Content-Type: text/html

                Media Coverage of Risk Events







Media Coverage of Risk Events:
A Framing Comparison of Two Fatal Manufacturing Accidents


By



Michael J. Palenchar
Ph.D. Student
University of Florida
College of Journalism and Communications
2039A Weimer Hall
PO Box 118400
Gainesville, FL 32611-8400
Tel: 352-379-5551
Fax: 352-392-3919
<a href="/cgi-bin/wa?LOGON=A3%3Dind0109B%26L%3DAEJMC%26E%3D8bit%26P%3D9388819%26B%3D--%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D%253D_279740104%253D%253D_%26T%3Dtext%252Fplain%3B%2520charset%3Diso-8859-1" >[log in to unmask]</a>


Competitive Paper:
Association for Education in Journalism and Mass Communication: Public Relations Division
Washington, D.C., August 2001


 Media Coverage of Risk Events:
A Framing Comparison of Two Fatal Manufacturing Accidents


This study extends analysis of the meaning contained in risk discourse as part of a community infrastructural approach to risk communication studies. The author deconstructed media narratives through a systematic, longitudinal content analysis of a major metropolitan newspaper's coverage (n=92) of two manufacturing crises. There was an increased use of risk bearers and a decreased use in risk generators as sources while the negative focus on risk generators increased, and a significant increase in the use of medical and legal sources. The implication of such analysis is that wise public relations personnel should understand how journalists use sources and narrative elements to frame situations of risk.
 Media Coverage of Risk Events:
A Framing Comparison of Two Fatal Manufacturing Accidents

        Risk communication addresses how individuals, groups or organizations frame their pictures of the world by addressing scientific evaluations of risks, the perceptions lay people have of them, and actions that are warranted in light of the degree of risk and people's tolerance. A growing body of research has examined the risk communication process from some key stakeholder positions (government officials, industry, scientists and ordinary citizens), but additional research is essential to ascertain the message content that journalists use to frame and discuss risk. The media - both at the micro (journalist) and macro (medium) level - affect with varying degrees how individuals develop their pictures of the world (e.g., McCombs, Einsiedel, & Weaver, 1991). This exploration has substantial implications for public relations practitioners who may be called upon to design messages about risks and to communicate on behalf of companies, activist groups, ordinary citizens and governmental
 agencies.
        Technological risk is one of the more hotly contested issues associated with living in a modern society, from privacy issues regarding reproductive processes to new sustainable energy sources.  Recent media reports concerning record-setting global warming, unprecedented forest fires in Brazil and Indonesia, and a host of other fallouts from our technological advancements bombard our attention daily.  Iconic incidents such as Bhopal and Mayak internationally, and Three Mile Island and Hanford nationally, will always remind people of the dangers involved with the production and use of hazardous materials.
Journalists and other communication practitioners who deal with risk communication need to achieve a delicate balance - to help people make sound judgments within a community of interest even though technical information is often difficult to obtain, assess and draw consensual conclusions about. On the other hand, however, those same people must not lose their sense of apprehension about risks to such an extent that they cannot or do not act to reduce their harm.  "Risks to health, safety and the environment abound in the world and people cope as best as they can.  But before action can be taken to control, reduce or eliminate these risks, decisions must be made about which risks are important and which risks can safely be ignored" (Covello & Johnson, 1987, p. vii).
        All aspects - input, process and output - are integral parts of risk communication. This paper, however, focuses on and extends analysis of the meaning contained in discourse as part of a community infrastructural approach to risk communication studies, specifically related to how the print media frames issues strife with health, safety and environmental uncertainty.
Risk Communication
        An essential element of human history is the drama of assessing, communicating about and adapting to risks. The origins of risk analysis have been traced to the Babylonians in 3200 B.C.  They used myths, metaphors and rituals to predict risks and to communicate knowledge about avoiding hazards.  Risk communication was embedded in folk discourse (Krimsky & Plough, 1988).  Explicit interest in risk communication in the United States can be traced back to the 1950s and the "Atoms for Peace" campaign. The anti-nuclear movement in the 1970s put risk communication into the limelight (Kasperson & Stallen, 1991).  Today, those drama messages have been replaced by information-based technologies and expert-centered messages.
        If exposure to risk is not new, then why is there a renaissance in risk assessment and risk communication?  According to Fischhoff (1990), "What is new in their [stakeholders and general public] response to the risk of modern technologies is their insistence on having a role in deciding how those risks will be managed" (p. 84).  Individuals, activist groups and communities insist on being part of the decision process, whether it concerns hazardous materials or new pharmaceutical products.  "A revolution in environmentalism and personal health is requiring that reasonable and responsible communication be employed to change personal and collective behaviors" (Heath & Nathan, 1991, p. 15).
Focus on risk communication within the chemical industry began, at least in the perception of key stakeholders, when private and public organizations failed to understand and exhibit appropriate levels of corporate responsibility, by failing to achieve proper control of risks associated with their activities.  When publics perceive this legitimacy gap, they take actions to impose standards on the offending organization (Sethi, 1977). For example, the chemical catastrophe in Bhopal motivated some citizens to worry whether similar risks loomed near their homes or at their work locations. Questions of that sort led federal legislators to create the Emergency Planning and Community Right-To-Know Act of 1986, section three of The Superfund Amendments and Reauthorization Act of 1986 (SARA Title III).  Legislators created SARA as a communication apparatus and strategic business planning process to empower people regarding risks.
        By codifying environmental risk communication, SARA requires chemical companies to inform citizens regarding the kinds and quantities of chemicals that are manufactured, stored, transported and emitted in each community. Newman (1988), in reviewing the Right-to-Know provision for the Public Relations Society of America, concluded "the theory behind these toxic laws is that this information not only will help answer citizen questions about [chemical] releases, but also will assist them in pressuring government and industry to correct practices that threaten their health and environment" (p. 8).
        Risk comprehension, however, is more than just understanding the actual risks involved, making a decision about such risks and communicating about them.  It deals with both the actual risks and the perception people have of them.  "We do not perceive risks, we perceive various features of decision problems and this, in turn, leads to feelings of risk" (Brehmer, 1987, p. 26).
        Heath and Nathan (1991) reasoned that risk communication is not just a matter of sharing data and interpreting it to draw science-based conclusions about risks. More accurately, risk estimates and policies related to the control of risks are socio-political decisions.  Risk communication has progressed beyond a source-oriented approach to a risk perception and a risk management approach. Typical of the theme that characterized the source-oriented era is Covello's (1992) view of risk communication "as the exchange of information among interested parties about the nature, magnitude, significance or control of a risk" (p. 359). Leiss (1996) called this the technical risk assessment period when, for example, industrial spokespersons were advised to appease or assuage publics' apprehension by being credible and clear.
