Content-Type: text/html
Source Diversity and Newspaper Size_
Source Diversity and Newspaper Size:
The Use of Sources in Local News
A Paper
Presented to
The Association for Education in Journalism and Mass Communication (AEJMC)
and for consideration of the
MacDougall Student Paper Award
By
Kristy H. Nichols
The University of Louisiana at Lafayette
103 Ridgewood
Lafayette, LA 70506
337-233-8171 (h)
337-989-0001 (o)
[log in to unmask]
AV Needs: Overhead Projector
INTRODUCTION
The formal study of media effects has been ongoing for more than 60 years. Initially, mass media were thought to have immediate and significant effects on their audience (Lowery & DeFleur, 1995). As mass-media studies progressed, researchers were unsuccessful in isolating these effects, and it became increasingly clear that a new approach to studying the long-term influences of mass media was necessary. In the early 1970s, one of these new approaches, the theory of agenda-setting, was developed.
The majority of agenda-setting research has been concerned with verifying the ability of the news media to influence the public's perceptual agenda; however, a smaller percentage of studies has examined how the media agenda develops (Berkowitz, 1988).
By the 1980s a new phase of agenda-setting research transformed the news agenda from independent variable, or a cause, to dependent variable, an effect. Weaver and Elliott (1985) described the process of examining who sets the media's agenda as agenda building. One way to study agenda building is to examine the routines from which published news stories seem to evolve (Berkowitz, 1988; Shoemaker & Reese, 1996).
Routines, those patterned, repeated practices that media workers use to do their jobs, are an intricate part of the systematic information-gathering of media systems. These routine practices often result in standardized, recurring patterns of news and entertainment content. Many routines are designed to help the organization cope with physical constraints. As complex organizations with constant deadlines, the news media cannot cope with an unpredictable and infinite number of occurrences in the everyday world without a system (Shoemaker & Reese, 1996). Shoemaker & Reese attempted to examine this system by establishing a theoretical foundation for the development of mass media content.
Shoemaker and Reese believed that it was important to understand the influences on mass media content. In order to understand these influences, Shoemaker and Reese developed a hierarchy of influences model. This model approaches media studies from five levels: the individual level, media routines, organizational influences, extramedia influences, and ideology. Routines are classified as a set of constraints on the individual media worker that "form an immediate context, both within and through which these individuals do their jobs" (p.105). Therefore, these routines have a tremendous impact on how the media worker collects news and ultimately on how the news is reported. How or why do these routines develop?
Many routines exist to simply improve the efficiency of the news organization. An unlimited supply of potential news information, combined with a limited amount of resources to collect this information, forces journalists to follow routines that better enable them to produce the news. For example, in order to 'manufacture' the news the media rely on external suppliers or sources for material. This reliance enables sources to become a part of media organizations' routines, ultimately giving sources the power to affect news content (Gans, 1979).
Evidence generated over the last 20 years indicates that Gans was correct in his assertion that powerful sources have an advantage in affecting news content. The literature clearly suggests that government and official sources tend to dominate the news (Sigal, 1973; Brown et al., 1987), but there may be specific characteristics of media systems that make media organizations more or less likely to use certain types of sources. Shoemaker and Reese (1996) said that media system characteristics have a definite impact on source use and the way stories are covered. The extramedia level of Shoemaker and Reese's hierarchical model recognized that characteristics within the mass media marketplace, including market characteristics that result from a community's size, can influence media content. To further investigate the impact that community size has on media content, specifically as it relates to what sources newspapers frequently utilize, this study will examine source use in newspapers. It will compare source usage in local news stories in daily newspapers of various circulation sizes.
THEORETICAL FRAMEWORK
What factors inside and outside media organizations affect media content (Shoemaker & Reese, 1996). Why is it important to study content? As noted, until recently, most mass media research presented studies dealing primarily with the process through which the audience receives mass media content or with the effects of content on people and society. These types of studies have typically used mass media content as a starting point for mass communication research instead as a focal point. Shoemaker and Reese believed that understanding the influences that shape content is equally as important as the typical process-and-effects theme that examines the effects that mass media have on their viewers or readers.
McManus (1995) and Shoemaker and Reese (1996) offer two major theoretical perspectives on the influences that shape media content. Both recognized that the factors that shape news content are both internal and external to the media firm. As noted in the introduction, Shoemaker and Reese have attempted to understand these influences by establishing an organized theoretical framework for studying the influences of mass media content. Within this framework, Shoemaker and Reese produced a hierarchy-of-influences model that identified five levels of influences on mass media content while drawing attention to the connections between them. The hierarchy-of-influences model categorized the forces working to influence media content into the five levels: the individual level, media routines, organizational influences, extramedia influences, and ideology. Studies have shown that systematic content patterns can be explained by each of these levels.
McManus (1995) took an economic approach in the analysis of factors that influence news production. According to McManus, market forces, which are contained within Shoemaker and Reese's (1996) extramedia level of the hierarchical model, have a strong impact on the production of news (McManus, 1995). McManus' model of commercial news production explained the production of news in terms of interactions that occur between nine parties: the general public, consumers, news decision-makers, newsworkers, the organizational culture, the news department, the media firm itself, the parent corporation, and investors/owners.
McManus (1995) and Shoemaker and Reese (1996) both identified constraints on the media organization that affect the production of news; these constraints result from internal and external factors. Whether to produce news that attracts consumers while maximizing profit (McManus, 1995) or to adapt to limited organizational resources and an infinite supply of news (Shoemaker and Reese, 1996), media routines are established and followed every day in the newsroom. Both McManus (1995) and Shoemaker and Reese (1996) agree that these routines are developed to meet the needs of media organizations; however, routines can act as a set of constraints, ultimately affecting the content of journalists' work.
Media Routines
Although the theoretical analysis of what affects news content is relatively recent, concerns regarding the development of news have been around since as early as the 1970s. For example, Sigal (1973) recognized that news-gathering organizations, like most other organizations, engage in and follow established routines. Sigal felt that these routines take on a dangerous life of their own. Although news organizations do alter their procedures and at times make innovative responses to certain situations, this behavior is not always the norm. Sigal identified the regularity of newspaper production, deadlines, legwork and contacts, and beats as standard operating procedures of newsgathering; having to fill the newspaper with news on a daily basis leaves journalists little time for reflection (Sigal, 1973).
