Americans use the mass media for much of their environmental information (Larson, Zimmerman, & Scherer, 1982; Pierce, Lee­Salmons, & Lovrich, 1988). It has been established that mass media reporting can influence which issues, including environmental issues, the public will perceive as important, a phenomenon known as agenda setting (e.g., Yagade & Dozier, 1990). Given this significant role of the media in public debate, it is worthwhile for environmental educators and others interested in the public’s relationship to the environment to ask exactly what the mass media are saying about environmental issues.

Problem Statement

The goal of this research is to catalog and compare profiles of the environmental reporting in two prominent representatives of American print media, Time and the New York Times. The profiles were produced through the content analysis of a portion of the environmental reporting from two years, 1991 and 1992. Three main questions were asked: 1) What are the characteristics of the environmental reporting in Time and the New York Times? 2) What differences in environmental reporting exist between Time and the New York Times? 3) What change is there in the character of environmental reporting in these publications in a presidential election year?

Academic interest in the character of environmental reporting has waxed and waned with the prominence of the environment in the public agenda. Content analyses on environmental reporting have been done occasionally, but it is a largely unexplored field. The following review presents studies of problems similar to the present one, to produce a broad profile of the environmental content of a national magazine and newspaper.

Interest in environmental reporting first surfaced in the aftermath of the explosion of environmental information which began in the late 1960s and peaked around Earth Day in 1970 (Rubin & Sachs, 1973). Clausing (1971) first applied content analysis to environmental reporting in fish and game magazines (e.g., Field and Stream). She asked whether these magazines were keeping up with broadening ecological awareness evident in the 1960s. To answer, she generated three dichotomous categories (i.e., Universal vs. Terrestrial, Urban vs. Rural, and Comprehensive vs. Fish and Game) with which to assess change in environmental information in these magazines from 1966 to 1968. She noted little change.

Belak (1972) built upon Clausing’s research. He looked for change in the environmental coverage over a longer period. He sampled fish and game magazines in 1908, 1934, and 1968 to 1970, “key years” in the environmental movement, and compared the results to Clausing’s. He also expanded the number of dichotomous categories into two sets of parameters, Natural and Cultural. The natural parameters were General versus Fish and Game, Universal versus Interior, Urban versus Rural and Multi-Regional versus Regional. The cultural parameters included six dichotomies: Aesthetic versus Non-aesthetic, Quality versus Quantity, Action versus Apathy, Integrate versus Subjugate, Humanistic versus Utilitarian, and Noneconomic versus Economic. Belak derived these categories from “leading environmental spokesmen.” He found that a broadening trend in environmental coverage was evident in 1968 to 1970, but until that point most environmental news in these magazines was related only to fish and game issues as it was in Clausing’s study.

Rubin and Sachs (1973) undertook a large-scale study of this trend toward increased environmental coverage. They analyzed the content of representatives of a variety of print media, including the San Francisco Chronicle, Reader’s Digest, and Time. They coded articles from 1965 to 1970 into environmental subject categories. They found that there was a definite increase in the number of environmental articles over the period. As for the subject categories, they noted that issues of population and air and water quality were discussed most often; issues of energy management were discussed very little.

Hoesterey and Bowman (1976) took the analysis of environmental coverage beyond quantifying the amount of coverage to ask what the environmental messages of certain magazines were. To answer this question, they applied the dichotomous categories constructed by Belak (1972), adding two dichotomous categories concerning the orientation of the organization, to two environmental magazines, Audubon and the Sierra Club Bulletin. They generated a profile of each magazine based on the categories. They found that both publications emphasized general, interior and rural problems, although the Sierra Club Bulletin included more multi-regional and urban issues than Audubon; both defined the environmental crisis aesthetically but still dealt with non-aesthetic arguments; and both showed strong interest in politics and increasing interest in political involvement.

Strodthoff, Hawkins, and Schoenfeld (1985) analyzed trends in environmental reporting with regard to a model of ideology diffusion. They hypothesized that environmental reporting would follow a pattern of increasing “routinization.” That is, environmental reporting would change in character from trying to describe a new and unfamiliar social movement to becoming a legitimate issue requiring ongoing coverage. They successfully supported this model by analyzing the content of two special interest magazines, Audubon and Environment, and two general audience magazines, Time and Saturday Review, from 1959 to 1979.

McGeachy (1989) found support for a similar hypothesis, that the placement of environmental articles in specialized “environment” sections decreased as environmental articles became “mainstreamed” and consequently included in general news and other sections. In addition, she extended part of the analysis begun by Rubin and Sachs (1973). She sought to study the change in the amount of environmental coverage in national consumer magazines, including Time, from 1961 to 1986. She used the five subject categories developed by Rubin and Sachs but added two, again coding by article. She found that environmental coverage increased during the late 1960s and early 1970s, then decreased through the later 1970s, and finally leveled off during the 1980s at a level slightly higher than the 1960s.

Howenstine (1987) studied a more narrowly-defined shift. Through content analysis of two national newspapers, the Washington Post and the New York Times, and two national newsweeklies, Time and Newsweek, he documented a distinct change from 1970 to 1982 in the viewpoint of environmental reporting. He found a shift away from reporting on environmental quality issues toward use of the environment as an resource for development.

In summary, broad analyses of the environmental reporting in national publications have been done successfully, but not often. The last broad study of national publications published, McGeachy (1989), covered up to 1986. Since then, there has been another increase in the amount of environmental reporting (Stocking & Leonard, 1990). It is time, then, for another look at environmental reporting. In addition to continuing the previous studies in this area, the present study also seeks to present revised categories for analysis by updating and expanding previous categories as well as adding new ones. Finally, this study seeks to present a mutually-exclusive category system to facilitate statistical analysis. Statistical analysis potentials have often been overlooked in content analysis in general (Stempel, 1989b); this is the case in environmental reporting analyses reviewed here as well.

Selection of Publications for Analysis

This study intends to contribute to a general picture of the environmental reporting of the national print media by constructing detailed profiles of two types of publications, a nationally-distributed newsweekly and a nationally-distributed daily newspaper. These dissimilar types were chosen for two reasons. First, both are prominent print media and therefore can contribute to the general picture of environmental reporting. Second, choosing two dissimilar publications offers opportunities for comparison, as follows.

A newsweekly was chosen because any editorial bias would likely be evident for two reasons. First, newsweeklies are limited in space in comparison to national newspapers and must therefore be more particular in which stories they select to publish. Second, newsweeklies traditionally offer more critical analysis of the news since much of that which they report has already been covered by the daily newspapers. Since story selection and critical analysis of the news are editorial functions, the significance of these functions to newsweeklies makes the appearance of editorial bias more likely.