        A second phase of risk communication featured a more interactive approach. The National Research Council (1989) described the process as the two-way exchange of information: "We see risk communication as the interactive process of exchange of information and opinion among individuals, groups and institutions" (p. 2). This definition suggests a movement from an asymmetrical approach of communication and decision-making toward what J. Grunig (1989) defined as two-way symmetrical communication.
        Leiss (1996) identified a third, current phase of risk communication that features social relations. Typical of this era, Otway (1992) discussed the complexity of risk communication: "Risk communication requirements are a political response to popular demands....  The main product of risk communication is not information, but the quality of the social relationship it supports" (p. 227). This approach based on a shared, social relations-community infrastructural approach works to achieve a level of discourse that can treat both the content issues of the risk - technical assessment - and the quality of the relationships.
        Palmlund (1992) argued against an authority-centered, linear communication approach. "I believe that a different vision of life is needed, one that emphasizes the role of social interaction, emotions and power in public life_. Societal evaluation of risk must be seen as a contest, where the participants offer competing views of reality.  They compete to define what should be viewed as the benefits and the risks of prevailing production practices" (p. 199).  Thus, risk controversy can be considered a drama enacted by many personae, each attempting to create and live a different view of reality.
        The nineteenth-century German philosopher Georg W. F. Hegel identified a process called dialectic, or the principle of contradiction. According to Hegel (Solomon, 1983), development is marked by the clashing and merging of opposites. In journalism, dialectic concerns such as objective versus subjective reporting and hard versus soft news influence the communication process (Merrill, 1989).  Dialectic controversies related to technology (ethical versus unethical, costs versus benefits) are a central topic of risk communication with far-impacting effects. According to Nelkin (1989), "The social movements organized to challenge science and technology are driven by a moral rhetoric of good and evil, of right and wrong.  They are attracting constituents who fear the misuse of science by major social institutions, who see the need to reassess the social values, priorities and political relationships underlying scientific and technological processes, and who see themselves as preserving
 the moral values lost in the course of technological change. Thus controversies matter and must be taken seriously as an indication of public attitudes toward science" (p. 456).
        Understanding such controversies has been one of the principles in the risk communication literature. It seems to be the central premise of the analysis and prescriptive advice offered by Covello, von Winterfeldt, and Slovic (1987).  Their analysis focused on: (1) experts as sources, (2) messages as information about risks, (3) channels as media reporters and editors, and (4) receivers as individual citizens.  They worried that messages suffer from a lack of scientific understanding, data, models, methods and technical analyses that are often unintelligible to laypersons.
        Covello et. al. (1987) noted that source problems include: lack of trust and credibility, disagreements among experts, limited authority and resources for addressing risk problems, lack of data addressing the specific fears and concerns of individuals and communities and failure to disclose research limitations. They also focused their research on the "limited understanding of the interests, concerns, fears, values, priorities, and preferences of individual citizens and public groups; and use of bureaucratic, legalistic and technical language" (pp. 110).  Channels are affected by "selective and biased media reporting that emphasizes drama, wrongdoing, disagreements, and conflict; premature disclosures of scientific information; and oversimplification and distortions of, as well as inaccuracies in, interpreting technical risk information" (p. 111).
        To effectively gain insight into how journalists frame media issues, risk and uncertainty related to risk should not be considered absolute concepts, but vary by individual perception in context. Palmlund (1992) contended that risk studies could lead to incorrect conclusions if researchers feel that risks can be explained exclusively by applying "quantitative estimates" that are "priced as commodities" (p. 189). He believed that a strategy is needed "that emphasizes the role of social interaction, emotions, and power in public life" (p. 199).  Part of that strategy could be using framing analysis to improve understanding of media coverage of events within communities of risk.
Framing
        Implicitly, framing plays a crucial role in public relations (Hallahan, 1999), and by extension risk communication. If the product of risk communication is not information but the quality of the social relationship it supports (Otway, 1992), understanding how the media frames risk is an important message element of the social relationship.
        Framing theories are often presented as explanations for the influence of media coverage on issue developments related to risk communication. The media are attracted to uncertainty and the risk (conflicts) about an issue.  Framing concepts related to risk communication, issues management and crisis management suggest that what is stated or omitted can define issues such as risk events for audiences (Bridges and Nelson, 1999). "Framing theorists suggest that the way an issue is presented - the frame - especially through the media, can affect public perceptions of the issue" (p. 100).  According to Nelkin (1989), the media establish a framework of expectations regarding risks associated with various technologies (e.g., manufacturing of chemicals) by making risks visible, providing a context for evaluation, and shaping expectations about similar or related risks.
        The concept of framing provides an avenue to describe the power of text.  "Analysis of frames illuminates the precise way in which influence over a human consciousness is exerted by the transfer (or communication) of information from one location - such as speech, utterance, new report, or novel - to that consciousness" (Entman, 1993, p. 52).
        Framing involves both selection and salience.  According to Entman (1993), "to frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation" (p. 52).  He also suggested that media frames could define problems, diagnose causes, make ethical judgments and suggest solutions. Some of the earliest work in understanding the concept of framing related to how receivers process information is Goffman's (1974) description of frames as a "schemata of interpretation" through which people organize their thoughts and the discourse related to their thoughts, relayed in experience and information (p. 21).
        Frames can also be used to describe how the communicator builds a message.  Entman (1993) described how text contains frames that are manifested in the presence or absence of key linguistic, syntactic, lexical and semantic elements that "provide thematically reinforcing clusters of facts or judgments" (p. 52).  The use of sources can have a significant effect on how the media frames an issue. This can depend of many variables, including the journalist's trust in source, familiarity with source and the ability of the source to gain the attention of the reporter, all affecting the amount and type of coverage (Einsiedel & Thorne, 1999).
Gitlin (1980) described how frames allow journalists to develop their stories within the infrastructure and requirements of journalism. "They sometimes generate, sometimes amplify a field of legitimate discourse that shapes the public's 'definitions of its situations,' and they [journalists] work through selections and omissions, through emphases and tones, through all their forms of treatment" (p. 9). Media frames involve the context, content, topic, coverage and package of news events.
One of Gans (1979, as noted by Gitlin, 1980) theories of news typologies included event-centered news. In this same vein, Shoemaker, Chang and Brendlinger (1987) noted that media coverage of events relates to the actions and behaviors of the community and world on that particular day, while being more of an index of the unusual aspects of individuals or organizations.
Among studies on media framing, focus on health, safety and environmental risks related to manufacturing hazardous materials have not been well developed, though media coverage of risk is part of the journalistic culture to inform readers about high-consequence events. The Chernobyl disaster of 1986 and the discovery of the ozone hole in 1985 gained the media's attention.  The practice of framing discordant news events is a dominant fact of contemporary journalism (McCombs, Einsiedel, & Weaver, 1991). Part of this comes from the common use of the story or narrative in the telling of an event. "Central to the genre of journalism is the story" (p. 35).  Work by McCartney (1987) has demonstrated that basic plots and themes used throughout literature are present in news stories.
