MULTIPLYING THE EFFECTS:
FACTORS INFLUENCING MEDIA COVERAGE OF BUSINESS CRISES

A Paper Presented at: The 6th Annual Conference on Crisis Management
University of Nevada at Las Vegas August 7, 1997

Robert B. Irvine, APR
Senior Consultant
Institute for Crisis Management
1161 East Broadway
Louisville, KY 40204
(502) 584-0402

Dan P. Millar, Ph.D., APR
Senior Consultant
Institute for Crisis Management
Coordinator of the Public Relations Program
Department of Communication
Indiana State University
Terre Haute, IN 46254
(812) 237-3257

Abstract
A growing number of organizations develop and maintain plans for the dual purposes of controlling the operational and communication chaos associated with business crises. A significant portion of a crisis communications plan advises how to react to, and cooperate with, members of the media sent to report on the crisis. Despite preparation, some crisis events receive more attention than others by the media. Why? What is the nature of crises which attracts the attention of the media? This paper seeks to answer the question of media attraction and to offer suggestions for exerting control on media coverage in a crisis.

The presentation divides into four sections. First, we present a brief description of the ICM database. Second, we report the nature of the multiplication factor for selected events, finding an overall ratio of 2.128. However, labor disputes generate 3.279 original stories per event ranking the highest for the 16 crisis categories monitored by ICM. Executive dismissal ranks 16 with a coverage to event ratio of 1.634. The third section suggests qualities of the crisis events which may attract the interest of the press; i.e., conflict, type and availability of news source. Finally, several strategies for reducing the highly focused attention of the media are recommended; i.e., recognize the potential news impact of different types of crises and gain centralized control of the information flow.

Presentation to the 6th Annual Conference on Crisis Management, Las Vegas, 8/7/97

Many organizations develop and maintain plans for the dual purposes of controlling the operational and communications chaos associated with business crises. A significant portion of a crisis communications plan advises practitioners on how to react to, and cooperate with, members of the media sent to report on the crisis. Despite preparation, some crisis events receive more attention than others by the media. Why? What is the nature of crises which attracts the attention of the media? This paper seeks to answer the question of media attraction and to offer suggestions for exerting control on media coverage in a crisis.

The Institute for Crisis Management maintains a database of records developed from the printed news coverage of business crises since 1990. The logic organizing the content of the database allows for an analysis of crisis occurrences by event, industry and company to produce both aggregate and individual data. While the content of the database is based on accumulated news stories, the records can be condensed to identify specific crises. Comparing crisis events to the extent of the news coverage uncovers a predictive multiplication factor. That factor may varies depending on the nature of the crisis event.

The presentation divides into four sections. First, a brief description of the ICM database. Second, a report the nature of the multiplication factor for selected events, industries and companies; i.e., an overall ratio of 2.128. The third section suggests qualities of the crisis events which may attract the interest of the press; i.e., conflict, type and availability of news source. Finally, several strategies for reducing the highly focused attention of the media are recommended; i.e., recognize the potential news impact of different types of crises and gain centralized control of the information flow.

METHODOLOGY
Content analysis of business crisis news stories published in the various print media comprises the methodology used by the Institute for Crisis Management. Using news stories rather than government statistics, ICM maintains a ever-growing database of more than 51,000 records from the printed press accumulated since 1989 from the Dialogue information service. More than 1500 sources are captured by Dialogue including nationally syndicated newspapers (New York Times, Los Angeles Times, Wall Street Journal, etc.), business news wire services, business magazines, business newsletters and trade publications, as well as selected regional newspapers (Louisville Courier-Journal, Indianapolis Star, San Francisco Chronicle, etc.). Records on business are down-loaded and analyzed quarterly. The evolution of the ICM database has paralleled the development of electronic databases and software for penetrating them. The improvements have permitted increased accuracy of story identification and decreased collection redundancy.

Presently the ICM database only uses printed stories. However, negotiations are underway which could add video news reports to the database. Such an addition may permit more accurate counting of the number and description of the content of the stories about newsworthy business crises.

