Barnhurst and Mutz-Decline of Event Centered Reporting

Barnhurst and Mutz-Decline of Event Centered Reporting -...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: American Journalism American Journalism and the Decline in Event-Centered Reporting by Kevin G. Barnhurst and Diana Mutz The definition of news has changed in the 20th century. Content analysis of the traditional five Ws in three American newspapers found that stories grew longer, included more analysis, expanded from specific locations to broader regions, placed more emphasis on time frames other than the present, and named fewer individuals and more groups, officials, and outside sources. These trends affected each newspaper’s coverage of three topics: crimes, accidents, and employment. Thus the basic recipe for news—the report of events new to the hearer—has acquired a third ingredient: For a story to qualify as news, journalists now supply a context of social problems, interpretations, and themes. This trend springs from the workings of the news market and the culture of journalism. Shit happens, but that is not necessarily news. Events become news when someone tells you. Even then, whatever happened is not news to you unless it is new to you. These two elements, the report of events and their novelty to the hearer, have constituted the core meaning or denotation of news in the 20th century (Burke, 1973; Schudson, 1982). In Western capitalism, news workers labor to find or manufacture events, and news industries compete to deliver those events first to the largest audiences. Despite general agreement about what makes news, there is a growing consensus that contemporary reporting has altered that definition to deemphasize events in favor of news analysis. These claims apply to television as well as newspapers. Some studies identify the change as fairly recent. Others see a longer historical trend. Almost all locate the changes in examinations of election coverage. Across differences in study context, time frame, and medium, closely related terms describe the shift from descrip- Kevin G. Barnhurst (PhD, University of Amsterdam, 1997) is an associate professor of public communications at Syracuse University. His research interests include the social consequences of the form of news. Diana Mutz (PhD, Stanford, 1988) is an associate professor of political science at the University of Wisconsin–Madison. Her research interests include political communication and public opinion. Earlier versions of this research were presented at the International Communication Association, Albuquerque, NM, May 1995, and at the American Association for Public Opinion Research, Salt Lake City, UT, May 1996. This project originated when the authors shared an office as research fellows at the Media Studies Center, Columbia University. The authors wish to thank Stephen Whitlock, Amy Hansen, and Mira Sotirovic for research assistance. Copyright © 1997 Journal of Communication 47(4), Autumn. 0021-9916/97/$5.00 27 Journal of Communication, Autumn 1997 tive to analytic journalism, from event-centered to interpretive reporting, or from episodic to thematic coverage. What changes might the trend include? Barnhurst (1994) suggested what the “new long journalism” looks like to a reader: Individuals are less important; stories get longer, and there are fewer of them; lines blur between current events and history; locations become more general; individual events are grouped into roundups; and analysis replaces coverage focused on individual events. In other words, “In the long journalism, the house across town didn’t burn; instead, society confronted a chronic wiring problem in its aging stock of housing” (Barnhurst, 1994, p. 15). Changes in newspaper design confirm a trend away from pages containing many small individual news items (Nerone & Barnhurst, 1995). How has the trend manifested itself in newspaper content? To find out, we first review previous work that led us to expect a changing definition of news. We next outline a study designed to measure the specific components of the new long journalism, before proceeding to test each component individually. Then we report the data drawn from coverage of three issues in three newspapers published over the course of a century. Finally, we offer some potential explanations for the historical trends we identified. Previous Studies Much of the evidence that newspapers have shifted away from events and toward analysis has been anecdotal. Some historians have quoted the recollections of prominent newspaper editors and the reports from trade journals and industry conferences (Donovan & Scherer, 1992). Other histories have focused on the social impetus for the change. Schudson (1978) identified two developments earlier in the century that led to what he called “the decline of ‘facts’ in journalism” (p. 134): the notoriety of public relations counselors and their “publicity stunts” and the emergence of government agencies charged with managing wartime propaganda. Schudson showed how these purveyors of the insubstantial came to cast doubt on the reality of events, leaving journalists suspicious of event-centered reporting. He argued that interpretive journalism reached an apotheosis in the 1960s as part of a larger move in America toward the culture of criticism. In his study of presidential election coverage, Patterson (1993) observed a similar trend. He found more than 90% of front-page election stories in the New York Times were descriptive in 1960, but that figure shrank as interpretive stories grew to more than 80% by 1992 (p. 82). In a study covering a much longer period, Schudson (1982) found a similar pattern in press coverage of the State of the Union address. In very early reporting (1790 to 1850), coverage was more or less a stenographic record of the speech. Later coverage (1850 to 1900) consisted of a chronology with commentary on congressional routines. From the turn of the century onward, coverage included message content along with analyses of the political implications. In other words, “the reporting in each successive period became more interpretative” (Schudson, 1982, p. 100). Schudson (1982) argued that television, entering the scene late, still “inherits the trend toward analytical reporting” (p. 110). 28 American Journalism Studies provide some support for a shift in broadcast coverage as well. Hallin (1994) found that between the 1960s and the present, the length of candidate sound bites shrank in the context of increasingly mediated, journalist-centered election reports. In his essay on the high modernism of American journalism, Hallin (1994) argued that this interpretive role for print and broadcast journalists expanded to fill the vacuum left by the loss of political consensus. Although these changes mark a shift away from passive reporting, they also increase the tendency for reporters to insert themselves into stories. Most of the studies containing evidence on interpretation and description in journalism have focused on politics. Schudson (1982) focused on presidential speeches. Patterson (1993), Hallin (1994), and others (e.g., Graber, 1993; Steele & Barnhurst, 1996) focused on the coverage of elections. Iyengar (1991) investigated a broader range of topics, including crime and terrorism, government corruption, poverty, unemployment, and racial inequality. However, Iyengar concerned himself exclusively with the episodic nature of television, and his studies did not address historical trends in either medium. Research on the form of news identified three general stages in the transformation of newspapers over the course of the century. In the first period, “Victorian” newspapers of the late 19th century were a compendium of individual events—many tiny news items without much order that “gave the impression of diversity, randomness and complexity, leaving the reader to make sense . . . of the world” (Barnhurst & Nerone, 1991, p. 803). In the second period, “transition” newspapers changed the form of news pages frequently, as journalists reconstructed the outward forms of their growing cultural authority (Nerone & Barnhurst, 1995). This period has been characterized as a precursor to the third period, when “modern” newspapers began to define the newspaper as an