A national newspaper was chosen to establish a baseline of the environmental news which was available for reporting. Because of its larger space for news and more frequent publication schedule it is able to report more, and a wider variety of, environmental stories than can a newsweekly. The articles in a newsweekly represent a subset of this larger body. The characteristics of this subset which differ significantly from the characteristics of the larger body, represented by the newspaper, are of greatest interest. The national newsweekly Time was chosen because it has the largest circulation of all newsweekly magazines (Audit Bureau of Circulations, 1992, cited in Hoffman, 1992); also, it has been used in similar environmental content analyses (Rubin & Sachs, 1973; Strodthoff et al., 1985; Howenstine, 1987; McGeachy, 1990 ). The national daily newspaper New York Times was chosen because of its wide distribution (Editor & Publisher Yearbook, 1991, cited in Hoffman, 1992), general recognition as a national standard of newspaper reporting and availability.

Selection of the Population

This content analysis was applied to six-month periods, March through August, in two years, 1991 and 1992. These years were chosen because they were likely to show dissimilarities for two reasons. First, 1992 was a presidential election year. Moreover, it was an election in which environmental issues were politically significant enough to support the nomination (and eventual election) of a vice-presidential candidate, Al Gore, with a reputation and record as an environmentalist. Second, in 1992 the United Nations Conference on the Environment and Development (UNCED, popularly known as the Earth Summit), sponsored by the United Nations World Commission on the Environment and Development, was held in Rio de Janeiro, Brazil. This event drew international attention toward environmental issues; there was no similar large-scale issue or event involving the environment in 1991. The question here is whether the profiles and comparisons will remain consistent for Time and the New York Times during these two dissimilar years.

Two six-month periods were chosen such that a six-month period (September 1991 through February 1992) separates the studied periods. This gap keeps the two periods more distinct from one another, reducing the problem of studying two consecutive years, which is that two years so near one another are more likely to share similarities than two years separated by more time. The same six months in each year were chosen to control for possible seasonal similarities in environmental reporting (e.g., spring storms).

Operational Definitions

Environmental article refers to those articles which discuss issues or events which involve the interaction of humans with the natural environment, either as they affect the natural environment (e.g., littering, logging, or camping), or are affected by natural phenomena (e.g., tornadoes, flooding, or insect plagues). Included are information and commentary items, such as editorials, news items, special reports, and so on. This inclusion of a broad range of articles reflects the diffusion of environmental articles throughout the publications which has been noted by Strodthoff et al. (1985) and McGeachy (1989). Letters to the editor, advertisements, and supplements directly sponsored by advertisers are not included because by their nature they are not necessarily reflective of the editorial philosophy of the publication. In the case of the New York Times, only items from nationally distributed sections were searched; this excluded the section entitled “Real Estate” as well as the regional journal sections. This was done to be consistent with the focus of this study on national media.

Articles which discuss activities of humans which affect other humans within the human-built environment but do not discuss the larger, natural environment are not included. For example, issues of human health which are related to a degradation in the general environment are included, for example, respiratory problems associated with smog, skin cancer from increased UV radiation, illnesses attributed to water pollution, and toxic dumps. In general, for health issues, the article is included if it refers to an environmental issue which influences more than human health alone. This distinction is made because studying the reporting of health and risks is a separate, large issue.

Category definitions. All articles were coded into a set of categories generated from the literature as well as during the pilot study, discussed below. In the interest of clarity, the categories are presented here in outline format.

The following category terminology is employed in the remainder of this proposal. Category refers to the six main topics, denoted by Roman numerals: Environmental Issue, Story Angle, Geopolitical Significance, Demographic Significance, Principal Player, and Concerns. Subcategory refers to the subdivisions of the categories, denoted below by capital letters.

I. Environmental Issue: the type of environmental issue which the article primarily discusses. Except where noted otherwise, all subcategories within this category but the last, Natural Phenomena, are taken directly from McGeachy (1989), who used, with modifications and additions, categories developed by Rubin and Sachs (1973).

A. Air Quality: articles discussing problems such as the following: smog, carbon monoxide, sulfur dioxide, and other pollutants from automobile exhaust, factory emissions, and other stationary sources; deterioration of the ozone layer; acid rain where the article concentrates on air quality rather than water quality; the effects of these things on animal and plant life; their costs to the economic system; and the methods of control (McGeachy, 1989).

B. Water Quality: articles discussing problems such as the following: factory wastes, sewage disposal and thermal discharges; acid rain where the article concentrates on water quality rather than air quality; the effects of these things on animal and plant life; their costs to the economic system; and the methods of control (McGeachy, 1989).

C. Human Population: articles discussing the concept of overpopulation and ways to prevent or cope with the increase (McGeachy, 1989).

D. Environmental Additives: articles discussing natural or chemical compounds artificially introduced into the ecosystem that concentrate through successive food-chain levels or cause upset in the ecosystem through destruction of a species with possible detrimental effects on plant and animal life (McGeachy, 1989). Examples include pesticides, herbicides, and radiation. Articles discussing toxic waste where the additive was the emphasis of the article rather than air or water quality are also covered by this category. Articles on intentional additives such as fluoride or food coloring of which only the effect on humans is discussed are not included in the study. Added for this study are wastes which present physical rather than toxicological problems, such as taking up landfill space (e.g., excess packaging) or physically injuring wildlife (e.g., beer can six-pack rings).

E. Energy Resources: articles discussing the management of energy supplies, including the supply of flowing water, coal, oil, natural gas, or fissionable materials available for the production of electricity (McGeachy, 1989). Specifically included here for this study are articles discussing oil spills where the article concentrates on the effect of the spill on energy production rather than the effect on the natural environment.

F. Wildlife or Wilderness Conservation: articles discussing the protection of endangered or threatened species or the protection of wilderness areas (McGeachy, 1989). For this study, wilderness is defined to include all natural communities such as forests, wetlands, prairies, coastal plains, coral reefs, and so on.

G. Environmental Movement: articles discussing the environmental movement or an environmental organization without concentrating on another specific environmental subject (McGeachy, 1989).

I. Natural Phenomena: articles discussing natural events (e.g., tornadoes, weather patterns) or properties (e.g., new drugs from natural sources) which involve or potentially involve humans but are not caused directly by human activities. This subcategory was added on the premise that, in addition to stories on environmental issues, reports on natural phenomena also influence humans’ attitudes toward the environment.