The database uses Standard Industrial Classification (SIC) codes to locate types of industries experiencing crises. With more than 8000 industrial categories, the SIC system allows for a relatively specific identification of types of organizations reported to have crises in print media. For example, SIC code #6021 identifies "banks," SIC code #6035 classifies "savings and loans associations-federal." By selecting SIC codes #60xx-67xx ICM can identify the crises in all types of financial institutions: depository, non-depository credit, security and commodity brokers, insurance carriers, insurance agents, brokers and service, real estate, holding and other investment offices. While the database carries reports on business crises around the globe, the primary focus is on institutions within North American.

The search logic governing the selection of the crises includes a second criterion -- impact. The logic is constructed so that the selection of database records is based upon a term identifying a crisis situation and the actual, or potential, impact of the crisis. The item enters the ICM database only if the nature of the crisis is connected to a consequent of the crisis, real or potential. For example, for "fire" or "explosion" to be included in the ICM database, "fatality," or "xxx dollar loss," or "shutdown" (the impact) would have to be connected in the same record.

Two consequences of establishing impact as a criterion. The database (a) includes only those situations for which negative outcomes to the business have been specified (improving the validity of a "crisis" database), and (b) probably underestimates the number of business crises because it skips stories in which neither the headline nor the descriptors may describe the impact (improving the reliability). The result is a governing logic based upon 73 sets of terms which representatives of Dialogue describe as one of the "most complex and terrifying" search logics they have ever seen.

The ICM database breaks into 16 categories of crises types as the basis for analysis:

Each type represents a variety of similar of events based upon language used by the print media to describe organizational crises. For example, the descriptors accident, blast, earthquake, collide, tornado, flood, derail, hurricane, blizzard, fire, sink, cluster into the "business catastrophe" crisis type. Each crisis category, then, is comprised of several characteristics.

The database uses the language of the story headline and the content of the story to determine the degree of suddenness of the crisis event. For example, some crises occur with no warning --the bombings at Oklahoma City or the Atlanta Olympics, or the Exxon Valdez oil spill. Other crises smolder, the sparks apparent but the organization’s management is unwilling or unable to rescue the situation before it goes "public" -- Dow Corning breast implants, AT&T insider trading, Department of Defense and Department of Energy disposal of chemical and radioactive waste. In short, the nature of the crisis -- sudden or smoldering -- can be identified.

The origin of the crisis can be identified; for example, teacher, worker, executive, vendor, customer, activist, investor. For analysis, the ICM database defines three origins of crises: management, employees, other (i.e., external agent-terrorist, natural phenomenon-hurricane, etc.). However, because the several descriptors remain in the database, the "other" category can be subdivided to provide more precise analysis.

Many records in the database also include company names and states in which the crisis occurred. Because of the manner in which ICM stores the data, the industry, the region (i.e. state), the company can be identified sharpening analysis of individual events and trends.

After downloading the thousands of records for a given period of time, we analyze each entry eliminating duplicates (an AP story may run several times in different publications), recording the sources of information mentioned in the story, and identifying whether or not the story is a continuation of a crisis or a new crisis event. For example, recently a gas pipeline exploded in Indianapolis killing one woman, injuring more than a dozen residents, completing destroying eight homes, and damaging scores of others. A story describing the grief of the family of the killed woman is another story about the explosion. A story revealing the construction pipe which laid the pipe has a history of safety violation and is under investigation by OSHA is recorded as a "new" crisis.

By eliminating duplication and identifying events, we can know the number of stories published about each crisis. By recording sources, we can know who provided information about the event.

Table One reports the total number of for each crisis category during the years 1995 and 1996, the percentage each category compared to the total number of crises and its rank with 1 being the largest). Notice the media published the greatest number of stories about white collar crime: fraud, insider trading, embezzlement, theft. Comparing white collar crime, the highest ranked category with whistleblowing, the lowest, reveal more than ten-and-a-half stories on crime. Because we were interested in determining if the media dispropor-tionately covered certain crises, we counted the number of crisis events.

Table One. Number and Rank of Crisis Categories: 1995-96

Table Two reports the number of events by crisis category, the percent of total events, and the rank order of the categories (with 1 being the largest). Notice the differences between tables one and two. Some categories generate a greater percentage of the total number of stories from fewer crisis events: hostile takeover, labor disputes, sexual harassment, whistleblowing, workplace violence and financial damages. Other categories of crises generate a lowered percentage of stories than warranted by the number of events: catastrophe, casualty accident, environmental damage, class action law suit, consumer action, executive dismissal, discrimination, mismanagement, white collar crime. The shifts in the percentages between the tables suggests the media differentially reports organizational crises.