        Frames have at least four locations in the communication process, including communicators, text, receivers and culture (Entman, 1993). This paper focuses on frames within the textual element of the risk communication process. "The text contains frames, which are manifested by the presence or absence of certain key words, stock phrases, stereotypical images, source of information, and sentences that provide thematically reinforcing clusters of facts or judgments" (p. 52).
        The language selected by journalists in describing events influences perceptions of those events. As Edelman (1988) argued, "It is language about political events, not the events in any other sense, that people experience; even developments that are close by take their meaning from the language that depicts them" (p. 104).
        Within the textual element of the risk communication process, framing can take place at multiple structural levels. As a form of language use, media texts can display linguistic or grammatical structures, such as phonology; syntactic structures may underlie ideological positions, such as using passive voice or deleting parts of news discourse; and lexical choices, such as using "Vietnam War" to describe a plant-explosion scene (van Dijk, 1988).
        Also important for public relations practitioners is interpreting media discourse at the semantic level.  "Semantically, sentence meanings (propositions) are mutually dependent and connected and form coherent sequences. Besides the meaning of words and sentences, world knowledge in the form of models, frames and scripts represented in memory, is brought to bear by the reader to understand a piece of news discourse as a coherent whole" (van Dijk, 1988, p. 177).
Van Dijk (1988) reasoned that framing elements also include rhetorical dimensions, such as structures or organizational principles. A rhetorical perspective for the study and practice of public relations entails the analysis of words and other symbols. "Meaning defines the identities and prerogatives of organizations, people associated with them, and their relationships" (Heath, 1993, p. 142). Specifically, words have propositional value (Burke, 1966), and the selection of those terms affects how information is considered, accepted, acted upon or altered. These propositions, according to Heath (2001), compete "in ways that help to inform judgments and actions, clarify and order the evaluative (value) dimensions of thought and choice, and justify or deny the expedient wisdom of competing policies" (p. 32).
Narrative analysis helps illuminate the meaning of discourse, including its cultural foundations. People think, communicate and behave in accord with their narratives (Fisher, 1987). Drama is dialogic and disputational - protagonist against antagonist. When telling a story, narrators feature themes of good and evil. Plots are means by which actors live out or enact themes of good and evil. People's sense of self and security is enacted in drama, narrative terms. In a similar way, each news story is the enactment of stories (plots and themes) by characters in scenes. To better understand a community or public, one can deconstruct its narratives, including media narratives of risk events.
        By deconstructing media narratives, framing ultimately can help public relations practitioners understand how key stakeholders navigate through the information environment by better understanding what sources of information and how narrative elements frame the risk. As Einsiedel and Thorne (1999) noted, a significant source of information for the public is mass media, people can be influenced by its content, "but simply that, over time, the media can be an important - if not the only - source of information for various publics on many issues" (p. 52).
Hallahan (1999) suggested that one of framing's inherent strengths might lie in the concept's "emphasis on providing context within which information presented and processed allows framing to be applied across a broad spectrum of communication situations" (p. 209).  This paper helps to extend the concept of media framing to the risk communication context. Examining the media frames regarding risk events can also supply important descriptions of the media's role in risk discourse related to one of Hallahan's (1999) seven models of framing - the framing of news.  Combining the perspectives of risk communication and framing, this research project analyzed the selection and salience of issues and character roles within two risk events by one community's major daily newspaper. The paper looked at the following research questions:
RQ1: How did the media frame the issues involved in the risk narrative?
RQ2: How did the media frame the characters involved in the risk narrative?
RQ3: Did the media frame each risk event differently?
Background
The Texas Gulf Coast, and Houston Ship Channel in particular, is an excellent location to analyze media coverage of the risks associated with the production, storage and transportation of chemicals. This region, often referred to as the "Chemical Coast," has the largest concentration of petrochemical plants in the world.  Chemical facilities possess some characteristics of risks.  Lindell and Earle (1983) found that high-risk chemical facilities were distinguished by elevated levels of perceived threat to workers, the public and future generations. Major governmental and non-profit associations advocate for, monitor and/or regulate the chemical industry within the Houston Ship Channel, such as the Environmental Protection Agency (EPA), the Chemical Manufacturers Association (CMA), the Occupational Safety and Health Administration (OSHA), the Texas Natural Resource Conservation Commission (TNRCC) and the Galveston-Houston Association for Smog Prevention (GHASP).
Of particular interest to this study is the Phillips Petroleum chemical complex located in Pasadena, Texas (outside the city of Houston). Two major incidents during the past 11 years provide an opportunity to analyze how the media -- the Houston Chronicle in particular -- frame the risks associated with the manufacturing of petrochemicals.  Phillips is an integrated petroleum company engaged in oil and gas exploration and production worldwide; refining, marketing and transportation operations primarily in the United States; chemicals and plastics manufacturing and sales around the globe; and technology development.
        On October 23, 1989, an explosion at the Phillips petrochemical plant killed 23 workers, injured more than 130 personnel injured and caused physical damage in excess of $1 billion.  The explosion, where approximately 1,150 workers operate on a 640-acre complex, was felt 25 miles away, causing fire, heavy black smoke and showered debris for miles around (Warren, 1989).
Numerous ambulances and medical helicopters responded to the plant located on the Houston ship channel. Six miles of the channel as well as a section of nearby State Highway 225 were closed immediately after the explosion. The U.S. Coast Guard and some Houston-based fireboats evacuated more than 100 people from the facility and took them across the channel.  No general evacuations were ordered but two nearby schools were evacuated (Warren, 1989).
        Regarding environmental pollution related to the explosion, the Texas Air Control Board stated that they did not detect any initial indications of hazards to the public. The air board found routine background levels of pollutants in trace quantities, except for higher-than-normal readings of compounds known as paraffin hydrocarbons, which are expected in a release by the fiery decomposition of polyethylene. Critics charged, however, that monitoring began more than 12 hours after the initial explosion and failed to include compounds such as dioxins and furans, suspected carcinogens that can be formed when chlorinated chemicals burn or other chemicals burn in the presence of chlorine (Dawson, 1989).
The blast occurred while workers were cleaning hardened plastic out of a polyethylene reactor loop, where chemicals are churned to make plastic.  "This accident occurred because of management failures by Phillips and Fish," said McMillian, then deputy assistant secretary at OSHA (Pearson & Mintz, 1990, p. A1). There was no backup to the key valve, which was improperly opened and caused the gas leak and blast, despite Phillip's corporate practice that called for the backup.