II. Story Angle: whether the article emphasizes the conflicts within or the solutions of the issue or the effects of the event. In the case of an oil spill, for example, does the article spend more time discussing the actual damage (Conflict) or future prevention efforts (Solution). This category is suggested by Lundberg’s study (1984) of the comprehensiveness of coverage in environmental issues in newsmagazines. Lundberg found that magazine coverage is weak in covering specific solutions, but strong in covering efforts of players in achieving or implementing solutions. The category as defined in the present study asks a broader question of whether the publications tend to emphasize conflicts in environmental stories rather than solutions.

A. Conflict: article emphasizes the problems associated with the issue or event, such as disagreements among players or negative effects.

B. Solution: article emphasizes efforts or explores options to resolve the issue or address the problem, such as compromises or cooperation among players.

III. Geopolitical Significance: the geopolitical range of the effect of the issue or event. The article is coded under the largest applicable area. This is an expansion of a category in Belak (1972) and Hoesterey and Bowman (1976).

A. International: article mentions more than one country, the event takes place in a country other than the U.S., or the global implications are discussed.

B. National: article mentions more than one region of the U.S. (but does not concern areas outside the U.S.) or is addressed in a general fashion stating or implying its significance to the average American. Region refers to one of the nine regional divisions determined by the U.S. Bureau of the Census, that is, Pacific (CA, WA, OR, HI, AK), Mountain (ID, MT, WY, CO, NM, AZ, NV, UT), West South Central (TX, OK, AR, LA), West North Central (IA, KS, MN, MO, ND, NE, SD), East North Central (IL, IN, MI, OH, WI), East South Central (AL, KY, TN, MS), South Atlantic (DE, FL, GA, MD, NC, SC, VA, WV), Middle Atlantic (NJ, NY, PA), and New England (CT, MA, ME, NH, RI, VT) (Congressional Information Service, 1992). Articles for which no location is specified are coded here on the assumption that these publications are written for a national audience; therefore, articles which mention no geopolitical designation are presumed national in significance.

C. Regional: article discusses the significance or effects of the issue or event for the population or area of more than one state, but all states mentioned are within one region.

D. Local: article discusses the significance of the event only in relation to the population or geographic area of one state or a smaller political demarcation.

IV. Demographic Significance: type of broadly-defined socioeconomic area in which the event or issue is based. If both urban and rural locales are mentioned, the article is coded under the locale discussed the most. The population of the locale mentioned was determined by 1990 U.S. Census Bureau figures. This category is based on a similar category in Clausing (1971), Belak (1972) and Hoesterey and Bowman (1976), although it has been operationalized differently here.

A. Urban: the article is mostly about a specific locale that includes a metropolitan area of at least 20,000 population.

B. Rural: the article is mostly about a specific locale that does not include a metropolitan area of at least 20,000 population.

C. Unspecified: the article is not about a specific locale.

V. Principal Player: the party with a role in the event or issue whom the article discusses the most. For example, if the article mentions government actions or involvement more than the actions or involvement of any other of the parties mention below, it is coded under government. The notion of principal player is included as a measure of who is considered newsworthy. The subcategories here were constructed during the pilot study. This interactive type of category construction, where categories are defined by analyzing a sample of articles, was used by Jensen (1989).

A. Government: the article primarily discusses the actions or stance of governing bodies or their representatives. This includes international, federal, state, and local governments and all agents thereof.

B. Business: the article primarily discusses the actions or stance of persons or organizations associated in the article with a commercial enterprise (e.g., oil company) or a pro-business agenda (e.g., Chamber of Commerce).

D. General Public: the article primarily discusses the effect of the issue or event on individuals or groups of private, unorganized citizens without mention of specific, organized action in response, or describes the actions of the general public without specific association to organized groups. Included here, for example, are descriptions of the effects of weather damage or contaminated water supplies on communities and discussions of the rate of participation of the general public in such things as recycling programs.

E. Scientists: the article primarily discusses scientific information concerning the issue or event or focuses on the professional actions or opinions of scientists. If the nonprofessional comments of the scientists are predominant, then the article is coded under the Principal Player subcategory with which the comments are most consistent. For example, if the article primarily discusses the technical analysis of a site by an environmental scientist, it is coded as Scientists. However, if the article primarily discusses the pro-environmental political activities of the environmental scientist, it is coded as Environmentalists.

F. Environmentalists: the article primarily discusses the actions or stance of persons or organized bodies associated in the article with a specifically pro-environment agenda.

VI. Concerns: the motivating factor which, according to the article, the players consider most important with regard to the issue, the writer cites as warranting informing the reader about the issue, or the writer discusses the most. The subcategories used here are a reduction of the more detailed systems of cultural parameters used by Belak (1972) and Hoesterey and Bowman (1976).

A. Economic: the article primarily discusses the economic concerns involved in the issue and all the factors which contribute to the economic concerns. Included here are articles which emphasize the expenses entailed (e.g., pollution control costs, energy taxes, and price increases), the profit motives (e.g., new technologies and new medicines), and the conservation of resources to assure future economic health (e.g., monitoring energy supplies and managing forest lands to assure continuous lumber supplies).

B. Intrinsic: the article primarily discusses concern for the intrinsic value of a natural component, characteristic, or phenomenon where the economic and safety concerns involved are not predominant in the article. Included here, for example, are articles which emphasize preserving nature for its spiritual value or its beauty and preserving species or habitats for their own sake or for the health of the ecosystem without reference to direct benefits for human health or wealth.

C. Safety: the article primarily discusses concern for the influences of the issue on human health and safety. Included here, for example, are articles discussing injury caused by severe weather and contamination of water supplies by toxic waste.

Detailing the operational category definitions. The following procedure was used to detail the above definitions before the selection of the final study articles began. This was done to assure that the categories are sufficiently defined such that others may understand and use them. Moreover, since the publications studied here are targeted at a broad audience, it is important that the coding is to some degree consistent with how the general public would classify the articles.

The author selected two articles from the pilot study and presented them along with the operational definitions given above to a group of independent coders. The coders were eight undergraduate students at Slippery Rock University. The coders were independent in that they were persons with whom the author has never discussed this study. The coders were similar to the general public in that they had no professional experience with content analysis or the relationships of the media and environmental issues. They were asked to code the articles.

The results of this exercise are presented in Table 1. Of the eight coding forms distributed, four were completed according to the instructions. The remaining four were not completed correctly (i.e., more than one subcategory for each category was selected) and were disregarded.