Table Two. Number and Rank of Crisis Events: 1995-96

To more clearly view the differences, we compared the coverage to events. The ratio, resulting rank and change in rank from the amount of coverage is reported in Table Three. The ratio reports the number of different stories each event generates while the direction indicates the change in ranking from the rank-order of events.

Table Three: Ratio and Rank-Order of Coverage/Events

If the ranking changed by moving from a low to a high ranking, we report the change as "up." If the crisis category moved "down, it changed from a high to a lower ranking. To indicate the amount of change, the 16 crisis categories were quartiled. The amount of change, then, is quantified as the number of quartiles the category moved. For example, in even ranking, sexual harassment appeared in the bottom quartile. In coverage, however, it appears in the top quartile. Sexual harassment, therefore, moved "up" in ranking "three quartiles."

Six categories increased the ranks although five categories climbed at least six rankings indicating coverage disproportionate to their occurrence: sexual harassment (15 to 2), whistleblowing (16 to 5), hostile takeover (13 to 3), workplace violence (10 to 4), and financial damages (12 to 6). Eight categories descended at least five ranking: white collar crime (1 to 14, mismanagement (2 to 11), class action lawsuit (4 to 12), executive dismissal (11 to 16).

At least two conclusions emerge from the data. First, an organizational crisis will engender at least two original stories (stories/events mean = 2.128) before either disappearing from the press or shifting to "aftershocks." By "aftershocks" we mean additional crises for the organization in much the same way that new shocks follow major earthquakes. While following, they are different, create new devastation, and, occasionally, pack more power than the original quake. The cost of litigation of the 1990s far exceeded the cost of environmental clean-up of Prince Albert Sound, Alaska by Exxon in the late 1980s.

Second, some crisis events themselves generate more stories than do others. More negative news stories are written about labor disputes than any of the other categories of crisis. Although many labor disputes occur their coverage exceeds the events by more than three-to-one suggesting that one labor dispute generates multiple stories. Sexual harassment events rank next to last in number of events but generate nearly three stories for every occurrence. Negative stories about hostile takeovers ranked in the bottom quartile of crisis events during the two years examined but ranked third in coverage, encouraging slightly less than two-and-a-half stories of each event. Workplace violence events ranked in the third-quartile but generated more than two-and-a-third stories. Whistleblowing ranked last among the sixteen crisis categories in total coverage and events but each incident led to two-and-a-third stories also. Financial damages ranked at the bottom of the third quartile of events but sixth in terms of coverage generating approximately two-and-a-quarter stories per event.

To determine if the differences were important we compared the two highest ranked ratios, labor disputes and sexual harassment, with the two lowest ranked, executive dismissal and consumer action. The chi-square statistic was significant at the .02 level suggesting the differences were not attributable to chance.

In addition, the rank-order of the coverage-to-event ratio reveals some clustering. Labor disputes differ from sexual harassment which differs from hostile takeovers with each difference statistically significant (chi-square exceeding .05). Workplace violence, whistleblowing and financial damages although showing no statistical difference among them differ from the three ranked above them (chi-square exceeding .10) as well as the cluster directly below them (chi-square exceeding .20) containing defects & recalls, catastrophes, discrimination and environmental damages. The remaining six categories--mismanagement, class action law suit, casualty accident, white collar crime, consumer action and executive dismissal--tend to cluster together. We intend to subject the data to more rigorous and powerful statistical analysis in the future attempting to tease-out the differences between crisis event coverage.

What elements of the crises with the highest proportion of negative news stories--labor disputes, sexual harassment, hostile takeovers, workplace violence, whistleblowing and financial damages--foster increased reporting by the press, reporting disproportionate to their occurrence? Answering that question is the purpose of the next section of this paper.

ELEMENTS OF CRISIS EVENTS STIMULATING MEDIA COVERAGE

One means of examining our data comes from the work of Berkowitz and Beach (1993). They posited that journalists both initially learn about, and gather subsequent information about, events and issues from news sources. Most news stories originate from news sources seeking to have their views of events distributed through the mass media. Government, business and interest groups tend to use top executives as sources. What Berkowitz and Beach label "unaffiliated sources" and low level employees appear in the news less frequently than do upper management personnel.