At the conclusion of the investigation (April 19, 1990), OSHA issued 566 willful and nine serious violations with a combined total proposed penalty of $5.7 million. Willful violations are those committed with an intentional disregard of, or plain indifference to, the requirements of the Occupational Safety and Health Act and OSHA regulations. A serious violation is defined as one in which there is a substantial probability that death or serious physical harm could result, and the employer knew or should have known of the hazard. Both sides eventually settled for a $4 million fine.  The proposed fine and settlement were the largest in OSHA's history (U. S. Department of Labor, 1991).
OSHA also issued 181 willful violations and 12 serious violations with a combined total proposed penalty of $729,600 to Fish Engineering and Construction, Inc., a maintenance contractor on the site.  The two companies were cited for willful violations under general requirements to provide a safe workplace.
The second explosion occurred March 27, 2000, which left one employee dead and 69 workers injured at the plant.  The employee killed had worked for Phillips for 19 years and survived the 1989 explosion. Injuries included burns, cuts from flying fragments of debris and wounds from falls. The explosion spurred area school officials to seal their buildings and keep children safe inside as a precaution against the possibility of toxic fumes.  Pasadena school buses were used to help evacuate employees (Rendon, Bryant, Hopper & Antosh, 2000, p. A1).
A chemical reaction at the bottom of a storage tank involving the chemical butadiene gave off enough heat to overpressure the tank and cause the explosion.  A chemical reaction resulted in a fire and damage to other nearby chemical tanks.  The butadiene tank was out of service for cleaning and had no pressure or temperature gauges that could have alerted workers in the control room to the impending hazard. Also, while the vessel was not in use, butadiene continued to flow into the tank through a non-functioning valve that had not been properly locked out (U.S. Department of Labor, 2000).  State and county officials didn't detect elevated levels of toxic gases in areas downwind of the fire, but noted that chemicals burned in the fire probably contained some of the same toxic components as diesel exhaust (Dawson, 2000).
OSHA proposed fining Phillips $2.5 million, resulting from allegations of 50 violations of safety standards. The primary violations were the failure to train workers properly in the hazards associated with the reaction of the butadiene chemical. "Unfortunately, this tragedy is not an isolated incident, but one in a series of incidents at this site," said U.S. Labor Secretary Herman. "Three workers lost their lives in explosions at this plant in less than a year's time, and 23 others were killed in a major explosion in 1989," (U.S. Department of Labor:00-277, 2000, p. 1).
Methodology
This paper examined media framing of these incidents - to describe what frames are used and what frames dominate news coverage.  Lindenmann (1983) suggested that public relations practitioners could use content analysis based on research questions related to any number of attributes of coverage, such as the presence or absence of particular themes.  Media framing research, however, utilizes but differs from content analysis because it moves beyond counting textual elements as equally salient and influential, and focuses on the most salient clusters of messages - the frames - and its relationship and influence on audience's schemata (Entman, 1993).
Second, this article provides some insight into how media frames change over time. Two major chemical explosions during the past 11 years at the same chemical complex provided an opportunity to analyze how the media, and the Houston Chronicle in particular, framed the risks associated with the production of hazardous materials.
        This research analyzed the content of all the news, feature, opinion and editorial articles in the Houston Chronicle related to two separate explosions at the Phillips Petroleum chemical complex in Pasadena, Texas. The Houston Chronicle was selected because it is now the only major daily newspaper in the area, and the only major daily in the area that was in operation during both explosions. The dominant local paper, albeit only one as opposed to including other major dailies or national papers, was selected due to the researcher's theoretical focus on a community, infrastructural approach to meaning in risk discourse. The time period analyzed was from October 23, 1989 (the day of the first explosion) until November 1, 2000.  The second explosion occurred March 27, 2000, and still received considerable coverage through October 2000.
        This study utilized the archival computer database of the Houston Chronicle to search for stories related to both explosions, including items related to the health, safety, environment, policy and resources affected by the explosions. Key terms used to search the database included "Phillips," "plant," "Pasadena," and "explosion."  All possible combinations of the words were searched. The search rendered 165 articles related to the first explosion, and 47 articles related to the second explosion. Repetitive or similar articles (e.g., for different sections/editions of the newspaper) and articles under 50 words were eliminated for a more accurate representation of the explosion. Full texts of the articles were analyzed for the study. Sixty-nine (41.8%) articles covering the first explosion and 23 (48.9%) articles covering the second explosion were coded and analyzed. No sampling was required (n=92).
The coding unit was the individual article about the explosions. Each article was coded for five variables, with each variable including numerous categories. The five variables are: news sources, risk narrative analysis features, associated risks, evaluation of the risks and contextual elements.
News source related to the source of information and consisted of 16 categories: (1) federal government, such as the Occupational Health and Safety Administration; (2) state and local government, such as the Texas Natural Resource Conservation Commission or the local mayor's office; (3) Phillips' management and spokespersons; (4) Phillips' contracted companies - management and spokespersons; (5) Phillips' employees, including contract; (6) Phillips' former employees; (7) Employees' families and friends (Phillips/contract); (8) industry, such as near-neighbor chemical facilities or related associations such as the Chemical Manufacturers Association; (9) unions, such as the Paper, Allied-Industrial, Chemical and Energy International Union; (10) medical, such as doctors, emergency technicians or hospital spokespersons; (11) attorneys - Phillips/contract companies; (12) attorneys - employees/families (Phillips and contract); (13) judges; (14) activist organizations, such Texas Citizen Action; (15) community leaders, such as pastors or principles; and (16) community residents.
Limited to concern over characters in a sociopolitical drama, Palmlund's (1992) risk narrative analysis identified six risk character features related to risk discourse: (1) Risk bearers perceive themselves as or are actual victims of the negative consequences of the risk. (2) Risk advocates are the heroic protagonists who speak on behalf of the risk bearers. (3) Risk generators create the risks, or are thought to do so, and therefore are likely featured as antagonists in the risk narrative. (4) Risk researchers apply science to determine whether risks occur and if so how they can be abated or mitigated. (5) Risk arbiters are the characters in the risk narratives who take actions to save the risk bearers from present or future risks. (6) Risk informers are persons, such as reporters, news directors, critics, protestors and book authors - persons who stand outside of the conflict as such and make statements about it, who comment on the actions, risks, policies, and outcomes.
Associated risks related to the negative consequence of the risk (explosion) and consists of ten categories: (1) personal health, such as long-term physical and mental health consequences; (2) personal safety, such as injuries or deaths; (3) environmental effects, such as air, water or soil contamination; (4) physical facility effects, such as property destruction; (5) physical community effects, such as home, business, and community disruptions; (6) economic community effects; (7) legal effects, such as lawsuits; (8) public policy effects, such as regulatory penalties and other public policy consequences; (9) ethical/moral effects, such as corporate social responsibilities; and (10) industry effects, such as market fluctuation and disruption of raw materials for other facilities.