orderly array or spatial map of the day’s events, interpreted in visual priority for the reader (Barnhurst, 1995). The Study This study took up the challenge to measure the trend in content: What direction did reporting take in the last century? We also set out to broaden the focus, not only from one moment in time, but also from one dimension such as politics. In his study of the long-term shift toward interpretation in political news, Schudson (1982) proposed that “similar kinds of transformations have occurred in other types of news stories” (p. 105), but that assertion has yet to be empirically tested. In general, we expected to find a greater emphasis on news analysis and the why part of the traditional five Ws formula. However, we expected the general trend to reveal itself in the other four Ws as well. Where might incorporate a broader context or broader range of locales than would event-centered coverage; when might show greater emphasis on times other than the present; what might include a larger number of events in single stories; and who might deemphasize individuals, so that people become identified less often by name and more often by demography or group affiliation. To test our hypotheses about changes in news coverage, we content- analyzed a 29 Journal of Communication, Autumn 1997 range of different types of newspapers that were continually available over a 100-year period. Because very few papers continuously published for such a long time, a random sample of newspapers was impossible. Instead, we purposely selected three papers with differing circulations that served geographically dispersed cities of varying sizes. These were the Portland Oregonian, the Chicago Tribune, and the New York Times. Our total sample of news stories comprised six independent random samples drawn at 20-year intervals beginning in 1894 and ending in 1994. For each year we wanted a reasonable number of stories on a variety of topics. To cover such a long period, the topics needed to be very general and not time bound, yet adequately well defined to produce reliable coding of stories. For these purposes we chose three topics: accidents, crime, and jobs. We defined these news categories by adapting the terminology from Graber (1980). When classifying stories, we used inclusiveness as our general principle.1 We selected a series of 5-day weeks from each time period. For each of the 6 years, we first used a random number generator to select a list of Mondays.2 To assess reliably how stories got covered (rather than how often a topic appeared), we instructed coders to continue progressing down the list of weeks until they had sampled 40 stories for each topic each year. This strategy produced a total of 240 stories per topic per newspaper, or 2,160 randomly selected stories in all. For each story, coders answered several questions about individual actors or groups in the text, the number of events reported, when these events took place (past, present, future, and whether change over time was mentioned), and where the events happened. They also identified whether the story reported any information about background context, implications, interpretations, or recommendations, and also whether the story reported any causes, general problems, collective social issues, or themes. On a 10point scale, they rated the emphasis from highly specific, event-centered coverage (1) to very general news analysis (10). Finally, they also rated the length of each article from 1 (very brief) to 5 (very long). After an initial coder progressed through a small sample of stories, we had an identically trained second coder go through the same procedures for locating and coding stories. Given the general complexity of the coding scheme, the reliability was surprisingly high. It ranged from .67 to .92, with an average coefficient of .80 across all of the measures reported in this study.3 Coding the number of single events reported 1 We defined crime as all criminal activity, whether violent or financial, against property or persons, and vice. This included crimes committed by individuals and by groups, such as the Mafia, syndicates, and gangs, violence, prison riots, jail breaks, fraud by government officials, and illegal drug-related and labor-related activity (when laws were broken). We specifically excluded ongoing criminal trials after the first report, terrorism and political violence, and ongoing prison reform. We defined accidents as any mishap, including man-made disasters involving traffic and transport and personal injury caused by chance. We also included natural disasters, such as fires, floods, earthquakes, storms, and contagious diseases or epidemics. We defined employment news as anything job related, including legal strikes and labor demonstrations, unemployment, and general economic coverage explicitly mentioning jobs, job creation, or job acquisition. We specifically excluded routine lists or notices of promotions or hires on the business pages. 2 Although the exact dates varied from year to year, the lists identically represented the 6th, 35th, and so forth to the nth Monday of that year, consecutively. 30 American Journalism in an article produced a coefficient of .79. For specific time frames in the articles, the reliabilities averaged .72. Whenever an article mentioned a location, coders marked the smallest site named (from street or address, to city or town, state, region, national, or international). By weighting these indicators according to their progressively broader definitions of location, and then dividing by the total number of events, we created an index of the extent to which events were defined in smaller versus larger geographical areas.4 The indexes produced by the coders resulted in a correlation coefficient of .83. Despite changes in fonts and formats over the period, rating an article’s length on a 5-point scale proved highly reliable, with the coding correlating at .92. Two potentially unreliable measures in the coding scheme required coders to assess whether the articles reported (a) context, implications, interpretations, or recommendations, and (b) causes, problems, or collective social issues or themes. Spearman’s coefficients indicated that coders succeeded in doing this fairly reliably, with coefficients of .77 in both cases. Finally, the most subjective characteristic simply asked coders to assess, on a 10point scale, whether the article emphasized event-specific coverage or news analysis. Despite the difficulty in making this kind of summary judgment, the coding produced a reliability of .72. This was roughly on par with the other coefficients. In short, our coding procedures produced reasonably to highly reliable indicators of some subjective characteristics of news coverage from the past and the present. Because our primary research question concerned the direction of historical change in newspaper coverage, we analyzed the data with an eye toward significant trends over time, which transcended individual topics or newspapers. We tested our hypotheses initially by looking for significant linear trends across the entire pooled data set. We subsequently analyzed coverage within the smaller data sets for individual newspapers and for specific topics, to see how consistent were the trends. The tests within subsets of coverage necessarily involved much smaller samples, and also required larger magnitudes of difference to establish statistical significance. For continuous variables, we used analyses of variance with linear trend tests on the means by year (thus the significance tests reflect the size of the F-values corresponding to year). For dichotomous variables, we employed logistic regression to test for the impact of year, after controlling for issue and newspaper (thus these tests reflect the 3 Based on a subsample of 51 stories, using Pearson or Spearman coefficients (as appropriate) between the scores assigned to each article by the two independent coders. Counts of the number of actors or victims in each article correlated .82 across the two coders. Similar counts of the number of individuals referenced by name only, and those both named and described, correlated at .87 and .91, respectively. Counts of the number of groups in each story produced a high reliability, correlating at .92. Coding the number of actors who were directly involved in a story and those who were outside sources proved more difficult, with coefficients of .67 and .70, respectively. The number of references to the past, the future, and to change over time within each article also correlated at slightly lower, but still acceptable, levels with coefficients of .71, .67, and .77, respectively. 