Tuchman (1978) agreed that in order for everyday occurrences to be recognized and reconstructed into news, news organizations typically engage in and follow routines. Tuchman asserted that news is arrived at by negotiation between source and reporter, between reporter and reporter, and between reporter and editor, and that as a result, the news presented to the public is largely news that legitimates the status quo.
Fishman's (1980) argument that routine methods of gathering news determine the 'ideological character' of the news product echoed the conclusions of Sigal (1973) and Tuchman (1978). Fishman explained how beats, "a complex object of reporting consisting of a domain of activities occurring outside the newsroom" (p.28), tend to route reporters exclusively through governmental agencies and corporate bureaucracies in their search for news. Reporters under daily deadlines come to depend upon these organizations for the predictable, steady flow of raw news material they provide. Therefore, news content is merely a reflection of this bureaucratically constructed reality, as opposed to a picture of reality itself.
Sources of the News
In constructing reality, or as Shoemaker and Reese (1996) described it, manufacturing symbolic content, " the media rely on external suppliers of raw material, whether speeches, interviews, corporate reports, or government hearings. These suppliers, or sources, have a major influence on media content" (Shoemaker & Reese, 1996, p.105). The media's reliance on these sources for content makes media organizations vulnerable to patterns that are dictated by sources. Journalists are forced to adapt to the constraints imposed by their sources.
Gans (1979) suggested that successful access to journalists is shaped by at least four factors: sources' incentives to be in the media, sources' power, sources' ability to supply suitable information, and sources geographic and social proximity to journalists. According to Gans, not all sources are equally likely to be contacted by journalists-those with economic or political power are more likely to influence news reports than those who lack power. More recent studies suggest that Gans is correct in his assertion that all sources are not equally likely to be used by journalists.
Government Sources in Print News
Sigal (1973) had found that The New York Times and The Washington Post reporters were more likely to rely on official sources for information than other types of sources. Sigal's study examined the use of routine channels for the news and how reliance on these routine channels resulted in an unfair advantage for official sources. This analysis of The New York Times and The Washington Post, found that 58.2 percent of these newspapers' front-page stories were received through routine channels, while 25.8 percent were coded as "enterprise" channels. Routine channels were defined as official proceedings, press releases, press conferences, and nonspontaneous events such as speeches or ceremonies, while enterprise channel news is developed through interviews, spontaneous events or independent research. Looking at sources, Sigal found that 81.3 percent of all sources were officials in some capacity. Sigal concluded that when journalists adhere to routine channels of newsgathering,
they leave news selection decisions largely up to official news sources.
Approximately 10 years later, Fico (1985) compared statehouse reporters' use of state government sources to wire service and local newspaper reporters. The study analyzed a 12-day sample of stories about the Michigan statehouse written in 1984 and included 270 stories. The stories in the sample were drawn from six newspapers with Capitol bureaus and the two wire services. Fico specifically investigated if constraints that arise from geographical location and medium type affect the number and diversity of sources used by reporters. Findings indicated that wire service and local newspaper reporters use fewer and less diverse sources than newspaper statehouse bureau reporters. However, the differences in source use among these reporters were not statistically significant. Conclusions drawn from these findings were that the more constrained reporters are by technological deadline factors and geographic location, the narrower the range of sources and the less often diverse sources will be used in stories. Still, the similarity in overall source priorities of the three different types of reporters suggeste
d that they all respond to similar opportunities and constraints in the way the institution generates sources.
Brown et al. (1987) took another look at the nature and extent of diversity in news sources used by newspapers. Brown conducted a content analysis of 846 front-page, straight-news stories that appeared in two national newspapers, The New York Times and The Washington Post, and in four North Carolina local newspapers. These results revealed that front-page news stories in both the national and local press, including stories from the wire services, rely heavily on government sources, who are primarily men in executive positions. Brown's results also compared the sourcing patterns in national and local news stories and found that local stories included fewer U.S. government sources and more state and local government sources than national stories. Stories in local newspapers also attributed more information to non-government sources than did stories in national newspapers, leading Brown to the conclusion that local newspapers, at least in North Carolina, appear to allow for mor
e diverse representation of sources. However, Brown did not compare the use of sources among these local newspapers, but rather compared the local group of newspapers to the national. There may be many variations in the way stories are reported from one local newspaper to the next. Furthermore, since these local newspapers were drawn from the same state and are limited in number, one cannot generalize about the use of sources in local newspapers.
Hallin et al. (1993) analyzed the sourcing patterns of 23 defense and national security reporters from seven major national newspapers, The New York Times, The Washington Post, the Los Angeles Times, The Boston Globe, The Wall Street Journal, The Washington Times, and the Chicago Tribune. As Hallin et al. had predicted, government sources strongly dominated the national security reporting of these seven papers. Hallin et al. said that these results confirm countless studies that have shown that government officials are consistently dominant in reporting by American journalists.
The literature suggests that both television (Berkowitz, 1988 & Whitney et al., 1989) and print news reports rely on official and government sources, especially in national media. Consequently sources, specifically powerful sources, have a very large influence on the way the news is presented in the media. The preceding studies all suggest that the American media fall short when it comes to reaching out to a variety of sources. Has this trend been increasing or decreasing over the past two decades? Soley (1994) examined whether today's newspaper journalists are seeking to increase the diversity of sources used in their stories. Soley asked whether experts are being used as sources in newspaper reports more in the 90s than 15 years ago. Soley's comparative content analysis of 945 front-page and jump stories published in 1978 and 1990 in The New York Times, the Los Angeles Times and The Washington Post, found a substantial increase in the use of 'experts' in newspaper reports during the last 15 years. Therefore, it seems that the standard practice of using official or 'expert' sources in news reporting was occurring more frequently than before.