Based on these results, only one category definition caused a problem for other readers, that of Principal Player, for which only 25% of the coding matched that of the author’s; for the other categories, 63% or greater of the coding matched that of the author. For Principal Player, there was no particular subcategory under which the coders consistently coded the stories when they did not code the same as the author. This seemed to indicate that the problem stemmed from an incomplete description of the concept of Principal Player rather than a misleading subcategory definition. Consequently, the category definition was clarified.

Population

The population included all environmental articles appearing in the regular weekly issues of Time for the months of March, April, May, June, July, and August for years 1991 and 1992 and the corresponding Sunday issues of the New York Times. Because the date which appears on the cover of Time is actually one week after the date on which the magazine is distributed, corresponding issue refers to the Sunday one week preceding the date appearing on the issue of Time. Only the Sunday edition of the New York Times was analyzed (rather than a random selection from all of the week’s editions as done in the pilot study) to reduce the effect of comparing two different publication formats. The Sunday edition includes a review of the week’s stories and has a broader circulation than the weekday editions. Thus, the Sunday edition is more similar to the newsweekly format of Time than the weekday editions are.

For both publications, all articles fitting the definition of an environmental article as given above were coded. All nationally-distributed sections of both publications were searched for relevant articles since it was noted in the pilot study that articles concerning the environment were found in all sections. This is consistent with Strodthoff et al. (1985) and McGeachy (1989) who documented the mainstreaming of environmental stories. The final sample sizes were 46 environmental articles from Time 1991, 49 from Time 1992, 148 from the New York Times 1991 and 155 from the New York Times 1992.

Data Collection

The data collection instrument was a checklist comprised of six categories consisting of two to eight mutually exclusive subcategories. All subcategories are mutually exclusive within their categories to allow for more statistical analysis than has been customary in previous environmental content analysis studies.

The unit of analysis is the article. Each article was coded into one of the subcategories under each category. This unit is large for content analysis. However, this study seeks to produce a broad profile of the publication for a large set of categories and subcategories. For this reason the article is considered sufficient to answer the questions asked. Furthermore, coding by article was done by two similar studies, Rubin and Sachs (1973) and McGeachy (1989). Only the headlines and text of the article were used. Pictures, diagrams, and their captions were not considered in the coding. If several subcategories appeared to share emphasis, then the paragraphs were individually categorized and totaled and the article was coded under the category which applied to the most paragraphs.

Reliability

Ten articles were randomly drawn from the population and coded by an independent coder (Glynn & Tims, 1982; Bowles & Bromley, 1992). The coder was independent in that this person had no prior experience with content analysis of environmental issues in the media nor had the independent coder discussed the present study with the author prior to coding. Reliability was calculated using Holsti’s formula (Howenstine, 1987): Reliability = M / N where M = number of items in agreement and N = number of items coded by each coder. A separate reliability was calculated for each category since the categories were independent of one another. Environmental Issue had a reliability of .80; Story Angle, .90; Geopolitical Significance, .90; Demographic Significance, .60; Principal Player, .80; and Concerns, .90. The average reliability for all categories was .80.

Data Analysis

The purpose of the study is to determine whether there are any significant differences in the environmental reporting between the publications and between the two years of the study for each publication. Descriptive and inferential data analysis were employed for the following comparisons: 1) Time 1991 versus the New York Times 1991; 2) Time 1992 versus the New York Times 1992; 3) Time 1991 versus Time 1992; and 4) the New York Times 1991 versus the New York Times 1992. The categories were also analyzed for correlations.

First, for each publication and year, percentages of articles in each subcategory were calculated by dividing the number of articles coded in the subcategory by the total number of articles coded. Significant differences were reported (p < .05).

Second, each category was analyzed for significant differences using chi-square (p < .05). The same four comparisons were made. The use of an inferential statistic is possible because all data are at the nominal level and all subcategories are mutually exclusive within their categories. The use of inferential statistics was not reported in any of the reviewed environmental content analyses. Indeed, Stempel (1989b) notes that the statistical analysis possibilities are often overlooked in content analyses. It is hoped that the present research will contribute toward remedying this oversight, at least with regard to the content analysis of environmental reporting.

Finally, to check for associations between the categories, Pearson’s product-moment correlation coefficient (Sproull, 1988) was computed for the categories of Geopolitical Significance, Demographic Significance, Principal Player, and Concerns. These four categories appeared to have the highest likelihood of association based on an informal analysis of the patterns of the individual articles. That is, for example, it appeared that articles falling into the Unspecified Demographic Significance subcategory tended to be International or National in Geopolitical Significance as well. Based on this informal analysis, a 1, 0, -1 dummy coding system was employed in order to prepare the data for correlational computations. Table 2 shows the values assigned to each subcategory. Correlation coefficients were calculated for these data sets: Time 1991, Time 1992, and for both Time samples combined; New York Times 1991, New York Times 1992, and for both New York Times samples combined; and for all samples combined.

Method

A pilot study was conducted to determine a satisfactory categorization scheme and method for analysis.

Sampling. The population for the pilot study was a total of thirteen weeks of Time and the New York Times: five weeks in 1993 and eight weeks in 1992. First, the population of Time was analyzed. The 1993 issues were those available in loose copies at the time the pilot study was begun. The 1992 issues were those present on the microfilm which corresponded to roughly one year previous to the 1993 issues. All articles which fit the definition of environmental article given above were selected.

Second, the New York Times was analyzed. First, the dates of all issues of the week previous to the corresponding date of issue of each selected Time were assembled. Then, one New York Times issue from each week was randomly selected by replacement sampling. All articles which fit the definition of environmental article were selected.

It was noted during this pilot study that the day of the week selected for the New York Times had an effect on the number of articles found. This raised the problem of picking the wrong day and thereby not having a fair standard to which to compare Time. Consequently, the Sunday New York Times was used in the final study, for the reasons discussed above (see Population).

Data Collection. The articles for both publications were coded into categories. The categories were taken from the literature or constructed during the pilot study as noted above (see Operational Definitions). The data are given in Appendix A

Results

Table 3 shows the results of the pilot study for each category. For each subcategory, the percent given in the “Time” and the “N. Y. Times” columns are the number of articles coded in that subcategory divided by the number of Time or New York Times articles coded respectively; it represents what percentage of the articles analyzed were coded in that subcategory. The column “Time­N. Y. Times” is the difference in percentage points between the publications for that particular subcategory.

However, for the Environmental Issue category, the results are also reported as percentages of total paragraphs. This weighted percentage was calculated by totaling the number of paragraphs in each story coded in the subcategory, then dividing by the total number of paragraphs coded for that publication. The weighting was done to determine if the size of the articles would significantly affect the results. The weighted and unweighted figures were not significantly different, so the weighting procedure was not used in the final study.