4 In the formula for this index, the entity in parentheses is the number of locations of that kind identified in a given story: 1 (street) + 2 (city or town) + 3 (state) + 4 (regional) + 5 (national) + 6 (international) y= sum (street, city, state, regional, national, international) 31 Journal of Communication, Autumn 1997 Figure 1. Length of articles. Mean length of articles by newspaper and issue, rated on a scale from 1 to 5. value and significance of the logistic regression coefficients). As our results demonstrate, the major findings in this study proved remarkably consistent across news topics and across newspapers. Results We began our analysis with the simplest of questions about changes in the nature of journalism over the past century: Is today’s journalism truly longer as predicted by the “new long journalism” hypothesis? The mean length of articles increased for each topic and each newspaper (see Figure 1). In addition to a positive and statistically significant overall trend (F = 279.88, p < .001), the length of articles increased clearly and consistently within each topic, and across all newspapers. Analyses of variance with linear trend tests confirmed a linear pattern for all six hypothesis tests (F = 112.24 accidents, F = 127.11 crime, F = 73.16 jobs, F = 142.13 New York Times, F = 162.36 Oregonian, F = 28.41 Chicago Tribune, p < .001 in all cases).5 Across newspapers, job stories tended to run longer, and accident stories shorter. Stories in the New York Times tended to be longer than those in the other two newspapers. Regardless of general differences between topics and newspapers, all gradually shifted toward a longer format. In short, the new journalism is at least aptly named. Next, we examined how much the emphasis on analysis and context increased over 32 American Journalism Figure 2. Events versus analysis. Percentage of articles explaining the how and why of events (left scale), and the mean content of articles rated on a scale (right scale) from 1 (most specific event-centered coverage) to 10 (most general news analysis). the century. We first compared the mean ratings on the 10-point scale assessing eventcentered versus analysis-centered coverage. Coverage did indeed include progressively more analysis, as the trend line demonstrates (see Figure 2). The scale scores, which were consistently less than 5, indicate that most stories were still judged to be more event-centered than analytic. However, the overall trend toward more analytic coverage was statistically significant (F = 67.58, p < .001), as were the linear trends within each of the three separate issue samples (F = 60.41 accidents, F = 22.91 crime, F = 15.97 jobs, p < .001 in all three cases). Within subsamples for individual newspapers, the trends were significant for the New York Times and the Portland Oregonian (F = 70.52 and 32.40, respectively, both p < .001), but not for the Chicago Tribune. The bar graphs (see Figure 2) show the percentage of stories that included references to how (whether stories reported any information about background context, implications, interpretations, or recommendations), or references to why (any causes, general problems, collective social issues, or themes). Using logistic regression, we analyzed these dichotomous measures controlling for issue and newspaper and 5 For all omnibus (full sample pooled) tests (often referred to as tests of the “overall trend” or “whole sample” in the text), the degrees of freedom are (1, 2154). For other tests (for specific subsamples by newspaper or issue), the degrees of freedom are (1, 714). 33 Journal of Communication, Autumn 1997 confirmed a positive trend through time toward greater emphasis on both how and why (b = .20, p < .001, and b = .21, p < .001, respectively). We found that the extent of context, implications, and interpretations differed tremendously among topics. Consistent with other research (e.g., Iyengar, 1991), coverage of jobs emphasized context much more than did coverage of crime or accidents. Nonetheless, the historical trend toward greater emphasis on context was significant for each issue. The trends were also statistically significant for each newspaper. For why, the measures reached statistical significance for the New York Times and Portland Oregonian, but not for the Chicago Tribune. Thus, three independent measures of the emphasis on events versus analysis—references to how and to why, and a subjective rating of whether articles centered on events—all showed highly consistent evidence of greater contextualization of the news. Journalism for all three topics in all three newspapers has gotten longer. Coverage has also become more oriented toward analysis, answering questions such as how and why. The share of stories reporting interpretations grew substantially since 1894, as did the share describing causes, problems, and social issues or themes. Compared to a 1-paragraph story on a factory adding jobs in 1894, a 1994 story on the revival of a port on Staten Island ran 27 paragraphs. It analyzed the return of jobs as part of an effort by New York City to regain world-class status lost in the 1960s (New York Times, Sept. 27, p. B1).6 A stenographic style is also evident in early reporting on crime and accidents. In 1894, the Chicago Tribune summarized an incident of assault without fanfare in a single paragraph: James McCune of 319 South Green street, a packer, is at the County hospital with a fractured skull. He was knocked down by William Warrington of 528 South Halstead street, a teamster. The men quarreled at West Congress and South Halstead streets. The police held Warrington without booking him. (April 15, p. 10) After the 1960s, accident stories showed a steep increase in interpretation. Two examples of accident stories illustrate how analysis increased. In 1897, the New York Times contained a two-paragraph item that began as follows: “Four-year-old Dora Cohen was run over before her father’s eyes by a horse and wagon in front of her home, at 87 Hester Street, at 7:30 o’clock last evening. The child’s ribs were crushed in and she died an hour later in her father’s arms” (April 17, p. 8). The second paragraph reported a chronology of the accident, describing the street, express wagon, and driver. Such items disappeared from the New York Times over the century, and by 1994, any accident reported had to have much broader significance. For example, a report on flooding in Fort Fairfield, Maine, included information on state emergency measures and past events, as well as the causes and consequences of the accident (April 19, p. A12). For example, suicide reports like the following were ubiquitous in the early years of the sample: “Louis Knorr, a theatrical man, whose home is at 1,101 Clark Avenue, St. Louis, fatally shot himself through the head in the Public Garden about 7:10 o’clock this evening” (New York Times, April 17, 1894, p. 1). The story went on 6 Quoted material from news stories in our sample is referenced in the text only. Citations are not included in the references. 