However, in the last decade, civic journalism, a movement toward a more community-oriented approach to news, has taken center stage in the source debate. Civic journalism seeks to encourage journalists to reach out to citizens as sources and resources, bringing citizen voices and ideas to the foreground. Proponents of civic journalism suggest that journalists can maintain their objectivity while simultaneously inspiring citizen involvement in community issues (Kramer, 1995). However, to do this, journalists must part ways with the usual routine of reporting only two sides of an issue and recognize that there are multiple sides that should be identified (Reeves, 1999). Recognizing these multiple perspectives means that journalists must look at local non-institutional sources and provide local perspective and connections on local, national and world events (Civic Journalism, 1999). Furthermore, in order to examine whether or not this is actually happening, researchers might nee
d to re-examine the issue of diversity of sources in the press at the local level.
Local Daily Newspapers:
As the literature suggests, government sources tend to dominate news coverage of national and international affairs; however, a number of gaps still exist in our knowledge of sourcing routines of American journalists. Many studies have examined the content of national newspapers, such as The New York Times and The Washington Post (Sigal, 1973; Brown et al., 1987; Hallin et al., 1993; Smith, 1993). However, until Phyllis Kaniss's (1991) study of the local newsmaking process, few studies had looked extensively at local newspapers to examine source use.
Kaniss attributed the neglect of local news media practices to a common misperception that local news is inconsequential. According to Kaniss, the truth is that local news has a major impact on the policy decisions of local officials and on the private
decisions of its readers. However, because local news media are so difficult to examine on a general basis, many researchers have overlooked it.
Kaniss' study was "based on a seven-year research effort that included analysis of data on local newspaper markets throughout the country as well as an examination of local news coverage of urban and regional developmental issues in a number of those markets (p.9)." It was also based on over 100 interviews conducted with journalists and local officials in Philadelphia, Houston, Detroit, and Seattle. A portion of the study was devoted to local journalists' professional standards and how these standards can influence local policy coverage. Kaniss specifically examined how these standards lead the journalist to accept without question the information provided to them by government sources.
Like Gans (1979), Kaniss (1991) found that the suitability and accessibility of city official sources was related to the positive coverage that city development projects received. Kaniss asserted that city government officials' access to media relations for the promotion of city development projects made it more likely that their projects would receive coverage than suburban development projects. Furthermore, deadlines that are placed on local journalists to produce stories prompted them to rely heavily on the information that was provided to them by these officials. As a result, news coverage of these projects tended to be positive.
Kaniss also found a particularly high reliance on official sources by local reporters for stories that were related to quantitative issues. She attributed this to journalists' 'fear of numbers.' Kaniss sited specific examples of journalists' tendencies to accept quantitative estimates by city-officials for development projects without questioning how they were derived. These journalists also presented city-officials' development estimates in news coverage much as they were packaged by the source that produced them, regardless of the source's self-interest in the estimate. Kaniss concluded that public officials' media skills, along with the complexities of quantitative issues, leads to a bias in the way public policies, positions, and officials are presented in the local press and that this is cause for concern.
According to Pollock (1995), this type of community focus is the most "promising analytical level to probe for systematic variations in newspaper reporting" (p. 168). Although Pollack was specifically referring to employment issues, the same focus could be applied to a wide range of issues. For example, Tichenor, Donohue, and Olien (1980) had concluded that newspapers in relatively large cities have greater access to resources and are unlikely to be easily controlled by a single set of political or economic elites. Similarly, Demers (1993) found that editors in larger newspapers have more autonomy. Therefore it is reasonable to conclude that the community size affects newspaper content.
Carroll (1989) looked specifically at the relationship between a community's size and the way news is reported. In Carroll's analysis of 161 news programs presented by 57 television news stations, the larger the market size, the more the television station covered spontaneous news events, while smaller-market stations relied on stories that could be planned. He attributed this difference to lower budgets and small staffs of smaller-market stations. Lacy and Berstein (1988) supported these results in their content analysis of 114 daily newspapers. They found that larger newspapers were more likely to generate their own content, while smaller papers relied more on wire-service copy and editorials. Beam (1993) also found that newspaper circulation size correlates positively with community size.
This research attempts to further the examination of the possibility that media in smaller markets differ in their use of sources from larger markets. Fico (1984) argued that the more constraints a reporter operates under, such as deadlines and geographic location, the narrower is the range of sources relied on for stories. Beam's (1993) findings suggested that in today's one-newspaper markets circulation size can serve as a proxy for market size in the newspaper industry. Therefore the question is restated: Are consumers in smaller markets receiving more or less source diversity than in larger markets?
Summary
Research suggests that there is a lack of diversity in the sources that are used by the national press. Journalists are predominately using government and expert sources to report the news; therefore, the public is exposed to more information from these select interest groups.
This study proposes to determine types of sources used by the local press and to examine whether circulation size is related to the sources these journalists use. While newspapers in general are prone to heavy use of government sources, is the size of a newspaper related to the extent of this use? This study asks the question: Do smaller daily
newspapers, in comparison to larger daily newspapers, depend more heavily on government sources? Stated as hypotheses:
Hypothesis One
H1: Circulation size will be related to the diversity of sources in local, staff-written, front-page newspaper stories.
Hypothesis Two
H2: The smaller the circulation, the greater the reliance on government sources for local, staff-written, front-page stories.
METHODOLOGY
Population and Sample
This study examined the use of sources in front-page stories from various-sized daily local newspapers in the United States. A pretest was conducted in 1998 on 27 front-page lead stories in six issues of three Louisiana daily newspapers to test the proposed methodology.
The actual population of the current study was sets of 70 daily newspaper front pages purchased from the Newseum at the Freedom Forum World Center headquarters in Arlington, VA. Three issues (Friday, December 4, 1998; Saturday, February 20,1999; and Thursday, March 18, 1999) were received from the Newseum. Because of the difference in the news base of Sunday newspapers versus weekday and Saturday newspapers, Sunday issues were not included in the sample (Stempel & Westley, 1989). This restriction allowed for a more consistent examination of the diversity of news coverage. Following its policy, the Newseum selected these three issues based on the availability of papers that they are prepared to send to the public. Although the Newseum's policy does not allow for a random sample, the large number of papers included allows for an extensive scope of news coverage. As noted, Brown et al. (1987) and Kern (1983) only used a combination of two national newspapers and seven local newspapers. By enlarging the quantity of local newspapers, large and small, this study's sample is more representative of the population than other samples in prior studies.