        Evaluation of the risks related to the degree of seriousness of the risk. The five categories are: major crisis, moderate crisis, minor crises and neutral coverage.
        Contextual elements related to the content elements of the articles. These include: type of article (news, feature, opinion, other), headline, lead and main topic.
To determine reliability, 10 articles were randomly chosen and coded by two graduate students, including the author. The author is familiar with both the community and industry being examined, working for more than seven years in the area of risk communication along the Houston Ship Channel. Both graduate students have prior experience conducting qualitative research, including coding, for local government and private industry. The coding was conducted jointly - discrepancies were addressed and compromised. Following the development of the coding scheme, the coders independently coded a 10% random sample of the articles to determine inter-coder reliability (89.3%).
Results
RQ1: How did the media frame the issues involved in the risk narrative?
News sources can play a critical role on the impact of information processing (Entman, 1991).  Fishman (1980) suggested that news sources play a large role in type of coverage and focus (definition) of risk.  Houston Chronicle reporters utilized a wide variety of sources (n=16) to cover both explosions, but their choice of sources and frequency differed. During the 1989 explosion the media relied on the following grouped news sources (see table 1): Phillips/contracted companies (24.9%), employees and their families (22%), government (21.8%), industry (14%), community (13.2%) and medical (3.3%). During the 2000 explosion, the media used the following news sources: employees and their families (31.6%), legal (20.8%), Phillips/contracted companies (17.5%), government (12.6%), medical (8.5%), industry (6.3%) and community (2.6%).
-------------------------
Insert table 1 here
-------------------------
        A detailed look at news sources (see table 2) revealed that during the1989 explosion, the Houston Chronicle relied on Phillips (22.5%), employees and their families (15.7%), federal government (11.2%), state and local government (10.6%), and industry (7.8%) as their sources. For the 2000 explosion, the five most frequently used news sources were employees and their families (21.9%), Phillips (17.1%), attorneys for employees and their families (10.8%), federal government (8.9%) and medical (8.5%).
-------------------------
Insert table 2 here
-------------------------
        The Houston Chronicle's coverage (analyzed by article) of the 1989 explosion focused on five major associated risks: personal safety (46.4%), physical facility effects (11.6%), public policy effects (11.6%), physical community effects (8.7%) and personal health (7.2%).  Headlines focused on personal safety (46.4%), physical facility effects (13%), public policy effects (13%), physical community effects (7.2%) and personal health (5.8%).  Story leads focused on personal safety (46.4%), personal health (13%), physical facility effects (13%) and public policy effects (10.1%) (see table 3).
-------------------------
Insert table 3 here
-------------------------
Concerning evaluations of risks, an overwhelming majority of the 1989 articles depicted the event as a major crisis (84.1%), followed by a moderate crisis (8.7%), a minor crisis (4.3%) and a neutral event (2.9%).  Headlines and leads followed a similar pattern, with an overwhelming majority of the headlines (81.2%) and leads (85.5%) describing the event as a major crisis (see table 4). Terms such as "inferno," "critical condition," "fatal explosion" and "something out of a Vietnam movie" were used to describe the severity of the event in the articles.  Headlines such as "Explosion in Pasadena/Outside Phillips: Shock Wave Stuns Homes, Schools" (Kreps, Campbell, Warren, 1989, p. A15) were used to describe the explosion, while leads, for example, referred to "a massive fire, showering debris for miles around and inuring up to 100_" (Warren, 1989, p. A1).
-------------------------
Insert table 4 here
-------------------------
        During the 2000 explosion, the media focused on four major areas of associated risk: personal safety (34.8%), legal effects (30.4%), physical facility effects (13%) and public policy effects (8.7%). Headlines focused on personal safety (43.5%), legal effects (30.4%) and physical facility effects (13%).  Leads focused on personal safety (39.1%), legal effects (26.1%), physical facility effects (13%) and personal health (8.7%) (see table 3).
Concerning evaluations of risks, an overwhelming majority of the 2000 articles (91.3%) depicted the event as a major crisis, with the remainder describing it as a moderate crisis (8.7%).  Headlines and leads followed a slightly different pattern, with an overwhelming majority of the headlines also describing it as a major crisis (78.3%), followed by neutral (13%) and moderate (8.7%).  Leads also focused on a major crisis (82.6%), followed by both neutral and moderate (8.7%) (see table 4).  Terms such as "huge flames," "toxic fumes" and "damn dangerous plant" were used to describe the severity of the event.  Headlines such as "Phillips Plant Blast Kills 1, Injures 52/Facility has an Infamous Safety Record" (Hanson & Kennett, 2000, p. A1) were used to describe the explosion, while leads, for example, referred to "an explosion and fire ripped through a section of Phillips_" (Rendon et al., 2000, p. A1).
RQ2: How did the media frame the characters involved in the risk narrative?
        Using Palmlund's (1992) typologies of risk characters, the media emphasized six different risk character features. Though not absolute categories (some characters may switch roles during a crisis), risk bearers typically included Phillips/contract employees and their families, community residents, and former Phillips employees. Risk generators included Phillips and their contracted company. Risk arbiters included government, medical community and attorneys for the families. Risk bearers' advocates included unions and activist organizations. Risk informers included industry and near-neighbor facilities, while risk researchers come from a variety of sources, including community leaders and local professors.
For the 1989 explosion, risk bearers were featured the most (49.3%), followed by risk generators (24.6%), risk arbiters (18.8%) and risk researchers (4.3%). For the 2000 explosion, risk bearers were also featured the most (39.1%), followed by risk arbiters (30.4%), risk generators (26.1%) and risk researchers (4.3%) (see table 5).
Risk character features in headlines and leads were also examined. Headlines regarding 1989 explosion focused on risk bearers (50.7%), risk generators (23.2%), risk arbiters (16.0%) and risk researchers (7.2%), while leads focused on risk bearers (49.3%), risk generators (23.2%), risk arbiters (17.4%) and risk researchers (7.2%).  For the 2000 explosion, headlines focused on risk bearers (34.8%), and risk generators and risk arbiters equally (30.4%). Risk bearers were also highlighted the most in leads (39.1%), followed by risk arbiters (30.4%), risk generators (21.7%) and risk researchers (8.7%) (see table 5).
-------------------------
Insert table 5 here
-------------------------
News sources by risk character features provided another measuring tool to examine how the media frames key stakeholders in risk events. Reporters used all six risk character features as sources during both the 1989 and 2000 explosion: 1989 - risk bearer (28.6%), risk arbiters (25.6%), risk generators (24.9%), risk advocates (13.5%), risk researchers (4.5%) and risk informers (2.8%); and 2000 - risk arbiters (33.8%), risk bearer (31.6%), risk generators (17.5%), risk researchers (9.7%), risk advocates (5.9%) and risk informers (1.5%) (see table 6).