34 American Journalism to give the text of the suicide note, a list of the contents of the man’s pockets, and the details of his arrival at the hotel in Boston, where he died. Stories like this, which do not lend themselves to larger interpretations, appeared much less often a century later. To understand the components underlying the general pattern of change, we looked at the ways specific aspects of news stories might also have changed. How did the change affect where stories got situated, who the relevant actors were, what event or events got reported, and when—whether the reporters stuck to the present or they addressed past and future, too. Where We first hypothesized that the general trend would incorporate a broader range of locales. Where older, event-centered reports would emphasize highly specific locations such as street addresses, more recent coverage would situate the topics more expansively in regional, national, or international contexts. Using an index constructed to represent the range from narrowly or broadly defined locations in the story, we compared means across years for each topic. The specificity of locations did change over the century (see Figure 3). The index shows a significant increasing linear trend (see the line graph) toward more broadly defined locations (p < .05). Two components of the index (see the bar graphs) moved in diametrically opposed directions. The number of the most specific locations, street addresses, significantly decreased (F = 8.25, p < .01), whereas the broadest locations, international settings, increased (F = 18.43, p < .001). Two crime stories can best illustrate this shift. The first, from 1894, envisions a robbery by focusing tightly on a highly specific location: The Post Office is in a small building a short distance from the Long Island Railroad station. There is a very little room outside of the partition. In one end of the latter is a door, and inside, near the door, stands a small safe. (New York Times, April 21, p. 1) The second, from 1994, reports a murder quite differently. Two maps accompanying the article show the Caribbean and the island of St. Thomas, with the caption, “The slaying of a tourist in Charlotte Amalie has become a symbol of crime in the Virgin Islands.” It begins with the following paragraph: According to the license plates on the cars and buses that hail tourists around this bustling port, the Virgin Islands are still an “American Paradise.” But a surge in violent crime over the last year, including the slaying of a San Diego swimming instructor this week, has put that in jeopardy. (New York Times, April 19, p. A17) This last example embodies the shift to the new long journalism. The story runs 26 paragraphs, the full news hole on its page. It takes the entire island as its location (and reaches the U.S. mainland). The event acquires historical development by linking all homicides related to the place. Government officials describe social problems, blaming the influence of television. A subhead provides the theme based on national origin: “Island officials say mainland values are leading the young astray.” 35 Journal of Communication, Autumn 1997 Figure 3. Distance and locations. Mean number of references to closest and most distant locations (left scale), and the mean distance of coverage, indexed on a scale (right scale) with low scores for events at street addresses, followed by towns and cities, states, regions, and nations. Who We also expected that the new long journalism would deemphasize individuals and increase attention to groups, particularly as they represent larger categories of actors. The results bear out those expectations (see Figure 4). The number of individuals identified by name only declined significantly (F = 23.03, p < .001). This trend proved extremely consistent. It was significant for each issue (F = 18.93 accidents, p < .001; F = 30.17 crime, p < .001; F = 8.28 jobs, p < .01), and for each newspaper (New York Times, F = 5.81, p < .05; Portland Oregonian, F = 7.52, p < .01; Chicago Tribune, F = 41.25, p < .001). In contrast, named individuals were identified increasingly by some kind of demographic information (F = 15.71, p < .001). For example, on April 16, 1894, a front-page New York Times report of a railroad accident in Hazelton, Pennsylvania, listed every casualty by name, occupation, and medical condition in a format that emphasized individuals: “The man killed was DAILEY, PATRICK, of Milton, Penn. He was riding on the freight train. The injured are: ARTHUR—Sunbury, Penn., conductor of freight train; back injured. BIDDLE— brakeman of express train; body bruised and back wounded . . . .” Likewise, an 1894 Chicago Tribune story about an accident at a school listed each injured child’s home address, age, and medical condition in a separate paragraph (April 10, pp. 1, 7). In 1914, individuals’ names still ran in the lead paragraph in a 36 American Journalism Figure 4. Who gets covered. Mean number of references to groups and mean number of references to individuals by name and by name and description (e.g., by gender, race, age) per article. Chicago Tribune article describing an accident: “Four men were seriously injured when their automobile rolled down an embankment near Gary, Ind., today. One of the injured men is F. W. Kurtz of Chicago. The others are Fred Hass, Thomas Murrey, and Frank Whitson of Knox, Ind.” (April 20, p. 11). By 1954, however, there was a greater emphasis on roles, rather than individuals, as illustrated by this lead: “A veteran West Seattle high school teacher who refused to say whether she ever belonged to the Communist party will be dropped from the Seattle school system, the school board announced Tuesday” (Portland Oregonian, April 21, p. 1). In stark contrast to the earlier examples of accident coverage, a jetliner crash in 1994 focused only on the national origins of those involved, and did not list names of casualties (New York Times, Sept. 28, p. A9). In other words, individuals no longer matter in stories as individuals per se, but they take on increasing importance as exemplars of particular human categories or types. The number of groups identified in news stories also increased (see Figure 4). Again, this trend was significant not only in the sample as a whole (F = 20.13, p < .001), but also for the subsamples by newspaper (Times and Oregonian, F = 51.05 and 22.76, p < .001 in both cases; Tribune, F = 5.08, p < .05), and by most issues (acci37 Journal of Communication, Autumn 1997 Figure 5. Roles of individuals. Mean number of references per article to individuals by role in the news as principal actors or victims (left scale), and as directly involved officials or as outside sources (right scale). dents and crime, F = 84.91 and 62.72, p < .001 in both cases; jobs, F = 2.67, p =.10). The trend is straightforward: More groups of all kinds appear in news articles, either identified by name or by description. The roles of individuals identified in news stories changed quite strikingly in several other respects (see Figure 5). First, the number of actors or victims involved in news stories declined considerably (F = 9.32, p < .01) between the 1930s and 1950s, and then leveled off. In contrast, the number of references to officials more than tripled. Again, the trend is supported not only in general (F = 103.73, p < .001), but also for every issue and newspaper (F = 70.81 accidents, F = 19.06 crime, F = 24.12 jobs, F = 63.10 Times, F = 25.24 Oregonian, F = 18.07 Tribune, p < .001 for all six tests). Although the increase appears to be more or less continuous, the upward slope is particularly steep from the 1950s to the present. In addition, we found more and more references to outside sources or commentators, in a trend consistent for all newspapers and issues (F = 22.78 Times, F = 8.43 Oregonian, F = 3.77 Tribune, F = 12.56 accidents, F = 4.71 crime, F = 20.20 jobs, p < .001 for all). Although the number of outsiders remained fairly low compared to directly involved officials, by 1994 stories referred to outside sources 16 times more often than before the turn of the century. Like the tendency to incorporate more officials, the growth occurred more or less continuously but got far steeper since the 38 American Journalism Figure 6. Number of events. Mean number of different events reported per article by newspaper and issue. 