Stratification
The sample analyzed in this study was the front pages of 18 local newspapers. To keep the focus of the study on local news, foreign, national, and collegiate newspapers were eliminated from the original sample of 70, narrowing the total number of newspapers to 46.
The papers were stratified by size. To stratify for size these three sets of 46 newspapers were grouped into three categories based on circulation. The Editor and Publisher Yearbook (1998) list of top 100 newspapers includes papers with a circulation size greater than 100,000. Small dailies, as defined by Beam (1993) include any daily with a circulation less than 50,000, and medium will be defined as newspapers in between small and large circulation sizes. Therefore, the three circulation categories are large- circulation dailies (100,000 and greater), medium-circulation dailies (50,000- 99,999) and small-circulation dailies (below 50,000). See Appendix I for a list of newspapers within each circulation category. Of the 46 newspapers, 24 were large-circulation, 12 were medium-circulation, and 10 were small-circulation. Six newspapers were then randomly selected from each group, resulting in a sample size of 18.
Coding
All local, staff-written, front-page stories from each of the three days were coded. To code the front-page newspapers, local was defined in terms of proximity. Stories were considered local if the story concerned people, events or institutions in the immediate coverage area of the newspaper organization (Bridges, 1987). Bridges defined coverage area as news coverage within a 30-mile radius. Because only the front page of each paper was purchased, jumps will not be included in the analysis. McCombs (1987) found that the front-page of a section was the key variable in predicting readership for all news items in the paper. Newspaper stories are also often written in an inverted pyramid style, where facts are listed in order of decreasing importance (Shoemaker, 1996). Furthermore, surveys consistently show that readers often stop reading stories if they jump from the front page to an inside page (Friend et al., 1999). Therefore, restricting this analysis to front-page stories
is appropriate, because not only is the information on the front page the
most important, but readers are less likely to read the information contained within the jumps.
Staff-written was defined as any story not clearly identified as coming from a wire service, syndicate or other non-staff source (Culbertson, 1975). A preliminary analysis of the December 4 set of newspapers estimated that local, staff-written stories on the front pages would make up approximately three stories per issue or 166 total stories.
For the purpose of this study two coders, a trained, independent coder and the author examined the content of each story. The trained coder identified each source cited within the stories and the context within which the identified source was speaking. Sources that were referenced more than once in a story were counted only once (Brown et al., 1987).
Next, to determine whether the content of local, daily front pages represents a distinctive pattern of source use among the three groups of dailies, each information source was coded for institutional affiliation. Using an adaptation of Brown et al.'s (1987) source affiliation categories, sources were grouped into seven possible affiliations:
Unaffiliated U.S. citizen: Sources categorized as unaffiliated U.S. citizens were any citizen who was not directly linked to any type of organization or affiliation associated with the story. Examples of unaffiliated citizens are sources who offer a specific
perspective for a story based on their role in society, such as a father or mother, homeowner, teenager, or average college student.
Affiliated U.S. citizen: Affiliated sources included any source who was connected to an organization or association related to the issue. Examples are business executives, attorneys, clergymen, law enforcement officials, members of associations, and organization representatives. Brown et al. (1987) looked at law firms, however, law enforcement officers were also included within this category when they were speaking about a crime event rather than speaking as a government representative. Only 10 of the total affiliated citizen sources (n=147) identified within the analysis were law enforcement officers that fell within this category, and they were distributed evenly among the newspaper circulation groups.
Local government official: Sources categorized as local government officials included elected and appointed local government officials speaking in their capacity. Examples are city council members and town mayors.
State government official: Sources categorized as state government officials included elected and appointed state government officials speaking in their capacity. Examples are governors and state representatives.
U.S. government official: Sources categorized as national government officials included elected and appointed national government officials speaking in their capacity. An example is the President of the United States.
Other Government Source: Government sources included any government workers or spokespersons who are paid by the government but were not included in the local, state, or U.S. government official categories speaking in their official capacity, e.g. director of public works, public information officer, or a member of the armed forces.
Other: This category was used when a source did not fit into any of the other six categories, e.g. an unidentified source, a report, etc. For examples of coded sources, see Appendix D.
The context in which the source was identified in the story determined the affiliation of that source. For example, if a government official who is also a parent was quoted in a story as a parent, he or she was coded as an unaffiliated U.S. citizen not a government official. If a local government official was speaking as a member of the Sierra Club, this source was coded as an affiliated U.S. citizen.
Reliability
To ensure coder reliability, Lacy and Riffe's (1996) formula for generating sample sizes necessary for valid reliability estimates was used to determine how many stories needed to be recoded serving as a reliability check. According to this formula, 36-45 content units are needed in the reliability test for a 95 percent assumed level of intercoder agreement and a 95 percent level of probability. Therefore, the author coded the first 10 stories of the sample analyzed by the trained coder and every fifth story thereafter. Upon the author's recoding of the first 10 stories, only two errors were found in the independent coder's source affiliation categorization of the stories' 17 coded sources. This resulted in a coder agreement of 88%. The documented errors were related to counting sources that were not directly quoted. The author advised the independent coder that sources that were not directly quoted, but whose comments were generally cited by the reporter, were considered to
be sources. Consequently, two additional sources were included within the analysis of the first 10 stories. The recoding of every fifth story after the initial 10 stories also revealed that the independent coder made four errors in documenting that local government officials were used as sources, resulting in a 99% coder agreement. On all four occasions local government officials' titles were unique to the geographic location of the newspapers, i.e. trustee and assemblyman, and were incorrectly coded as affiliated citizens by the independent coder. In all cases these sources' affiliations were changed to local government official.