-------------------------
Insert table 6 here
-------------------------
RQ3: Did the media frame differently each risk event?
        Comparisons of the two explosions demonstrated some differences in the selection and salience of sources, issues and characters. Concerning grouped news sources, between 1989 and 2000 the use of legal sources (+1980%) and medical sources (+158%) increased dramatically. Other changes include an increase in the use of employees and their families as sources (44%). At the same time, there was a decrease in the use of community sources (-80%), industry (-55%), government (-42%) and Phillips/contract companies (-30%) (see table 1).
        Regarding specific news sources (see table 2), there was a noticeable increase in the use of legal sources representing employees and their families (+980%), legal sources representing Phillips and their contract companies (+720%), medical (+158%), employees (+39%) and employees' families (+32%). There was a marked decrease in the use of community residents (-100%), industry (-81%), state and local government (-65%), community leaders (-48%), unions (-23%), Phillips
(-21%) and federal government (-20%).
Concerning news sources identified by risk character features (see table 6), there was an increase in the use of risk researchers (+116%), risk arbiters (+32%) and risk bearers (+10%), while a decrease in the use of risk advocates (-56%), risk informers (-46%) and risk generators (-30%). Comparison of articles, headlines and leads by risk character features (see table 4) yielded some variations. Articles focused on risk arbiters increased 62 percent, while articles focused on risk bearers decreased 21 percent. Headlines focused more on risk generators (+31%) while decreasing focus on risk bearers (-31%). Leads reduced focus on risk bearers (-21%) and risk generators (-6%).
        Comparison of articles, headlines and leads by associated risks (see table 3) showed several differences. Articles focused on personal health (-40%) and personal safety (-25%) decreased, while headlines focused on personal safety (-6%) and leads focused on personal safety (-16%) also decreased.  Regarding evaluations of risks, more articles described the 2000 explosion as a major crisis (+9%) as opposed to the 1989 explosion. Comparing headlines, there was a four percent reduction in describing the 2000 explosion as a major crisis, while leads showed a similar (-3%) reduction (see table 3).
Discussion
        The Houston Chronicle uniquely framed each risk event, with some considerable differences in the selection and salience of issues and characters. Regarding news sources, those relied on by the media during these risk events can either be good news or bad news for practitioners, depending on perspective. For practitioners working for organizations that create risks, for example, sources from Phillips, contracted companies, industry and their attorneys were utilized slightly more (one in three) during the 1989 explosion than in the 2000 explosion (three out of ten). At the same time, community sources (e.g., activists or community leaders) were used 13 percent of the time during 1989 and less than three percent of the time during 2000. Employees and their families (risk bearers) also appear to be an important choice as news sources, utilized 22 percent during 1989 and 32 percent during 2000.
         The Houston Chronicle clearly framed the issue as a major, personal safety crisis. During 1989, personal safety was the focus in articles, headlines and leads (46.4% for all three categories). Similar results for the 2000 explosion, with articles (34.8), headlines (43.5%) and leads (39.1%) focused on personal safety. Only legal effects related to explosion rated nearly as important during the 2000 explosion in articles (30.4%), headlines (30.4%) and leads (26.1%). The focus of risk events on legal concerns might have considerable affect on key risk process variables such as uncertainty, control, support/opposition, and relationship dynamics within risk discourse. Prior research has demonstrated that different types of spokespersons with varying levels of credibility, such as plant managers, local safety officers or third-party chemical experts, affect how lay person's perceive risks (e.g., Heath & Palenchar, 2000)
        The language of the articles, headlines and leads all demonstrated a major crisis in more than 80 percent of their respective categories. It is interesting to note that 13 percent of the headlines and almost nine percent of the story leads covering the 2000 explosion indicated a neutral risk event, once again demonstrating Heath (1993) and Otway's (1992), among others, argument that risk communication is about perception, values and relationships more than about information and science as a product of discourse.
        Overall, the 1989 explosion primarily used Phillips/contracted companies (24.9%), employees and their families (22%), and government (21.8%) for sources, which is consistent with the extended analysis that risk bearers (28.6%), risk arbiters (25.6%) and risk generators (24.9%) were the primary news sources. Articles, headlines and leads clearly focused on risk bearers facing a major crisis related to personal safety.
        Overall, the 2000 explosion primarily used employees and the families (31.6%) and legal sources (20.8%) for sources, which is consistent with the extended analysis that risk arbiters (33.8%) and risk bearers (31.6%) were the primary news sources. Articles, headlines and leads focused on risk bearers facing a major crisis related to personal safety. This time however, headlines tended to focus nearly as much on risk generators (30.4%) as risk bearers (34.8%), and associated risks related to legal issues became an important topic covered by the media in articles (30.4%), headlines (30.4%) and leads (26.1%).
        The data provided interesting insights related to non-personal safety issues. Despite the presence of risks related to chemical explosions, neither personal health effects (1989-7.2%; 2000-4.3%) nor environmental effects (1989-2.9%; 2000-0%) were major issues covered in either event.  Other issues that received minor coverage included physical community effects (2000-4.3%), economic community effects (1989, 2000-4.3%), ethical moral effects (1989, 2000 <2%), and industry effects (1989, 2000 <2%). It appears that the Houston Chronicle's coverage of the event focused on the crisis at hand, related to personal safety, while providing little coverage of serious environmental and long-term personal health issues, and other secondary effects.
        The second research question asked how the media framed risk characters involved in these events.  During both the 1989 and the 2000 explosions, risk bearers received almost half of the attention in articles (1989-49.3%; 2000-39.1%), headlines (1989-50.7%; 2000-34.8%) and leads (1989-50.7%; 2000-34.8%). Risk generators also were highly featured, ranging from 30 percent of 2000 headlines to 22 percent of 2000 leads. As one might expect, the central characters in the risk event, bearers and generators, received the most attention.
        It is important to note, however, the lack of attention focused on risk advocates, risk researchers and risk informers following both explosions.  Risk advocates, such as activist groups, were emphasized in less than two percent of articles, leads and headlines during both explosions. This, despite the large presence of sophisticated local, state and national activist groups and other concerned organizations in the community. Risk informers were used as sources less than three percent of the time during both explosions, while less than two percent of the articles, headlines and leads during the 1989 explosion focused on this character, while during the 2000 explosion, they were not a focus of any article.  Concerning effective risk communication strategies, this contradicts recent research by Heath and Palenchar (2000) that demonstrated that risk informers are preferred sources by residents in a community of risk.