1950s. What We began with the assumption that the new long journalism was simply grouping news differently, so that each article reported more events than a century ago. This hypothesis seemed consistent with longer articles that emphasize context and the connections among events. However, the pattern we found did not meet these expectations at all (see Figure 6). In fact, the number of events per article declined significantly (F = 36.10, p < .001). Regardless of which newspaper we looked at, the trend proved significant (Times and Tribune, F = 17.62 and 12.48, p < .001 in both cases; Oregonian, F = 6.64, p < .05), and it was likewise for coverage of accidents (F = 9.87, p < .01) and jobs (F = 27.02, p < .001). The sole exception, crime, stayed about the same. No topic went against the trend. The general consistency for newspapers and topics suggests a fundamental redefinition in reporting, not a simple change in the way reporters group the events they cover. When Has the trend toward a new long journalism influenced the temporal settings of news as well? Did news stories become less centered in the present? As an initial test of this 39 Journal of Communication, Autumn 1997 hypothesis, we looked at the number of references to different time points—past, present and future—within each article. Some issues always contained more temporal references (see Figure 7). Coverage of jobs, in general, referred to far more points in time, and accident news involved far fewer. However, the number of time references per article consistently increased. This conspicuous trend was significant for each issue (F = 12.69 accidents, F = 22.25 crime, F = 33.52 jobs, p < .001 in all three cases) and for each newspaper (Times and Oregonian, F = 45.02 and 20.59, p < .001 in both cases; Tribune, F = 4.70, p < .05). Moreover, the changes occurred with virtually identical timing in each subsample. After a small dip between 1894 and 1914, the slow but steady increase continued through 1994. To understand the specific changes underlying the pattern, we looked for explicit references to change over time, and to either the past, present, or future. Almost without exception, stories dealt with the present, in 1994 as well as before. The lack of variance in this measure made clear that no fundamental change occurred in this regard. Journalists also showed no greater proclivity for speculating about the future (no topic showed a significant linear change). Instead, the curvilinear pattern (see Figure 8) suggests that speculation about future events increases shortly before the turn of each century (a significant curvilinear trend). What drove up the number of temporal points within news articles was a significant increase in references to the past (b = .25, p < .001). In fact, more than twice as many articles now include references to the past than they did a century ago (see the bar graph). Likewise, the number of articles describing changes over time (see the trend line) increased significantly (b = .41, p < .001). In 1894, less than 2% of the articles incorporated the notion of change. By 1994, more than 7% did so (down a bit from the peak in 1974). Considerably more articles now talk about temporal changes than did those a century ago. Reports in 1894 that hardly mentioned the past contrast strikingly with recent coverage. The 1994 shipping article mentioned earlier contains a background section that begins by calling the reopening of the port “a development that was having stark consequences for its maritime history” (New York Times, Sept. 27, p. B1). It goes on to summarize the history of the shipping industry in Manhattan and Brooklyn since the 1600s, reporting specific events from the 1960s and 1970s. Our results suggest that the new long journalism affected many components of newspaper content. In their effort to provide context and interpretation, journalists identified individuals less often by name and more often by demographic group. Fewer ordinary people played roles as actors and victims, replaced by a cast of official sources, outside experts, and commentators. The number of current events went down in these longer stories. References to history and temporal change went up. The index of location also grew, as journalists abandoned the particular street address in favor of broader geographical frames of reference. Who, what, when, where, why, and how— hardly any ingredient in the five Ws formula for journalism remained constant. Clearly the content of news stories changed along with their length; such long stories could not simply offer the same old content. Could the change in length cause these other changes? To answer this question we reanalyzed our findings, controlling for the increased length of stories over time.7 In the summary analyses incorporating data from all newspapers and issues, we found that only two of the previously reported findings became statistically insignificant. The increased number of references to how 40 American Journalism Figure 7. Points in time. Mean number of references per article to different time periods by newspaper and issue. and why no longer constituted significant increasing linear trends, after controlling for length. Surprisingly, all of the other findings persisted, including the trend toward less event-centered coverage and more news analysis. These analyses persuaded us that, whatever facilitating role increased article length may have played, the changes still represent a fundamental redefinition of news. Discussion Having found such robust evidence of the redefinition of news at every turn, we asked whether our results could possibly apply beyond the particular newspapers and topics we selected for study. No three newspapers can represent the entire U.S. press. The fact that the New York Times, the Portland Oregonian, and the Chicago Tribune endured a century makes them extraordinary. Peculiarities in the history of one paper might account for its changing in isolation, but not for such concerted action in newspapers of such differing sizes and reputations, occupying three distinct markets widely separated by geography. Their qualitative span adequately represents the urban U.S. daily press. For most of our results, identical trends occurred at all the newspapers. To attribute such uniform findings to the vagaries of local history would stretch credulity. The few differences in the results match the well-known differences among 41 Journal of Communication, Autumn 1997 Figure 8. Change over time. Percentage of articles referring to past and future time periods (left scale) and percentage of stories including some reference to change over time (right scale). newspapers. Besides the already mentioned tendency to run longer stories, the New York Times made references to more points in time, especially in the last 20 years (see Figure 7). This seems to point to its ascendancy as a national newspaper of record, as well as to its market position as other quality papers closed down in New York. No other notable differences between newspapers turned up over the century. Although a paper might reflect some changes more strongly than others, no single instance contradicts the general trend. All three of the newspapers acted almost simultaneously to change their standard of news. Topical Differences All three topics also changed historically in very similar ways. The differences in emphasis on analysis versus events, and on groups versus individuals, make sense and enhance our understanding of the new long journalism. We expected newspapers to report accidents mostly as isolated events. Generally, 7 For the continuous variables, we used analysis of variance with linear trend tests and length included as a covariate. For dichotomous variables, we used logistic regression controlling for the impact of length. In both cases, we also included the main effects of newspaper and topic. The presence of either a significant linear trend test for year (in the case of analyses of variance) or a significant coefficient for year (in the logistic regressions) confirmed the trends over time. 42 American Journalism that is what they did through the beginning of the century. By 1934, the Portland Oregonian devoted two paragraphs to describing an accident, still in a fairly stenographic style: The engine of Southern railway passenger train No. 135 was derailed tonight at Harrisburg, 14 miles north of here, upon striking a coal tender which broke away from train No. 32 on a parallel track and careened into its path. The engine of train 135 overturned and the engineer and fireman were bruised and cut. None of the cars left the rails and the passengers, although badly shaken up, escaped injury. (April 9, p. 1) Accidents proved the topic least amenable to larger interpretive frames. Especially in American parlance, accidents occupy a largely denotative level: Accidents happen. Even after the new long journalism became firmly entrenched, accident