Statistical Analysis
Using the Statistical Package for the Social Sciences (SPSS), version 9.0, and following Fico (1984) and Brown et al. (1987), crosstabulations identified the number of sources within each source affiliation category for each newspaper group. Chi square tests assessed the significance of differences in total number of sources by newspaper group. One-way chi-square tests determined the relationship of individual source affiliation category by newspaper group. Pearson's correlation coefficient was used to detect whether a significant, linear relationship existed between each source affiliation category and raw circulation size.
RESULTS
Description of the Sample
As stated, the 18 newspapers used for this study were grouped into large, medium, and small newspapers, based on circulation size. These three newspaper groups contained six newspapers each. The mean circulation of large newspapers was 342,849, medium circulation mean was 65,043, and small newspaper mean size was 26,933. Standard deviations indicated a wide range in newspaper size among groups (see table 1).
Table 1: Mean Circulation of Newspaper Groups
Mean
Std. Deviation
Large
342,849
100,101
Medium
65,043
12,230
Small
26,933
10,253
The front pages of the three issues of these 18 newspapers yielded a total of 171 locally produced stories. When the stories were compared by newspaper size, 31% of the total number of stories were found on the front pages of the large newspapers, 30% were found on the front pages of the medium newspapers, and 39% were found on the front pages of the small newspapers.
Frequencies were run to determine the total number of sources found in all of the stories included in the sample. The total number of sources in the sample of stories was 331. When the sources were compared by newspaper size, large newspaper stories contained 31% of the total sources in the sample; medium newspapers contained 30% of the total number of sources; and small newspapers used 39% of the sources. As these percentages reveal, small newspapers used slightly more sources than larger newspapers, however, the proportion of sources used was closely distributed among groups.
Of the 331 sources found in the local stories, almost half (45%) were affiliated citizens (see table 2). The second most common sources identified in local stories were unaffiliated citizens (20%). The third and fourth most frequently used sources were state and local government officials (13% and 11% respectively). The most infrequently identified sources included other government sources (6%), other sources (4%), and national government officials (1%).
Hypothesis Testing
In analyzing the use of sources in the sample data there were notable differences in the types of sources used by group. Chi-square analysis of the original 7x3 matrix approached significance, but low expected cell counts for national government sources decreased the applicability of the chi-square test. National government sources made up only 3 of the total 331 sources within the sample (2 for large newspapers, 1 for medium, and 0 for small). Theoretically this is understandable because the sample included only local stories. In order to get a better picture of the relationship between circulation size and source use, national government sources were removed from the chi-square analysis, and the resulting chi-square was significant (p= .048). These findings support the expectation that circulation size is related to the use of sources in local newspapers. Further analysis was conducted to identify where the differences exist in source use among the three newspaper groups.
In order to further explore source patterns, one-way chi-square tests were used to examine the relationship between each source-affiliation category by circulation group. As table 2 indicates, significant differences were found in both unaffiliated citizen source use (p=.055) and local government official source use (p=.017). Small newspapers used 23.4% unaffiliated citizen sources, medium newspapers used 22.2% unaffiliated citizen sources, while large newspapers used only 13.9%. These results indicate that smaller newspapers are more reliant on unaffiliated citizen sources than large newspapers. In terms of pluralistic expectations (Brown et al., 1987), smaller newspapers' use of unaffiliated sources implies that they are doing a good job of using different types of sources.
A different pattern was found in the examination of the newspaper groups' use of local government officials. The study revealed that small newspapers used 15.6% local government sources and large newspapers used only 10.9%. However, the smallest percentage of local government sources was found in medium newspapers (6.1%). In this case, the pattern of local government official source use is curvilinear. It is also important to note that smaller newspapers used fewer state government officials (10.9% and 12.1%) and fewer other government sources (1.6% and 5.1%) than large newspapers (17.8% and 5.9%). However, these differences were not significant.
The use of affiliated citizen sources was also interesting. As previously noted, almost half of the 331 sources were affiliated citizens. As table 2 indicates, the use of affiliated citizen sources by small, medium, and large newspapers was virtually the same.
Table 2: Source Affiliation by Newspaper Circulation Group
Large
Newspapers
(n=101)
Medium
Newspapers
(n=99)
Small
Newspapers
(n=128)
Total
(n=total # of sources)
One-Way
Chi-Square
Affiliated Citizen
45.5%
43.4%
45.3%
44.8%
(n=147)
(( = 2.571, df = 2, p = .276
*Unaffiliated Citizen
13.9
22.2
23.4
20.1
(n=66)
(( = 5.818, df = 2, p = .055
State Government Official
17.8
12.1
10.9
13.4
(n=44)
(( = 1.273, df = 2, p = .529
*Local Government Official
10.9
6.1
15.6
11.3
(n=37)
(( = 8.162, df = 2, p = .017
Other Government Source
5.9
11.1
3.1
6.4
(n=21)
(( = .156, df = 2,
p = .156
Other
5.9
5.1
1.6
4.0
(n=13)
(( = 2.0, df = 2,
p = .368
% of Overall Sample
30.8
30.2
39
100
(N=328)
Pearson Chi-Square ((( = 18.425, df =10, p = .048*); * = p ( .05
The small size of the individual newspaper groups (n=6) led the researcher to believe that Pearson's correlation coefficient may useful in testing the relationship between source affiliation use and circulation size. As table 3 indicates, Pearson's correlations supported the patterns indicated in the one-way chi-square analysis of unaffiliated citizen source use (r = -.363) and approached significance (p= .069). At first glance, correlational analysis of local government official source use was disappointing (r = -.097); however, it supports the non-linear pattern of the grouped data. There was no significance in any of the other source affiliation categories. The analyses supported the expectation that
circulation size will be related to source use, but these differences are limited to unaffiliated citizens and local government officials. Therefore, hypothesis 1 was partially supported.
