scientist, 40(5),
March/April, 632-644. Price, V., & Zaller, J. (1993). Who gets the news?
Alternative measures of news reception and their implications for research.
Public Opinion Quarterly, 57, 133-164. Putnam, Robert (1993). The prosperous
community: Social capital and economic growth. The American Prospect, Spring,
35-42. Putnam, Robert (1995). Bowling alone: AmericaUs declining social
capital. Journal of Democracy, 6(1), 65-78. Putnam, Robert (1995). Tuning in,
tuning out: The strange disappearance of social capital in America. PS:
Political Science and Politics, December, 664-683. Putnam, Robert )(1996). The
strange disappearance of civic America. The American Prospect, 24,
1-18). Rodgers, Shelly, Thorson, Esther, & Antecol, Michael (1998). Reality as
it appears in the St. Louis Post ?Dispatch: A content analysis. Paper
presented at the Annual Meeting of the Association for Education in Journalism
and Mass Communication, Baltimore, August. Schultz, Theodore (1961).
Investment in human capital. American Economic Review, 51(March),
1-17. Siebert, Fred S., Peterson, Theodore, & Schramm, Wilbur (1963). Four
Theories of the Press. Urbana: University of Illinois Press. Seipel, Michael,
Hobbs, Daryl, Beaulieu, Lionel, & Elder, Bill (1996). Social capital in the
production of human capital: Implications for rural economic development.
Paper presented at the Annual Meeting of the American Agricultural Economics
Association, San Antonio, TX, July 28-31. Smith, Erna (1994). Transmitting
race. The Los Angeles riot in television news. Cambridge, MA: Harvard
University, John F. Kennedy School of Government. Stamm, Keith R., Emig, Arthur
G., & Hesse, Michael B (1997). The contribution of local media to community
involvement. Journalism and Mass Communication Quarterly, 74(1),
97-107. Stamm, Keith R., & Fortini-Campbell, Lisa (1983). The relationship of
community ties to newspaper subscribing and use. Journalism Monographs, no. 84
(Columbia, SC: AEJMC). Tan, Alex, & Tan, G. (1979). Television use and
self-esteem of blacks. Journal of Communication, Winter, 129-135. Thorson,
Esther, & Leshner, Glenn (1998). Participation in community organizations and
consumption of TV and newspaper news. Paper presented at the Annual Meeting of
the Association for Education in Journalism and Mass Communication, Baltimore,
August. Tichenor, Phillip J., Donohue, George A., & Olien, Clarice N. (1980).
Community conflict and the press. Beverly Hills, Sage. Verba, S., Schlozman,
K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism in
American politics. Cambridge, MA: Harvard University Press.