The use of routine sources evolves from the reporter’s need to gather information effectively (i.e., gather accurate, timely material) and efficiently (i.e., at least cost). Journalists commonly use routine sources for planned events, those which produce predictable news at an expected time.

However, when reporters create their own stories, as in the case of a crisis when "routine" is disturbed, they have more source options available to them. Berkowitz and Beach expect that non-routine news will have a different mix of sources than will routine news.

Drawing on the database of negative news stories, we can contribute to argument that different stories lead to a different mix of sources. Each item entered into the database includes descriptive terms. From those descriptors can be gathered the important sources of information reported in each article. We have clustered those sources into eight categories: employee, executive, union leaders, government (elected and regulatory), judicial activist, consumer and customer. Table Four reports the percentage of "mentions" for all crisis categories combined compared with the two highest ranking coverage/event ratios and the two lowest.

Table Four: Percent and Rank-Order of Sources Mentioned in Crisis News Stories

Their assertion that different events breed a differential mix of sources is supported by our data. What might be surprising is the relative unimportance of the official and routine source of information from an organization, the executive.

Berkowitz and Beach also predict that conflict news will bring a different mix of sources. They develop the theme in the following manner:

In non-conflict stories, the range of views about a subject is relatively unidimensional, so that the reporters can effectively cover a story through official channels . . . For controversial issues, though, a more diverse range of voices makes an attempt to be heard. Journalistic concern for objectivity then lead journalists to try to provide information from this controversy, first, because the community has already acknowledged that a disparity exists and, second, because covering conflict news can benefit a media organization by drawing greater readership. In conflict news, then the mix of news sources should differ from non-conflict situations. (6)

When a crisis arises from conflict between competing parties, then, a journalist will be more likely to find non-routine news sources from which to build the story and or stories. More sources may reasonably generate more stories seems a reasonable assumption.

Berkowitz and Beach introduce a second variable, proximity of the journalist to the event. Journalists "proximate" to the event have access to a larger pool of news sources, both official and unofficial, than have "non-proximate" journalists.

The accessibility of routine and non-routine sources, the appearance of conflict, and the proximity of the reporter to the event may enhance our insight into the elements of crises which encourage disproportionate negative news coverage.

Another element which may influence reporting arises from the journalists role as a storyteller. Travis Linn poses a dilemma resulting from the storyteller nature of the journalist:

If a story is to be meaningful and interesting (editors note: attracting greater readership), it has to be about people, specific people. But when that is done, imbalance is introduced. It is similar to the dilemma of the researcher. When the researcher uses anecdotal evidence, the research has life and meaning, but it loses "significance" because anecdotes can and do distort; they leap out from their contexts and assume unwarranted importance. (17)

When a crisis event centers on people, as cause or as victim, then we might expect reporters to focus upon those people in the telling of the story. And, by such a focus, the press "decontextualizes" the event giving great emphasis to the individual behavior than to the political, social, economic, and technological context within which the event occurs.

The "decontextualizing" of crisis events is noted by several writers in the collection of essays call Bad Tidings: Communication and Catastrophe (1989). The edition examines the means by which the press tell the stories of disasters such as Bhopal, TMI, hurricanes, volcanoes, Chernobyl, the AIDS epidemic and terrorist activities. Quarantelli, in another essay in the same book, notices that emphasizing individual behavior tends to give the appearance that individuals behave in erratic, bizarre and non-adaptive manner to crisis. However, he also notes that psychological research suggests just the opposite.

One theme in the research literature is that human beings respond remarkably well to extreme stress. Those threatened by disasters do not break into panic flight. Likewise, they seldom engage in antisocial or criminal behavior such as looting. Similarly, on the whole victims neither go "crazy" or psychologically break down, nor do they mani-fest severe mental health problems as a result of disasters. Those officials and others with community responsibilities do not abandon their work roles to favor their family roles. In the aftermath of a disaster impact, survivors do not passively wait for out-side assistance, but actively initiate the first search-and-rescue efforts, taking the inju

ICM
455 S. Fourth Street, Suite 1490
Louisville, KY 40202, USA
Phone: 502-587-0327
Fax: 502-587-0329

Available 24 hours/day,
7 days/week.

Contact us for more information

copyright © 2008
View our website terms of use