Discussion

Since the purpose of the pilot study was to establish a category system and methodology, no statistical analysis was performed. Therefore, the following discussion is strictly a qualitative comparison. Nevertheless, the data pointed to some interesting differences which indicated things to look for in the final study, as well as some potential problems with the method which were corrected in the final study.

For the category of Environmental Issue, the largest difference was in the subcategory Air Quality. This is due in part to a four-article spread in one issue of Time on ozone depletion which discussed the implications of recent research findings indicating an ozone hole over North America. The New York Times had only one article which was sampled in this study on the same news. This may be a consequence of sampling only one New York Times issue per week. In the final study, this effect of missing the big story in the New York Times was reduced by sampling only the Sunday New York Times with its weekly news review which was more likely to include the big news of the week.

In Geopolitical Significance, Time had noticeably fewer Local articles. This is not surprising given that the New York Times has a distinctly local readership as well as a national one. Similarly, it seems reasonable that Time would lend more space to the other subcategories. This seems to be the pattern.

The differences in the Demographic Significance category are not as easily explained. While it makes sense that both publications would tend to favor urban issues because most Americans, hence most readers, live in urban areas, Time had noticeably more Urban and less Rural articles, while the New York Times was fairly balanced between the two. Here is an example of a potential bias on the part of Time toward environmental issues affecting urban populations. The bias can be asserted because the New York Times, with a selection of articles presumably more representative of what stories are available by virtue of its larger newshole, indicates a more balanced body of news than that which Time is reporting. Of course, this pilot study is inadequate justification for the assertion of bias; hence, the final study uses a larger sample.

Another large difference appeared in the Principal Player category. Time seemed to include more information on scientists and their viewpoints than the New York Times; meanwhile, the New York Times focused more often on Government’s role in environmental issues. This may have been a function of Time’s more analytical style as a weekly newsmagazine relative to the New York Times’ style as a news-item oriented daily paper. Perhaps more interesting was a slightly larger emphasis by Time on Business in environmental articles than Government’s role, while the New York Times placed greater emphasis on Government than Business. Here too is a possible bias that warrants further research.

There are other smaller differences between the publications and within the publications among the subcategories (e.g., fewer Solution articles in Time than the New York Times, a different ratio of Concerns subcategories covered in the publications), but this pilot study was too limited to justify any conclusions about these. Nevertheless, there were enough differences to warrant the larger, more rigorous analysis of the final study.

A significant difference was found in only one category for one comparison. For Time 1991 versus the New York Times 1991, the category Demographic Significance was significantly different (c² = 10.40, 2 df, p < .05). The difference between the two publications in this category is that Time had more articles which did not discuss a specific location, that is, either Urban or Rural, while more New York Times articles discussed specific locations. In particular, the New York Times had a larger proportion of articles discussing Urban locations. There was no significant difference for this category in 1992.

Although no other significant differences were found for whole categories, there are significant percentage differences among individual subcategories.

Time

Table 4 summarizes the percentage differences for the two periods of Time analyzed and notes which are significant. Time dedicated just under 50 articles to environmental subjects in each six month period, or about two articles per issue. For Time 1991 versus Time 1992, a difference of 11% or greater is significant.

There were some differences for Time between 1991 and 1992. For Environmental Issue, there was a significant increase in the number of articles on the Environmental Movement (+22%), accompanied by a significant decrease in the number of Energy Resources articles (-13%). No other environmental issue changed significantly in coverage. No changes in the emphasis of Conflict versus Solutions were significant. Nor were there any significant changes in the proportion of Urban, Rural, or Unspecified story locations.

In 1992 Time coverage included significantly more articles of International Geopolitical Significance (+14%). In Principal Player representation for 1991 versus 1992, Environmentalists’ coverage did not change, but Government’s and Scientists’ significantly increased in 1992 (+16% and +13%, respectively), while coverage of Business and the General Public significantly decreased (-16% and -12%, respectively). Coverage of Economic Concerns significantly increased in emphasis in 1992 (+14%) while Intrinsic and Safety Concerns remained unchanged.

The New York Times

Table 5 summarizes the percentage differences for the two periods of the New York Times analyzed and notes which are significant. The New York Times dedicated around 150 articles to environmental topics per period studied, which is slightly over five and a half articles per Sunday edition. For the New York Times 1991 versus the New York Times 1992, a difference of 6% or greater is significant.

The New York Times had slightly fewer significant subcategory differences between 1991 and 1992 and these were generally of lesser magnitude than those in Time. For Environmental Issue, coverage of Natural Phenomena had a barely significant drop in 1992 (-6%) and the Environmental Movement, a barely significant rise (+6%). The drop in articles of National interest and increase in those of Local interest were also weakly significant (-7% and +7%, respectively). There was a decrease in articles of Urban focus and an increase in articles Unspecified in location (-8% and +13%). In the Principal Player representation, there was a significant increase in the number of Government-focused articles (+16%). The New York Times covered significantly fewer Business-oriented environmental articles in 1992 (-12%). There were no significant differences in Story Angle or Concerns between 1991 and 1992.

Time versus the New York Times

Table 6 summarizes the percentage differences for the two comparisons of Time to the New York Times and notes which are significant. (Appendix B shows the frequencies of articles and Appendix C shows the distributions.) For Time 1991 versus the New York Times 1991, a difference of 7% or greater is significant; for Time 1992 versus the New York Times 1992, a difference of 6% or greater is significant.

Most differences between the publications varied by the year. For 1991, the New York Times printed significantly fewer articles on Energy Resources and more Environmental Movement articles than Time (-12% and +9%, respectively). In 1992, the New York Times printed significantly fewer Environmental Movement articles (-7%) while the number of Energy Resource articles was not significantly different. In 1992 Time printed more articles on Natural Phenomena than did the New York Times (+15%). On Geopolitical Significance, the New York Times and Time were not significantly different in 1991, but in 1992 Time had significantly more International stories (+16%) and fewer Local stories (-17%) than the New York Times. In both years, the New York Times had significantly more articles Urban in focus (+23% in 1991 and +18% in 1992) and significantly fewer stories Unspecified in location than Time (-25% in 1991 and ­12% in 1992). As noted above, the publications were significantly different for the whole category of Demographic Significance in 1991. In both years, the New York Times had significantly more articles which reported primarily Government activity (+8 to +9%). For Concerns, the New York Times had significantly fewer articles on Safety both years (-13% in 1991 and -7% in 1992). In 1991, this difference was offset by a greater number of articles on Economic Concerns (+10%); in 1992, it was offset by a greater number of articles on Intrinsic Concerns (+13%). No other significant percentage differences between the publications were found.