stories remained less analytic than the other topics. The potential for increased interpretation was a bit greater for crimes than for accidents. On the whole, crime reports followed the overall norm in the study, except that the locations became more general only gradually over the century. Crime remained a local, denotative event until becoming identified with social problems after the 1960s. The reporting of crime analyses, interpretations, and themes surged by 1974. These changes accompanied shifts in emphasis away from the perpetrators and toward social institutions such as the courts, prison system, and police departments (see Graber, 1980). The title of a 1974 story in the New York Times suggests the pattern: “April Killings Set Detroit Record and Bring New Focus on Police” (April 22, p. 24). International stories in the same paper tied crimes to the Red Brigades and to Ulster “sectarians” (p. 11), and the next day’s paper covered the origins of the Symbionese Liberation Army (p. 82). On the level of personal experience, losing or getting a job ranks with being the victim of crime or accident as newsworthy. However, a specific individual’s job status rarely enters the realm of news (unless driven by other factors, such as celebrity). Thus, we thought employment would get more thematic treatment as a broader group or social phenomenon. As expected, the coverage was consistently more interpretive. An early job story illustrates how event reporting always included at least some implied analysis: The Hartford City Window Glass Company, which pulled the fires from its larger tank two months ago, resumed blowing glass last night, furnishing employment for 300 additional men. Five hundred men are now employed at the works. The company was prevented from employing 100 more men by the arbitrary ruling of the Secretary of the union prohibiting two gatherers at one “ringhole.” (New York Times, April 18, 1894, p. 2) In 1994, the explanations of how and why take center stage. For example, a story on two union contracts with hospitals introduces a long analytical passage this way: 43 Journal of Communication, Autumn 1997 Labor relations and health care experts agreed that the unions’ strength were [sic] major reasons for the divergent fortunes of the two groups of workers. But they also said that marked differences in the private and public hospitals’ financial health, management structure and ability to adapt to a changing market-place were more fundamentally responsible. (New York Times, Sept. 26, p. B1) A lengthy exposition of these causes and responsibilities then fills out the remainder of the article. Of the three topics, coverage of jobs showed the least dramatic shift toward analysis and interpretation. The reporting of issues and themes, already present in 1894, still grew over the century. The same applies to other aspects of job coverage. Employment and unemployment did not experience the same magnitude of shift toward more actors and more locations, but the general direction of change did follow the overall pattern. The differences we measured in content make sense for these topics, but what about others? Features, business, society, and sports require further study. These “soft” genres partake less in the values of news and may have participated less in its redefinition. Most research already cited dealt with political news and found an overall shift toward analysis. Accidents, employment, and crime, along with politics, encompass a large share of the core of hard news, suggesting a redefinition that affected the news genre in general. Parallel Trends Our results revealed a pattern that parallels the three general stages (described previously) in the form of news over the course of the century. News stories from the Victorian period were the shortest, and even long articles often reported more than one event. For example, Portland Oregonian coverage of a rail strike was divided into six articles, each under a separate dateline: from Great Northern headquarters in St. Paul, from the local lodge in Seattle, from the office of the superintendent in Spokane, from a local paper in Chicago, from an official in Great Falls, Montana, and from the U.S. court of appeals in Milwaukee (April 16, 1894, p. 1). The news content we observed before the turn of the century suggests a journalism secure within its definition of news and self-sufficient on the two pillars of events and novelty. The second period, evident in our sample through the middle of the 20th century, represents the transition, as the new long journalism began to establish itself. Reporting moved in the general direction of analysis and interpretation. However, the period also saw considerable experimentation in content, just as it did in the form of news pages. The third, and most dramatic, period occurred in the last 2 decades in our data. The emphasis on interpretation and social issues increased substantially between 1954 and 1974, and reports became longer. Those years also showed a steep growth in the particulars of the new journalism, including more references to history. The stories already cited on hospital labor negotiations in New York and murders in the Virgin Islands typify this period. They emphasized groups rather than individuals and depended on outside expert sources. They explained a single event by referring to other time periods. They focused on larger regions rather than particular addresses, and they emphasized the how and why rather than the event itself. These changes in 44 American Journalism content paralleled the emergence of modernist newspaper design. Form and content intertwine, and one might expect that the space for articles, the so-called news hole, would have grown proportionately as newspapers grew thicker, but it did not. The increased size of pictures, headlines, and text typography, the larger numbers of graphics, indexes, and self-promotional items, and the bulk of publicity and advertising supplements, together, absorbed much of the added room in newspapers (Barnhurst, 1994). These changes in the form minimized growth in the actual amount of space left for news. Although news stories and newspapers indeed grew fatter, their girth cannot account for the changes we have documented. News today represents a substantial departure from previous styles of reporting. Our findings replicate themselves in several distinct kinds of newspapers and cross various areas of coverage. The same trends appear in other studies of political news and of newspaper form. These facts enhance confidence in our results. In light of the evidence, it seems likely that the new long journalism represents a fairly general change in the core definition of news. Explanations What could possibly explain the changes we observed? Why did reporting move in this particular direction? Several possible reasons fall into two categories, one based in the market for news and the other in the culture of journalism. News Market Several arguments from different quarters suggest that television had a major influence on the content of print journalism. The most common, from journalists and historians, suggests that newspapers adapted to the new competitor by differentiating their product. With its capacity to outstrip the fastest presses, television crippled one of the newspaper’s legs—its ability to report events first. Most predictions of the demise of newspapers depend entirely on the loss of novelty (Katz, 1994). Upon losing their monopoly on timeliness, newspapers found a way to prop themselves up. They made their narratives richer by adding more news analyses (Donovan & Scherer, 1992). If they could not scoop competitors with what editors call the “first day story,” which tells what happened, they could instead write a “second day story,” which keeps the story alive and maintains its value in the marketplace, by telling the story better, with greater depth, explanation, and background. This argument suggests that newspaper reports would become longer and analytical, just as our research confirmed. Another argument often advanced by journalists, and by scholars of visual communication, suggests that newspapers attempted to compete with television head to head. With its capacity to show filmed events in detail