Table 3: Correlations of Source Use by Raw Circulation Size
Source Affiliation Category
Correlation
Affiliated Citizen Source
r =-.204, p =.208
Unaffiliated Citizen Source
r = -.363, p = .069
State Government Official
r =.089, p =.3625
Local Government Official
r = -.097, p = .351
Other Government Source
r =-.095, p =.3540
Other Source
r =.127, p =.308
To examine the overall government source use by newspaper group the source affiliation categories were collapsed into two groups, government and non-government sources. Government sources were defined as local, state, and national government sources and the other government source category. Non-government sources included all unaffiliated and affiliated citizens and sources coded as "other". When these two variables were analyzed, more than twice as many non-government sources (68%) were used in local, staff-written stories than government sources (32%).
Sample percentages show a pattern that is the reverse of hypothesis 2. Large newspapers used fewer non-government sources and more government sources than medium or small newspapers (see table 4). Chi-square analysis indicated there was no significant difference in source use among groups for either category. Neither the chi-square nor the Pearson's correlation analysis supports the expectation that smaller local newspapers will rely more heavily on government sources than larger newspaper (see table 5). Therefore, hypothesis 2 was not supported.
Table 4: Government vs. Non-government Sources by Circulation Group
Large
Medium
Small
Total
One Way
Chi-Square
Total
Non-government Sources
64.1%
70%
70.3%
68.3%
(n=226)
(( = 4.389, df = 2, p = .111
Total
Government Sources
35.9
30
29.7
31.7
(n=105)
(( = 1.086, df = 2, p = .581
Pearson's Chi-Square ((( = 1.22, df = 2, p = .543)
Table 5: Correlations of Government and Non-government Source Use by Raw
Circulation Size
Source Affiliation Category
Correlation with Circulation Size
Non-government Sources
r=-.332, p =.0895
Government Sources
r=-.048, p =.4255
Although no hypothesis was stated regarding the newspaper circulation groups' use of unaffiliated citizen sources versus total government sources, the researcher thought it would be interesting to compare these uses to see how smaller newspapers fared. To make this comparison, sample data for hypothesis 1 and sample data for hypothesis 2 were compared. Table 6 indicates that small and medium local newspapers used similar amounts of unaffiliated citizen sources (23.4% and 22.2%) and government sources (29.7% and 30%). Larger local newspapers used over twice as many government sources (35.9%) as unaffiliated citizen sources (13.6%).
Table 6: Unaffiliated Citizen Source Use Compared to Total Government Source Use by Circulation Group
Large
Medium
Small
Unaffiliated Citizen Sources
13.9%
22.2%
23.4%
Total Government Sources
35.9
30
29.7
DISCUSSION
The present study's findings contribute to the local newsmaking process body of knowledge. This study specifically compared the use of 331 sources for 171 front-page, local, staff-written stories in 18 small, medium, and large non-national newspapers. The inclusion of size as a factor adds a new dynamic to the discussion of source use in local news. Although patterns were not uniform, this analysis supported the expectation that local newspaper size would be related to source use. A summary of the study's specific findings is as follows.
ù Almost half of the overall sources identified in the front-page, local, staff-written stories were affiliated citizens.
ù The high overall use of affiliated sources was virtually the same for small, medium, and large local newspapers.
ù The second most common sources identified in the front-page, local, staff-written stories were unaffiliated citizens.
ù Smaller newspapers used significantly more unaffiliated citizen sources than large local newspapers.
ù There was a significant difference in local government official source use among the three circulation groups; however the pattern was curvilinear, and the smallest newspapers used the greatest number.
ù Overall, twice as many non-government sources as government sources were found in the front-page, local, staff-written stories.
ù The pattern of government source use in the sample data was opposite of the pattern predicted. Smaller local newspapers used more non-government sources than large newspapers. This relationship approached significance in the Pearson's analysis. There was a weak pattern of decreasing government source use from large to small that was not significant.
ù Large local newspapers use more than twice as many total government sources as unaffiliated citizen sources.
As noted, the examination of the use of government sources vs. unaffiliated citizen sources by each group of newspapers revealed some surprising results. Research has also shown that government or official sources are typically the most accessible types of sources (Sigal, 1973). Therefore, it was reasonable to expect smaller local newspapers to use government sources more frequently than unaffiliated sources because of perceived greater accessibility. However, small local newspapers' use of unaffiliated citizens as sources did not confirm this assumption. Although the results did find that smaller newspapers were more likely to use local government officials as sources than larger newspapers, overall smaller local newspapers faired better in their use of unaffiliated citizen sources. Smaller local newspapers used similar amounts of unaffiliated citizen sources and overall government sources, while larger local newspapers used more than twice as many government sources as unaff
iliated citizen sources. These findings suggest that the limited resources of small newspapers do not impede their ability to contact a variety of sources when developing stories.
Although findings indicated that smaller newspapers use significantly more local government official sources than larger newspapers, findings did not support the expectation that smaller newspapers would be more likely to use significantly more total government sources. In fact, the pattern was reversed. Sample data indicated that smaller newspapers used fewer government sources than large local newspapers, although this difference was not significant. However, it is important to note that the overall use of
government sources in local newspapers was relatively low, which is also a significant finding.
Prior to Kaniss' (1991) research, limited data had been collected documenting source use among local newspapers. However, studies focusing on sourcing patterns of national newspapers were quite common. In general, these studies consistently over time found national newspapers to be heavily reliant on government sources (Sigal, 1973, Kern, 1983, & Brown et al., 1987). Kaniss (1991) also found a high reliance on official sources by local reporters. The results of this study did not confirm Kaniss's findings. On the contrary, the results indicate that small, medium, and large local newspapers are not heavily reliant on government sources, at least not in their showpiece, front-page stories. Overall, only 32% of sources used for local, front-page, staff written stories were government officials. In comparison to Sigal's (1973) study in which 81% of national newspaper sources were officials in some capacity and Brown et al.'s (1987) report that over half of the sources used in t
wo national newspapers were affiliated with some governmental body, local newspapers appear to be doing a good job of reaching out to non-governmental sources.
Recommendations for Future Research
This study prompts several suggestions for future research. A similar study analyzing the context in which sources are cited in a larger, random sample of local newspapers would be useful to the body of knowledge of local newspapers use of sources. This would allow us to make more concrete conclusions about the meaning of the data.