Correlations between Categories

Table 7 shows the correlation coefficients for the categories of Geopolitical Significance, Demographic Significance, Principal Player, and Concerns. For the two sample years combined, Geopolitical Significance and Demographic Significance were moderately correlated in Time (r = .32, p < .01) and weakly correlated in the New York Times (r = .19, p < .01). Similarly, Principal Player and Concerns were somewhat more strongly correlated in Time (r = .34, p < .01) than in the New York Times (r = .22, p < .01). While Demographic Significance and Concerns were positively associated in the New York Times (r = .31, p < .01), they were negatively associated in Time (r = ­.25, p < .05). Geopolitical Significance and Concerns were weakly associated in the New York Times (r = .12, p < .05) but were not significantly associated in Time. Principal Player was not associated with either Geopolitical Significance or Demographic Significance.

General Characteristics of Environmental Reporting

The first goal of this study was to outline the characteristics of the environmental reporting in Time and the New York Times. For both publications, it is clear that human population issues are no longer a priority topic, representing less than five percent of the environmental issues covered in each of the four samples. This is a continuation of a downward trend noted by McGeachy (1989) in the early 1980s. It is also apparent that while conflict is the story angle of choice, solutions are well represented. The national and international emphases are consistent with the national nature of these particular publications. Also, environmentalists and their viewpoints receive the least attention. However, no one other player is dominant in either publication, so accusations of bias toward business, for example, are not supported by this study.

While intrinsic concerns (e.g., aesthetic, philosophical or religious views of nature) were the least covered of the three concerns considered here, they represented from at least 12% of the articles in Time to as much as 25% in the New York Times. It can be argued that issues of economic well-being and safety are of a higher priority for the average person, hence the mass media’s attraction to such issues from a marketing standpoint. Nevertheless, while Time and the New York Times in this study consistently covered economic and safety concerns more often, they did not exclude intrinsic concerns.

Changes between 1991 and 1992

The second goal of this study was to determine what changes there might be in the environmental reporting for each publication between an election year and a non-election year. Not only was 1992 a presidential election year, but also the Democratic vice-presidential candidate had a reputation as an environmentalist. In addition to the election, the Earth Summit took place. No significant differences between whole categories were found for the comparisons of 1991 and 1992, but several significant percentage differences were found for individual subcategories. Although these significant percentage differences were not large enough to produce significant differences for any entire category in the 1991 to 1992 comparisons, they are instructive.

The significant increase in the number of articles on the environmental movement issue in 1992 in Time may be linked to the Earth Summit. Time reported on the Earth Summit extensively; 20% of its environmental articles were about the Earth Summit and ancillary issues. Only Energy Resources coverage was significantly reduced. The total number of environmental articles increased by only three in the 1992 sample (a 7% increase over 1991), so Time apparently chose not to devote more of its newshole to articles on environmental issues despite the extra coverage of the Earth Summit.

The increase in the number of articles of International Significance can probably be traced to the Earth Summit as well since the Earth Summit was an international event. The increase in Government articles in 1992 could be attributed to the Earth Summit, primarily a governmental activity. The increase in attention to economic concerns in 1992 may also be linked to the Earth Summit. While many groups, including many with concerns of an intrinsic and a safety nature, had a deep interest in the Earth Summit, Time nominally mentioned these others while focusing on the economic aspects.

On the whole, the Earth Summit appears to have had an appreciable effect on Time magazine’s environmental coverage; most of the significant changes between 1991 and 1992 can be traced to it. Furthermore, that it was an election year may have contributed to this interest in the Earth Summit; that the environment was an important issue was indicated by Al Gore’s nomination for Vice-President.

The New York Times had slightly fewer significant subcategory differences between 1991 and 1992 than did Time, and these were of lesser magnitude in most cases. For environmental issue coverage, the barely significant increase in coverage of the environmental movement may be traced to the Earth Summit as it was in Time, but the change is not as large. New York Times devoted 12% of its environmental articles to the Earth Summit and ancillary issues, compared to 20% for Time.

The increase in articles unspecified in location and decrease in urban articles may indicate an increase in policy and issue-oriented articles as opposed to event reporting. This may be linked to the election year increase in political debate and the Earth Summit, which involved many policy and philosophy issues. No increase in Time was seen in this category, however.

The significant increase in the number of government-focused articles may be an indication of the Earth Summit or the election year; a significant increase was also seen in Time. Both Time and the New York Times significantly decreased the number of business-oriented environmental articles. The decrease in business articles does not appear to be linked to a specific trend identifiable here.

The significant decrease in national significance articles and increase in local significance appear not to be linked to a particular event such as the Earth Summit. The New York Times had no significant differences in story angle or concerns, indicating no apparent change in editorial policies with regard to these characteristics.

These changes in the New York Times are on the whole less profound than those in Time. This may be due to the difference in emphasis on the Earth Summit. The New York Times’ environmental coverage included a smaller percentage of stories devoted to the Earth Summit than Time’s included (12% and 20%, respectively). The larger newshole of the New York Times, in addition to the effects of daily of publication discussed above, may have had a leveling effect on the New York Times’ environmental coverage by reducing the influence which big spreads on single topics (e.g., the Earth Summit) had on the results.

In neither the New York Times nor Time does there appear to be any large scale editorial shift in environmental reporting between 1991 and the election year 1992 since those differences which do exist are small and not in a pattern consistent with such a shift.

Differences between Time and the New York Times

The third goal of this study was to determine the ways in which the two publications differ in their environmental reporting as a whole. It is important to reiterate that only one significant difference was found between the two publications for whole categories; however, there were some significant percentage differences among individual subcategories.