and to incorporate graphics and color, television made its version of the news not only fresher but more appealing. To compete, newspapers began to enlarge their photographs, insert subheads and other explanatory display type, and produce more elaborate graphics, eventually adding color. These design devices affected the content of news indirectly, in ways journalists did not anticipate (Barnhurst, 1994). Information graphics shifted emphasis from individuals and toward statistical groups, and located events within maps of larger 45 Journal of Communication, Autumn 1997 geographic regions. The redesigns assigned more space to fewer stories, and incorporated themes and issues into the new summaries, headings, and other typographic display. The dominance of pictures, which seem to depict events directly, made describing events in the text redundant. These changes in the form of news likely contributed to the trends we measured. One problem with the explanations forwarded by news executives and historians is that they focus narrowly on television in the short run, assigning it the lion’s share of responsibility for change. Our data do not support so central a role. Most of the trends we measured began long before the introduction of television. The published recollections of newspaper editors do say they perceived television as a threat and changed their attitudes and news practices in response. Our research, consistent with other longitudinal studies of the period, shows that the trends accelerated after television became a serious competitor. The perception of a competitive threat from television probably gave a boost to the new long journalism, but cannot fully explain its origin. Another problem is the apparent contradiction between these two responses to television. Did newspapers move aside to make room for the faster competitor, or did they confront the competition head-on? This contradiction might be resolved by taking a long-term view, reducing television to one technological innovation within the broader market for news. The apparent contradictions then become competitive strategies, one adaptive (as newspapers made room for another competitor) and the other confrontational. It is possible that newspapers followed both (and perhaps all available) strategies, adjusting their visual structure while differentiating their content to adapt to and compete with television in the news market. Both strategies have antecedents during the century. The latter strategy, confrontation, is an iteration of an old chestnut, in which the Goliath newspaper faces a more adroit challenger, whose technology delivers its blows through the air. Radio, television, and now the new interactive media have all been assigned the role of David. In each case, newspapers may have responded by differentiating their product, adopting a longer, more explanatory form. This long-range view would help account for the early beginnings of the trends we measured. The former strategy, adaptation, goes back at least to the appearance in the 1930s of news magazines. Their relatively leisurely production timetable did not allow them to compete with daily newspapers on novelty. Instead, news magazines found a gap or opening not served by newspapers, and established a new niche in the market. News magazines have made a virtue of the generally long lead time they can give reporters. According to Wills (1983), They learned to stand off from the flow of discrete items filling daily newspapers, to look for longer trends. Something in the very format of such papers suggests that knowledge merely agglutinates—that you stick discrete new items onto an unchanged body of past knowledge. (p. xvii) Of course, the economic logic would seem to get reversed here, unless newspapers are viewed broadly within the market for news. In competitive markets, products often come to resemble each other, and as the news market became more crowded as the century progressed, magazines may have contributed to the general trend toward 46 American Journalism interpretation that affected all the competitors. By defining television as only the most recent in a long series of competitive threats leading newspapers to adopt an expanded and elaborated narrative style, the market economy of news may account, at least to some degree, for the changes we observed. A related argument forwarded by scholars concerns the ways narrative structures on television influenced newspaper stories. The quickly digestible narrative frames used on television news (i.e., story plots with a formulaic beginning, middle, and end) also may have helped move news practices toward interpretive and thematic coverage. Patterson (1993) said the interpretive narrative frames in television news, such as the election as a horse race, influenced the shift he identified in the New York Times. We found that this explanation fell short in the face of the stories themselves. In the articles we read, narrative structure was, in fact, a stronger, not weaker, facet of early reporting. Consider this passage from an 1894 crime report, “Battle with a Polish Mob”: Sheriff Collins waved his hands wildly, and shouted to the mob to stand back, but no attention was paid to what he said, and he drew his revolver and aimed it point blank at the crowd. One! two! three! the shots came from his weapon as fast as he could pull the trigger. Then a long-handled shovel, upraised behind him, descended swiftly, and a corner of it pierced his skull. As he fell another shovel struck him on the right leg, and a stream of blood spurted from this new wound. A dozen gathered around and were aiming fresh blows at him when a man rushed in, and, under the uplifted shovels, cried out: “He is dead; leave him alone.” (New York Times, April 19, p. 1) This example of reporting from a century ago describes the scene and conveys the observable action with a clear plot line. By comparison, the complex analyses of 1994 lack narrative drive. Instead of individual actors described in their setting, the new style offers abstract conceptual frames to explain events, which are described only when emblematic and otherwise merely mentioned, not narrated at all. Television images supposedly have a greater capacity to show the scene and action of the news event. Studies of political news on television describe an increasingly fragmented use of imagery and greater reliance on interpretive frames over the course of the medium’s short history (Hallin, 1994; Steele & Barnhurst, 1996). Neither newspapers nor television newscasts offered fuller narratives in 1994, because, by then, the emphasis throughout the journalism market had moved away from the narrative of individuals, scenes, and events with a beginning, middle, and end. The old journalism market had many newspapers competing for readers, whose purchases responded to particular stories hawked on the street corner. The writing needed a story line to carry the reader through to the end. The new long journalism developed as monopoly news markets became the rule in the U.S. If consumers have only one newspaper (or only one of quality) to choose from, then that paper must aspire to complete, authoritative coverage. Journalism then becomes a reference tool, and consumers use the paper not by reading entire narratives but by scanning and collecting bits of information. The transfixed and captivated reader changes into a 47 Journal of Communication, Autumn 1997 captive but autonomous consumer, and the news event changes from a compelling story into one of a line of goods in a department store. The market thus produces news meant to be referred to, not read. Journalism Culture This longer, more interpretive form of reporting makes sense as part of the broad development of the culture of journalism over the century. Early in the period, an emphasis on science and expertise from within the Progressive Movement found its way into journalism through several avenues, especially journalism education. The 1930s through the 1950s saw many leading journalism schools founded. Despite the predominantly humanistic view of journalism of the time, most began in U.S. universities as social science departments. They tended to hire PhDs to teach their courses, which brought social science