A closer examination of the use of affiliated citizen sources in local newspapers would also be a worthwhile research topic. The most common source used among all groups of local newspapers in this study was affiliated citizens. By definition, affiliated sources were individuals who were connected to an organization or association related to the issue. This affiliation included business executives, members of associations, law enforcement officials, etc. Although local newspapers are doing well in terms of government source use, their heavy reliance on affiliated sources could indicate a bias toward the business community or associations. This in turn could lead to a bias in reporting. Therefore, an investigation of local newspapers' connections to affiliated sources may be useful. If this investigation were done, the author suggests reanalyzing the affiliated sources without law enforcement officials. This study included both law firms, following Brown et al. (1987), and law enforcement officers as affiliated citizen sources. An analysis without this group would be more conservative.
References
Altheide, D. (1976). Creating Reality: How TV News Distorts Events. Beverly Hills, CA: Sage.
Beam, R. (1993). The impact of group ownership variables on organizational
professionalism at daily newspapers. Journalism Quarterly, 70 (4). 907-918.
Berkowitz, D. (1988). TV news sources and news channels: A study in agenda building. Journalism Quarterly, 64 (2,3), 508-513.
Bridges, J. (1987). News use on the front pages of the American daily. Journalism Quarterly, 332-337.
Brown, J., Bybee, C., Wearden, S., & Straughan, D. (1987). Invisible Power: Newspaper news sources and the limits of diversity. Journalism Quarterly, 64 (1), 45-54.
Carroll, R. (1989). Market size and TV news values. Journalism Quarterly, 66, 49-56.
Culbertson, H. (1975). Veiled attribution-An element of style? Journalism Quarterly, 52, 456-465.
Demers, D. (1993). Effect of corporate structure on autonomy of top editors at U.S. dailies. Journalism Quarterly, 70 (3), 499-508.
Editor & Publisher International Yearbook. (1998) New York: Editor & Publisher, Inc.
Fico, F. (1984). News coverage of part-time and full-time legislatures. Newspaper Research Journal, 6, 49-57.
Fico, F. (1985). Replication of findings in newspaper and wire service source use in statehouse coverage. Newspaper Research Journal, 7. 45-51.
Fishman, M. (1980). Manufacturing the News. Austin, TX: University of Texas Press.
The Freedom Forum Online. http://www.freedomforum.org/ (2/8/1999).
Friend, C., Challenger, D., McAdams, K. (1999). Contemporary Editing. Chapter 8: Making a long story short. Lincolnwood, Il: NTC/Contemporary Publishing Group.
Gans, H. (1979). Deciding what's news: A study of CBS Evening News, NBC Nightly News, Newsweek and Time. New York: Pantheon Books.
Hallin, D., Manoff, R., & Weddle, J. (1993). Sourcing patterns of national
security reporters. Journalism Quarterly, 70 (4) , 753-766.
Kaniss, P. (1991). Making Local News. Chicago, IL: The University of Chicago Press.
Kramer, S. D. (1995). Civic journalism: Six case studies. CPN Journalism.
http://www.mediainfo.com/ (8/16/99).
Lacy, S., & Bernstein, J.M. (1988). Daily newspaper content's relationship to publication cycle and circulation size. Newspaper Research Journal, 9, 49-58.
Lacy, S. & Riffe, D. (1996). Sampling error and selecting intercoder reliability
samples for nominal content categories. Journalism and Mass Communication Quarterly, 73 (4), 963-973.
Lowery, S. & DeFleur, M. (1995). Milestones in Mass Communication Research. White Plains, NY: Longman Publishers USA.
Mapping seminars peel back layers of civic life. (1999). Civic Catalyst, Summer, 12-13.
McCombs, M. & Shaw, D. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36, 176-187.
McCombs, M., Mauro, J., & Son, J. (1987). Predicting newspaper readership from content characteristics: A replication. A paper presented to the national meeting
of the Association for Education in Journalism and Mass Communication, San Antonio, TX, August, 1987.
McCombs, M. (1992). Explorers and surveyors: Expanding strategies for
agenda setting research. Journalism Quarterly, 69 (4).
McManus, J. (1995). Market-Driven Journalism: Let the Citizen Beware?
Thousand Oaks, CA: Sage Publications.
Norusis, M.. (1990). SPSS: Base System User's Guide. Chicago, IL: SPSS Inc.
Pollock, J. (1995). City characteristics and nationwide newspaper coverage of the NAFTA agreement. Mass Communciation Review, 22 (3,4), 166-177.
Reeves, R. (1999). Rends or mends. Trust, Spring, 8-14.
Sansom, L. (Summer, 1990). A little survey shows improvement. The Professional Communicator, 6-7.
Severin, W., & Tankard, J. (1997). Communication Theories: Origins, Methods, and Uses in the Mass Media. White Plains, NY: Longman Publishers USA.
Shoemaker, P. & Reese, S. (1996). Mediating the Message: Theories of Influences on Mass Media Content. White Plains, NY: Longman Publishers USA.
Sigal, L. (1973). Reporters and Officials. Lexington, MA: D.C. Heath and Company.
Soley, L. (1994). Pundits in print: "Experts" and their use in newspaper stories. Newspaper Research Journal, 15 (2) , 65-75.
Stempel, G. , "Content Analysis," in Stempel, G. & Westley, B., eds. (1989), Research Methods in Mass Communication., 35, Upper Saddle River, NJ: Prentice Hall.
Tichenor, P., Donohue, G., & Olien, C. (1980). Community Conflict and the Press. Beverly Hills, CA: Sage.
Tuchman, G. (1978). Making News: A Study in the Construction of Reality. New York, NY: Free Press.
Weaver, N., & Elliot, S. (1985). Who sets the agenda for the media? A study of local agenda setting. Journalism Quarterly, 62 , 87-94.
Whitney, D., Fritzler, M., Jones, S., Mazarella, S., & Rakow, L. (1989). Geographic and source biases in network television news, 1982-1984. Journal of Broadcasting & Electronic Media, 33, 137-77.