For 1991, the New York Times had significantly less coverage of energy resources and more of the environmental movement than Time, but all other issues were given about the same play. In 1992, the New York Times had significantly fewer environmental movement and natural phenomena articles than Time. This may be linked to the differences between a daily newspaper and a weekly newsmagazine. It is Time’s pattern to take the big story of the week and run several articles summarizing and analyzing the event in depth. The New York Times, however, is a daily publication for which only one issue per week was selected for this study. The author assumes it will distribute its articles on a major event throughout the week as events unfold. While the Sunday edition was selected for this study to minimize this effect (see Method), the inconsistent nature of the differences in issue reporting between the publications (e.g., the environmental movement articles were significantly greater one year, significantly less the next) indicate that the effect may not have been eliminated. Two events in particular correspond to this pattern. Two major environmental events in 1992 were the Earth Summit and Hurricane Andrew. Time devoted extensive coverage to each, while the New York Times Sunday edition did not devote as much coverage proportionately. Since both events took place largely in the middle of the week, the Sunday New York Times would be less likely to include as many articles on these events having already covered them as they unfolded during the week. This would support the result of significantly fewer articles for the New York Times than Time in 1992 in the subcategories of Environmental Movement and Natural Phenomena, which included most Earth Summit and Hurricane Andrew articles, respectively. Likewise, the result that the New York Times and Time were very similar for Geopolitical Significance in 1991, but that Time had significantly more international stories in 1992, may also be a result of Time’s coverage of the international Earth Summit.

The New York Times had more urban and fewer unspecified articles both years, especially in 1991. The New York Times is the local newspaper for New York City, therefore a greater percentage of urban articles is not surprising. Also, this difference may be due to the nature of the New York Times as a daily newspaper which consequently reports more events as they occur, whereas the newsweekly Time can exercise the option to analyze broader issues in which location is less important and devote less space to reporting specific events.

Each year, the New York Times had significantly more articles which reported government actions and viewpoints. Again, this may be linked to the event-reporting nature of the newspaper. Specifically, government actions are newsworthy items which the New York Times as a daily paper has more space to include. Time, on the other hand, may be more selective in which government actions it reports due in part to its limited space.

As for concerns, the New York Times had fewer articles on safety both years, instead focusing on economic concerns in 1991 and on intrinsic concerns in 1992. This difference cannot be directly linked to a difference in editorial selection. Apparently, Time sees a greater need to discuss safety issues in relationship to the environment than does the New York Times.

Further conclusions concerning these comparisons are limited in this study because of the lack of data with which to compare the results found here. Ruben and Sachs (1973) and McGeachy (1989) studied only environmental issue coverage and reported results as an average number of stories per issue, which is not directly comparable to the results described here. The present method was chosen because it yields more quantitative data; reporting the average number of articles per issue largely yielded numbers from zero to five, which are too small for most statistical analyses. Therefore, despite the limited comparability to past studies, the results as given here will be more useful to future quantitative studies.

Associations between Categories

That Geopolitical Significance and Demographic Significance were correlated in both Time and the New York Times was expected mostly because of the characteristics of particular subcategories. A large portion of articles were Unspecified in Demographic Significance, that is, they did not mention a specific locale. Since they did not deal with a specific locale, they often dealt with broader issues such as policy. Furthermore, when policy and other broad issues were discussed, the articles tended to be of National Geopolitical Significance since Time and the New York Times are national publications. Therefore, it was reasoned that Demographic Significance would be associated with Geopolitical Significance

Also, that Principal Player and Concerns were correlated was expected because interest groups are organized around concerns which the members share. It is logical, then, that an article tending to emphasize a certain Principal Player would tend to talk about a Concern important to that Principal Player.

Demographic Significance and Concerns were negatively correlated for Time but were positively correlated for the New York Times. In general, this indicates that Time did not tend to discuss Concerns and Demographic Significance in the associations expected, while the New York Times did. For example, Time articles which were Unspecified in location did not tend to discuss Economic Concerns while the New York Times Unspecified articles did. Since Unspecified articles tended to be about policy and broader issues, it appears that the New York Times tended to discuss economics when reporting on policy issues which involve the environment, while Time tended not to discuss economic concerns when reporting on policy.

A correlation between Geopolitical Significance and Concerns was not necessarily expected. The scope of the issue or event does not necessarily mean that it will tend to be focused on a particular Concern. Issues which involve Economic, Safety, or Intrinsic Concerns may arise on any Geopolitical level from the Local to the International. Nevertheless, the weak correlation for the New York Times indicates that at least for this publication there may a pattern of tending to discuss certain Concerns when dealing with issues of certain Geopolitical Significance. This correlation for all of New York Times is a result of the correlation in 1992; there was no correlation in 1991. This 1992 correlation may be caused in part by a tendency to discuss Economic Concerns when reporting in the Earth Summit, which was of largely International Geopolitical Significance. However, one would have to look specifically at the differences between Earth Summit articles and other articles to confirm this.

That Principal Player and neither Geopolitical Significance nor Demographic Significance were correlated (except for Geopolitical Significance in New York Times 1992) was expected because the Principal Players studied here are involved in issues on all Geopolitical and Demographic Significance levels. One logical exception might be that articles of International Geopolitical Significance may tend to involve Government more than other Principal Players because international relations are often political. The correlation between these two categories in New York Times 1992 may support this exception. This may be a unusual case linked to the articles on the Earth Summit, an event of largely International Geopolitical Significance which may have tended to discuss Government actions. Nevertheless, Time did not reflect such an association in 1992, so this assertion cannot be confirmed here.

In summary, that most category pairs were correlated when expected in most instances indicates that the asserted logical links between the categories are to some degree being preserved by the categorization scheme developed for this study.

If future studies are to be based on the method presented here, certain issues need to be addressed. Foremost, the one categorical significant difference which was found, the category Demographic Significance for the comparison Time 1991 versus the New York Times 1991, is suspect because of the large number of articles coding in the Unspecified subcategory, especially for Time (70%). This adds a measure of uncertainty to the coding. This category also had the lowest reliability result (.60). This category needs to be more precisely defined. In particular, the Unspecified subcategory needs further explanation to produce clearer picture of what those articles are discussing when not addressing an urban or rural issue.

Nevertheless, the Unspecified subcategory is meaningful. It indicates how often a publication includes an article which does not concern a specific location, for example, articles on environmental policy or philosophy. In the case of Time 1991 versus the New York Times 1991, the New York Times chose to print significantly more articles which dealt with a specific locale: Urban. Time, meanwhile, ran more articles unrelated to a specific locale. The New York Times, as a daily publication, bases more of its articles on specific events as they unfold; Time, on the other hand, as a newsweekly, can spend proportionately less space discussing specific events and more space analyzing broader topics such as policy issues. Perhaps a better name for the subcategory might be Generalized rather than Unspecified. This name reflects the assumption used in this study that issues of policy or philosophy not associated with a specific Demographic group apply to all groups, Urban and Rural.

The lack of significant differences in other categories and comparisons was not expected and also may indicate areas for improvement of the method. Of course, there may not be any real differences to be found. The lack of differences between 1991 and 1992 may be because environmental issues are not an appreciable factor in elections. The lack of differences between the two publications may be because both are national news publications; therefore, the desire to select articles which appeal to the broadest audience may cause them to report on similar environmental issues in similar ways.