Concerning sources, there was a marked trend in the use of legal (+1980%), medical (+158%) and employees and their families (+44%) as sources of information, while community
(-80%), industry (-55%) and Phillips/contract companies (-30%) decreased.  Combine this data with marked differences among the increased use of risk researchers (+116%), risk arbiters (+32%) and risk bearers (+10%) as sources, and the decrease in sources from risk advocates (-56%), risk informers (-46%) and risk generators (-30%), and it paints a convoluted picture.  Medical and legal sources increased, even though there were dramatically fewer deaths, injuries, penalties and allegations of wrongdoing during the later explosion. Possible explanations could include a possible increased importance of the medical and legal community in risk cultures (demonstrated by an increased use of them as sources by the media). Another explanation could be that the 2000 explosion was the fourth incident in the past 12 months at that plant. A K-Resin section of the plant exploded and killed two workers and injured four others in June 1999. In April 1999, a rail car containing polypropylene blew u
p, and in August 1999 there was an explosion in the polypropylene section of the plant. This analysis is consistent with Coombs (2000) recent work demonstrating that relationship history (in this case a facility that has a history of problems) in large part affects perception of organizational reputation and crisis responsibility to a lesser degree.
Concern regarding the increased use of legal sources during a crisis is also important.  During the 1989 explosion, legal sources were used less than two percent, while during the 2000 explosion attorneys were used as sources 19 percent of the time. Specifically, the Houston Chronicle's use of attorneys on both sides of the debate was similar (legal sources representing risk generators, 1989-<1%, 2000-8.2%; legal sources representing employees and their families, 1989-<1%, 2000-10.8%).  These findings provide additional insight regarding whether public relations is holding its own against the legal field encroachment, specifically as spokespersons, during a crisis. It also supports the viewpoint that public relations practitioners should and must learn to work with legal departments during a crisis (Fitzpatrick, 1993).
Clearly, however, there is an increased use of risk bearers and a decreased use in risk generators as sources. At the same time, the focus on risk bearers in articles (-20.7%), headlines
(-31.4%) and leads (-20.7%) decreased, while focus on risk generators in articles (+6.1%) and headlines (+31%) increased. This information may be disconcerting to practitioners who represent organizations such as chemical companies. Not only is the focus on them increasing but also the reliance on them as sources of information is decreasing. Overall, the framing analysis represented a murky picture and several major inconsistencies, not unlike the uncertainty, perceptions and values that surround risk communication.
Conclusion
        Tuchman (1978) argued more than 20 years ago that the mass media actively set the frames that media users employ to analyze, interpret and discuss public events (e.g., chemical explosions). As such, public relations practitioners need to focus on understanding how the media frames issues and their affect on an organization's key stakeholders.
        Hallahan (1999) rightfully argued that public relations practitioners fundamentally operate as "framing strategists, who strive to determine how situations, attributes, choices, actions, issues and responsibility should be posed to achieve favorable outcomes for clients" (p. 224). To operate as framing strategists, it is also crucial to understand how the key media outlets frame issues or events related to strategic goals and objectives of the organization. Though, as this paper notes, some work has been accomplished in this area, how these messages relate to public relations and specifically risk communication, needs to be further explored. The implication of such analysis is that wise public relations personnel should understand how journalists work with "characters and plots" to increase the community fit leading to the development of mutually beneficial relationships.



References
        Brehmer, B.  (1987).  The psychology of risk.  In W. T. Singleton & J. Hovden (Eds.), Risk and decisions (pp. 25-39).  Great Britain: John Wiley & Sons.
        Bridges, J. A., & Nelson, R. A.  (2000).  Issues management: A relational approach.  In J. A. Ledingham & S. D. Bruning (Eds.), Public relations as relationship management: A relational approach to the study and practice of public relations  (pp. 95-115).  Mahwah, NJ: Lawrence Erlbaum.
        Burke, K.  (1966).  Language as symbolic action.  Berkeley: University of California Press.
        Coombs, W. T.  (2000).  An extended examination of the crisis situations: A fusion of the relational management and symbolic approaches.  Paper presented at the National Communication Association conference, Seattle, WA.
        Covello, V. T.  (1992).  Risk communication: An emerging area of health communication research.  In S. A. Deetz (Ed), Communication yearbook 15 (pp. 359-373).  Newbury Park, CA: Sage.
        Covello, V. T., & Johnson, B. B.  (1987).  The social and cultural construction of risk: Issues, methods, and case studies.  In B. B. Johnson & V. T. Covello (Eds.), The Social and cultural construction of risk: Essays on risk selection and perception (pp.vii-xiii).  Dordecht, Holland: Reidel.
        Covello, V.T., von Winterfeldt, D., & Slovic, P. (1987). Communicating scientific information about health and environmental risks:  Problems and opportunities from a social and behavioral perspective. In J.C. Davies, V.T. Covello, & F.W. Allen (Eds.), Risk Communication (pp. 109-128). Washington, DC: Conservation Foundation.
Dawson, B.  (2000, March 29).  Smoke from plant fire concerns some experts.  Houston Chronicle, p. A23.
Dawson, B.  (1989, October 25).  Explosion in Pasadena: No sign of hazardous pollution.  Houston Chronicle, p. A22.
Edelman, M.  (1988).  Constructing the political spectacle.  Chicago: University of Chicago Press.
Einsiedel, E., & Thorne, B.  (1999).  Public response to uncertainty.  In S. M. Friedman, S. Dunwoody, & C. L. Rogers (Eds.),  Communicating uncertainty: Media coverage of new and controversial science (pp. 43-57).  Mahwah, NJ: Lawrence Erlbaum.
Entman, R.M.  (1993).  Framing: Toward clarification of a fractured paradigm.  Journal of Communication, 43, 51-58.
Fischhoff, B.  (1990).  Risk issues in the news: Why experts and laymen disagree.  News Backgrounders: Foundation for American Communities.
Fisher, W. R.  (1987).  Human communication as narrative: Toward a philosophy of reason, value, and action.  Columbia, SC: University of South Carolina Press.
Fishman, M.  (1980).  Manufacturing the news.  Austin, TX: University of Texas Press.
Gitlin, T.  (1980).  The whole world is watching: Mass media in the making & unmaking of the new left.  Los Angeles: University of California Press.
        Goffman, E.  (1974).  Frame analysis: An essay on the organization of experience.  Cambridge, MA: Harvard University Press.
Grunig, J. E.  (1989).  Symmetrical presuppositions as a framework for public relations theory.  In C. H. Botan and V. Hazleton, Jr. (Eds.), Public relations theory (pp.17-44).  Hillsdale, NJ: Lawrence Earlbaum.
Hallahan, K.  (1999).  Seven models of framing: Implications for public relations.  Journal of Public Relations Research, 11(3), 205-242.
        Hanson, E., & Kennett, J.  (2000, March 28).  Phillips plant kills 1, injures 52: Facility has an infamous safety record.  Houston Chronicle, p. A1.
        Heath, R. L.  (1993).  A rhetorical approach to zones of meaning and organizational prerogatives.  Public Relations Review, 19(2), 141-155.