(Appendix A)
GROUP I- LARGE Above 100,000
Newspaper
City, State
Circulation Size
1. Chicago Tribune
2. Dallas Morning News
3. Boston Globe
4. The Arizona Republic
5. Philadelphia Inquirer
6. Star Tribune
7. The Denver Post
8. St. Louis Dispatch
9. Sun
10. Denver Rocky Mtn. News
11. The Atlanta Constitution
12. The Courier Journal (KY)
13. Omaha World Herald
14. The Indianapolis Star
15. Richmond Times Dispatch
16. Arkansas Democrat Gazette
17. The Providence Journal
18. The Tulsa World
19. Dayton Daily News
20. The News and Observer
21. The State
22. Albuquerque Journal
23. The Clarion Ledger
24. The Honolulu Advertiser
Chicago, IL
Dallas, TX
Boston, MA
Phoenix, AZ
Philadelphia, PA
Minneapolis, MN
Denver, CO
St. Louis, MO
Baltimore, MA
Denver, CO
Atlanta, GA
Louisville, KY
Omaha, NE
Indianapolis, IN
Richmond, VA
Little Rock, AR
Providence, RI
Tulsa, OK
Dayton, OH
Raleigh, NC
Columbia, SC
Alberquerque, NM
Jackson, MS
Honolulu, HI
653,554
481,032
476,996
437,118
428,233
387,412
337,372
313,594
312,826
302,953
296,669
228,185
225,761
224,372
209,690
170,766
170,292
162,186
159,072
153,408
121,699
113,694
104,375
102,236
GROUP II - MEDIUM 50,000-99,000
Newspaper
City, State
Circulation Size
25. Arizona Daily Star
26. The Oakland Tribune
27. The Christian Science Monitor
28. The Times
29. Anchorage Daily News
30. Reno Gazette -Journal
31. State Journal Register
32. Statesman Journal
33. Fairfax Journal
34. The Forum
35. Argus Leader
36. The Charleston Gazette
Tuscon, AZ
Oakland, CA
Boston, MA
Shreveport, LA
Anchorage, AL
Reno, NV
Springfield, IL
Salem, MA
Fairfax, VA
Fargo, ND
Sioux Fall, SD
Charleston, WV
91,561
86,017
78,233
75,552
71,273
66,682
63,590
58,381
57,823
54,714
51,197
51,000
GROUP III - SMALL below 50,000
Newspaper
City, State
Circulation
37. The Day
38. Sun Journal
39. The Anniston Star
40. Sun
41. Concord Monitor
42. Rutland Herald
43. The Leaf Chronicle
44. The Spectrum
45. Boca Raton News
46. Press and Dakotan
New London, CT
Lewiston, Maine
Anniston, AL
Bremerton, WA
Concord, NH
Rutland, VT
Clarksville, TN
St. George, UT
Boca Raton, FL
Yankton, SD
41,676
39,898
27,153
38,360
22,003
22,186
21,264
19,207
14,979
8,653
(Appendix B)
The Sample
GROUP I- LARGE Above 100,000
Newspaper
City, State
Circulation Size
1. The Arizona Republic
2. The Philadelphia Inquirer
3. Star Tribune
4. The Denver Post
5. The Atlanta Constitution
6. The Providence Journal
Phoenix, AZ
Philadelphia, PA
Minneapolis, MN
Denver, CO
Atlanta, GA
Providence, RI
437,118
428,233
387,412
337,372
296,669
170,292
GROUP II - MEDIUM 50,000-99,999
Newspaper
City, State
Circulation Size
7. The Oakland Tribune
8. Anchorage Daily News
9. State Journal Register
10. Greenbay Press-Gazette
11. Statesman Journal
12. The Charleston Gazette
Oakland, CA
Anchorage, AL
Springfield, IL
Greenbay, WI
Salem, MA
Charleston, WV
86,017
71,273
63,590
60,000
58,381
51,000
GROUP III - SMALL below 50,000
Newspaper
City, State
Circulation
13. Sun Journal
14. Sun
15. The Anniston Star
16. Concord Monitor
17. The Spectrum
18. Boca Raton News
Lewiston, Maine
Bremerton, WA
Anniston, AL
Concord, NH
St. George, UT
Boca Raton, FL
39,898
38,360
27,153
22,003
19,207
14,979
(Appendix C)
Coding Sheet
Name of Paper ________________________________ ID ______ Week Day ______
Issue
Mo-Day-Yr. _____/_____/_____ Circulation ________________
Story ID
01
02
03
04
05
06
07
08
09
10
Unaffiliated
U.S. Citizen
Affiliated U.S.
Citizen
Local Government Official
State Government Official
U.S. Government Official
Government Source
Other
(Appendix D)
Coding Examples
Unaffiliated U.S. citizen: Sources categorized as unaffiliated U.S. citizens were any citizen who was not directly linked to any type of organization or affiliation associated with the story.
Examples are:
ù Town residents
ù Consumers
ù A young driver
ù Neighbors
ù A witness
Affiliated U.S. citizen: Affiliated sources included any source who was connected to an organization or association related to the issue.
Examples are:
ù Attorneys
ù An insurance consultant
ù A National Football League Head Coach
ù A spokeswoman for the Texas Lottery Commission
ù Union officers
ù A Motorola spokesperson
Local government official: Sources categorized as local government officials included elected and appointed local government officials speaking in their capacity.
Examples are:
ù District Judges
ù City Managers
ù Chairman of County Boards
ù Town Mayors
ù Town Officials
State government official: Sources categorized as state government officials included elected and appointed state government officials speaking in their capacity.
Examples are:
ù Secretary of the State Department of National Resources
ù State Representatives
U.S. government official: Sources categorized as national government officials included elected and appointed national government officials speaking in their capacity.
Examples are:
ù President of the United States
ù United States Senators
Other Government Source: Government sources included any government workers or spokespersons who are paid by the government but were not included in the local, state, or U.S. government official categories speaking in their official capacity.
Examples are:
ù Director of County Planning
ù Superintendent of Public Schools
ù Governors' Spokespersons
ù Public Information Officers
Other: This category was used when a source did not fit into any of the other six categories.
Examples are:
ù Reports
ù Policies