Nevertheless, the pilot study did indicate grounds for expecting more differences based on the rationale given (see Problem Statement). This expectation was supported by significant percentage differences between subcategories in the pilot study (whole categories were not tested for significant differences). The final study does show many significant percentage differences among subcategories as did the pilot study. However, these were not enough to produce a significant difference for entire categories in the final study except in one instance.

This lack of significant differences may be due to the categorization scheme being too finely divided for the sample size. There were numerous sparse matrices in the statistical analysis. Many subcategories, mainly for the Time samples but also some for the New York Times samples, had five or fewer articles. A larger sample might have eliminated these sparse matrices, but that would not necessarily have produced more significant differences because few of these sparse categories were close to being significantly different as they were. Rather, the problem might be better addressed by collapsing subcategories. For example, in Geopolitical Significance, the subcategory Region had only one sample with more than five articles. This could be collapsed with Local to form one subcategory, eliminating all sparse matrices for the Demographic Significance category; the two subcategories have a logical link as well.

Environmental Issues could similarly be collapsed; however, one subcategory presents a particular problem here. Human Population was included as a subcategory because it is traditionally an environmental issue and was included in previous content analysis studies (e.g., Rubin & Sachs, 1973, and McGeachy, 1989). However, very few articles were written on this topic in the Time and the New York Times for the samples studied. McGeachy noticed a drop from previous levels as well. Given this trend, it is unlikely a larger sample would have had enough articles on this issue to eliminate the sparse matrices. For example, the largest sample in this study, the New York Times 1992 (n=155), had fewer than five articles in this subcategory. In future studies, this topic will have to be included in another subcategory or its articles coded under the next best subcategory to preserve statistical validity.

Collapsing, like increasing sample size, would have been unlikely to produce a large change in the results of the present study because few of the comparisons were close to being significantly different beforehand. Nevertheless, at the very least it would simplify coding and make statistical analysis more reliable.

Another explanation for the lack of significant differences when comparing each publication to itself may be that the sampling periods are too close in time. The 1991 and 1992 samples are only six months apart. Despite the reasons for expecting differences stated above (see Problem Statement), six months may not be enough time for differences to become evident. Coding by article may be too coarse a method for detecting these gradual changes, or there may not be any significant differences to be found in so short a period. Perhaps a longer gap of several years (e.g., Rubin & Sachs, 1973, and McGeachy, 1989) may uncover more differences.

Concerning the lack of differences between Time and the New York Times (except for Demographic Significance), the anticipated effect of different editorial viewpoints on the environmental reporting may also be too subtle for the article-based coding method.

This content analysis accomplished the following: It quantitatively profiled the characteristics of environmental reporting in Time and the New York Times. It found only one major significant difference between the environmental reporting in Time and the New York Times, which is that the New York Times in 1991 included more articles dealing with Urban and Rural locations while Time in 1991 had more articles without a specific location identified. The study found no major significant differences between in the environmental reporting in 1991 and an election year, 1992. However, numerous significant percentage differences among subcategories indicate that there may be differences which may become evident in future studies which use samples more separated in time or employ a method with finer resolution. Finally, the present study provided an example of a method as well as data for use in future quantitative studies of environmental reporting in the mass media.

Belak, E. R., Jr. (1972). The outdoor magazines revisited. Journal of Environmental Education, 4(1), 15-19.

Bowles, D. A., & Bromley, R. V. (1992). Newsmagazine coverage of the Supreme Court during the Reagan administration. Journalism Quarterly, 69, 948-959.

Clausing, J. (1971). The ecological message of the outdoor magazines. Journal of Environmental Education, 2(4), 10-12.

Congressional Information Service (1992). American Statistics Index Annual. Washington: Author.

Glynn, C. J., & Tims, A. R. (1982). Sensationalism in science issues: A case study. Journalism Quarterly, 59, 126-131.

Hoesterey, J., & Bowman, J. S. (1976). The environmental message of Audubon and the Sierra Club Bulletin. Journal of Environmental Education, 7(3), 61-65.

Hoffman, M. S. (Ed.). (1992). The World Almanac Book of Facts 1993. New York: Pharos Books.

Howenstine, E. (1987). Environmental reporting: Shift from 1970 to 1982. Journalism Quarterly, 64, 842-846.

Jensen, D. W. (1989, August). Magazine coverage of issues of nuclear warfare. Paper presented at the Annual Meeting of the Association for Education in Journalism and Mass Communication, Washington, DC. (EDRS Document Service No. ED 308 563)

Larson, M. A., Zimmerman, D., & Scherer, C. (1982). Communication behavior by environmental activists compared to non-active persons. Journal of Environmental Education, 14(1), 11-20.

Lundberg, L. J. (1984). Comprehensiveness of coverage of tropical rain deforestation. Journalism Quarterly, 61, 378-382.

McGeachy, L. (1989). Trends in the magazine coverage of environmental issues, 1961-1986. Journal of Environmental Education, 20(2), 6-13.

Pierce, J. C., Lee-Sammons, L., & Lovrich, N. P., Jr. (1988). U.S. and Japanese source reliance for environmental information. Journalism Quarterly, 65, 902-908.

Rubin, D. M., & Sachs, D. P. (1973). Mass Media and the Environment: Water Resources, Land Use and Atomic Energy in California. New York: Praeger.

Sproull, N. L. (1988). Handbook of research methods: A guide for practitioners and students in the social sciences. Metuchen, N. J.: Scarecrow Press.

Stempel, G. H., III. (1989a). Content analysis. In G. H. Stempel III, & B. H. Westley (Eds.), Research methods in mass communication (pp. 124-136). Englewood Cliffs, N. J.: Prentice Hall.

Stempel, G. H., III. (1989b). Statistical designs for content analysis. In G. H. Stempel III, & B. H. Westley (Eds.), Research methods in mass communication (pp. 137-149). Englewood Cliffs, N. J.: Prentice Hall.

Stocking, H. & Leonard, J. P. (1990). The greening of the press. Columbia Journalism Review, 29, 37-44.

Strodthoff, G. G., Hawkins, R. P., & Schoenfeld, A. C. (1985). Media roles in a social movement: A model of ideology diffusion. Journal of Communication, 35(2), 134-153.

Yagade, A., & Dozier, D. M. (1990). The media agenda-setting effect of concrete versus abstract issues. Journalism Quarterly, 67, 3-10.