        Heath, R. L.  (2001).  A rhetorical enactment rational for public relations: The good organization communicating well.  In R. L. Heath (Ed.), Handbook of public relations (pp. 31-50). Thousand Oaks, CA: Sage.
        Heath, R. L., & Nathan, K.  (1991).  Public relations' role in risk communication: Information, rhetoric, and power.  Public Relations Quarterly, 35(4), 15-22.
Heath, R.L., & Palenchar, M, J. (2000).  Community relations and risk communication: A longitudinal study of the impact of emergency response messages.  Journal of Public Relations Research, 12(2), 131-161.
Heath, R.L., & Palenchar, M. J. (2001).  Risk Communication and Multicultural Publics: Competing, Compatible or Conflicting Cultural Narratives.  Paper to be presented at the Intercultural Communication Conference, Miami, FL.
        Kasperson, R. E., & Stallen, P. J. M.  (1991).  Risk communication: The evolution of attempts.  In R. E. Kasperson & P. J. M. Stallen (Eds.), Communicating risks to the public (pp. 1-14).  Netherlands: Kluwer.
        Kreps, M.A., Campbell, J. T., & Warren, S.  (1989).  Explosion in Pasadena/outside Phillips: Shock wave stuns homes, schools.  (1989, October 24).  Houston Chronicle, p. A15.
        Krimsky, S., & Plough, A.  (1988).  Plant closure: The ASARCO/Tacoma Copper Smelter: Environmental hazards: Communicating risks as a social process. Dover, MA: Auburn House.
        Leiss, W.  (1996).  Three phases in the evolution of risk communication practice. Annals of the American Academy of Political and Social Science, 545, 85-94.
        Lindell, M. K., & Earle, T. C.  (1983).  How close is enough: Public perceptions of the risks of industrial facilities.  Risk Analysis, 3(4), 245-253.
        Lindenmann, W. K.  (1983).  Content analysis.  Public Relations Review, 39(7), 24-26.
        McCartney, H. P.  (1987).  Applying fiction conflict situations to analysis of news stories.  Journalism Quarterly, 64, 163-170.
        McCombs, M., Einsiedel, E., & Weaver, D.  (1991).  Contemporary public opinion: Issues and the news.  Hillsdale, NJ: Lawrence Erlbaum.
        Merrill, J. C.  (1989).  The dialectic in Journalism: Toward a responsible use of press freedom.  Baton Rouge, LA: Louisiana State University Press.
        National Research Council.  (1989).  Improving risk communication.  Washington, DC:  National Academy Press.
        Nelkin, D.  (1989).  Communicating technological risk: The social construction of risk perception.  Annual Review of Public Health, 10, 95-113.
        Neuman, W. R., Just, M. R., & Crigler, A. N.  (1992).  Common knowledge: News and the construction of political meaning.  Chicago: University of Chicago Press.
        Newman, K. M. (1988). Toxic chemical disclosures: An overview of new problems, new opportunities for the professional communicator. New York: Public Relations Society of America.
        Otway, H.  (1992).  Public wisdom, expert fallibility:  Toward a contextual theory of risk.  In S. Krimsky & D. Golding (Eds.), Social theories of risk (pp. 215-228). Westport, CT:  Praeger.
        Palmlund, I.  (1992).  Social drama and risk evaluation.  In S. Krimsky & D. Golding (Eds.), Social theories of risk (pp. 197-212). Westport, CT: Praeger.
Pearson, A., & Mintz, B.  (1990, April 20).  Phillips fined in explosion: Violations to cost firm $5.7 million.  Houston Chronicle, p. A1.
        Rendon, R., Bryant, S., Hopper, L., & Antosh, N.  (2000, March 28).  Phillips plant kills 1, injures 71: Investigators seek reason for inferno.  Houston Chronicle, p. A1.
        Sethi, S. P.  (1977). Advocacy advertising and large corporations: Social conflict, big business image, the news media and public policy.  Lexington, MA: D.C. Heath.
        Shoemaker, P., Chang, T. K., & Brendlinger, N. (1987).  Deviance as a predictor of newsworthiness: Coverage of international events in the U.S. media.  In M. L. McLaughlin (Ed.), Communication yearbook 10 (pp. 348-365).  Newbury Park, CA: Sage.
        Solomon, R. C.  (1983).  In the spirit of Hegel.  New York.
        Tuchman, G.  (1978).  Making news: A study in the construction of reality.  New York: Free Press.
        U. S. Department of Labor.  (2000).  Investigation of explosion that killed one worker, injured 69 (National news release USDL: 00-277).  Washington, D.C.: Author.
        U. S. Department of Labor.  (1991).  Investigation conducted by OSHA at the Houston chemical complex (OSHA standards and interpretation and compliance letters:7.21.91).  Washington, D. C.: Author.
        van Dijk, T. A.  (1988).  News as discourse.  Hillsdale, NJ: Lawrence Erlbaum.
        Warren, S.  (1989, October 23).  Explosion rocks Pasadena plant.  Houston Chronicle, p. A1.

Table 1: News Sources by Grouped Categories
1989
frequency
(n=706)
1989
percentage
2000
frequency
(n=269)
2000
percentage
Percentage
change
Government
(local, state, federal)
154
21.8%
34
12.6%
-42.2%
Phillips/contract companies
176
24.9%
47
17.5%
-29.7%
Employees and families (Phillips/contract, former)
155
22.0%
85
31.6%
+43.6%
Legal
6
< 1%
56
20.8%
+1980%
Medical
23
3.3%
23
8.5%
+157.6%
Industry
(associations, unions, near-neighbor facilities)
99
14.0%
17
6.3%
-55.0%
Community
(activists, community leaders, residents)
93
13.2%
7

2.6%
-80.3%
slight discrepancies in the totals of the percentage columns are due to rounding error.


Table 2: Specific News Sources
1989
frequency
(n=706)
1989
percentage
2000
frequency
(n=269)
2000
percentage
Percentage
change
Federal government
79
11.2%
24
8.9%
-20.5%
State and local government
75
10.6%
10
3.7%
-65.1%
Phillips' management and spokespersons
159
22.5%
46
17.1%
-21.3%
Phillips' contracted companies - management and spokespersons*
17
2.4%
1
< 1%

Phillips' employees, including contract
111
15.7%
59
21.9%
+39.5%
Phillips' former employees
2
< 1%
5
1.9%
Employees' families
(Phillips/contract)
42
5.9%
21
7.8%
+32.2%
Industry
55
7.8%
4
1.5%
-80.8%
Unions
44
6.2%
13
4.8%
-22.6%
Medical
23
3.3%
23
8.5%
+157.6%
Attorneys - Phillips/contract companies
2
< 1%
22
8.2%
+720%
Attorneys - employees/families
(Phillips/contract)
3
< 1%
29
10.8%
+980%
Judges*