approaches into the training of journalists. As a result, “More emphasis was placed on generalizing from specific observations” (Weaver & McCombs, 1980, p. 481). About the same time, the Hutchins Commission recommended that the press provide “a truthful, comprehensive, and intelligent account of the day’s events in a context which gives them meaning” (Commission on Freedom of the Press, 1947, p. 20). The recommendations centered on the idea that journalists should go beyond merely acquainting people with facts of specific events and, instead, put the events into context. Although many professionals demurred, these recommendations still influenced journalistic training, by encouraging interpretative reporting. One popular reporting textbook argued that interpreting the news “involves recognizing the particular event as one of a series with both a cause and an effect,” and that a good journalist will at least be aware of the fact that an item of news is not an isolated incident but one inevitably linked to a chain of important events . . . . The interpreter of the news must see reasons where ordinary individuals observe only overt happenings. And he must study them as the scientist scrutinizes the specimen in his microscope, scientifically. (MacDougall, 1977, p. 12) This concern echoed in the surge of interest in investigative journalism in the late 1950s and early 1960s. Although different from interpretive reporting, the investigative turn pushed journalists to do more than passively transmit events (Guba, 1978). In a 1964 textbook, Copple argued that journalists should go “beyond the reporting of specific, discrete events by providing a context which gives them meaning” (quoted in Weaver & McCombs, 1980, p. 488). Historical events also seemed to underline the importance of going beyond stenographic reporting. In particular, coverage of the McCarthy hearings called into question whose interests were served by an absolute emphasis on objectivity (Meyer, 1973). Of course, investigative reporting gained widespread appeal among journalists during and after the Watergate coverage. A third trend confirmed that the dominant humanistic view of journalism had shifted toward applied social science. In the 1970s, advocates of precision journalism made this synthesis explicit. Meyer (1973) argued that, “To cope with the acceleration of social change in today’s world, journalism must become social science in a hurry” 48 American Journalism (p. 14). Precision journalism made reporters familiar with social science methods and encouraged them to investigate the broad context of related events rather than passively describing individual ones in superficial detail. The social indicators movement gained momentum during this same period and further highlighted the important connections among events. The rise of quantitative data collection, the spread of computers, and the availability of statistical information encouraged journalists to consider the news value of what previously interested only social scientists. The journalistic interest in social indicators lagged about 15 years behind the academic interest, and as the new kind of information became ubiquitous, social trends began “to compete with more circumscribed news events for time and space in the news” (Weaver & McCombs, 1980, p. 479). Besides ideas from these movements, the growing professionalism of journalists themselves also may have encouraged the trends we observed. Over the period, journalists’ levels of education rose, and that may have led them to seek higher status in other ways. Increases in analysis and interpretation might reflect a desire to participate and influence the political process. That sort of clout generally comes with occupational prestige. According to Grant (1971), What separates most journalists from the few great ones is that the latter are not content with knowing what their sources think. They exhibit an independent intelligence that seeks to wrest meaning from the torrent of events rather than acting as mere transmission belts. (pp. 368–369) Journalists have also argued that the sheer scale and complexity of contemporary society make it necessary and desirable for them to identify larger patterns and themes for readers. People now conduct more of their day-to-day lives through complex indirect social relationships, influenced by media, markets, and various bureaucracies. The massive scale makes the lessons of local social and political life difficult to translate directly into the knowledge people need (Calhoun, 1988). With the centers of power more distant from everyday life, journalists take on the task of helping readers understand “the big picture.” A professional responsibility to interpret complex contemporary U.S. institutions also depends on journalists’ concepts of the audience. In the survey research emerging at midcentury, journalists and politicians found what looked like a tool not only for taming the menagerie of information (Herbst, 1993), but also for understanding the masses. Surveys seemed to leave little doubt about the low levels of information held by most Americans. If journalists of the 19th century considered the public attentive and astute enough to draw connections among events, this new evidence did not sustain that respect. Instead, it most likely encouraged them to think that newspapers needed to organize the world for the reader. Of course, the familiar arguments for professionalization run through this rationale. Increased complexity requires advanced training and specialized knowledge, which erect barriers to entry into the profession. The resulting monopoly over knowledge elevates news workers in status, justifying increased power and privilege. The consequent asymmetry also denigrates the citizenry, which becomes dependent upon the patronage of professional news interpreters. The culture of journalism thus stands in 49 Journal of Communication, Autumn 1997 contradiction to the market for news. Although structured to serve autonomous, selfdirected consumers, the new long journalism takes a condescending attitude. The experts talk down to a public considered lazy and ill informed. Like other professionals, journalists must write to impress their fellows first. The new long journalism also reflects a larger cultural transformation. Historian Gordon Wood (1991) argued that the way people attempt to understand the world has changed generally. Wood described the United States of past centuries as one in which: the question asked of events was not “how did they happen?” but “who did them?” Specific identifiable individuals did things and were personally accountable for what happened. If the price of bread rose suddenly, then a particular baker or merchant could be blamed. The political and social world still seemed small and intimate enough to hold particular men morally responsible for what occurred within it. (pp. 60–61) Changes in the culture of news, as seen in the education of journalists, their status as practitioners, and their concepts of audience and evidence, parallel the more often cited changes in the technology and market for news, as well as much broader cultural transformations. These explanations describe the setting where the new long journalism emerged. Conclusion We began with a flip rendition of news denotation, to point out that it stands on two legs—the report of events and their novelty to the consumer. Our study suggests yet a third leg has joined that core meaning. To qualify as news these days, an event also must fit into a larger body of interpretations and themes. Many things that happen to people, although novel, must now get ignored unless reporters can link them to something bigger. We thus conclude with one more question: Is the news better, more compelling or lively, with this third, supernumerary leg? Based on our experience reading 100 years of news, we must answer no. The recent stories are fiercely dull. Even with the luxury of paper copies, every coder complained of boredom. The length and complexity of the reports made them difficult to quote or recount. Turning to the older content, we all anticipated the task of scrolling through microfilm with dread. That form makes the head spin, and the content promised no current appeal. The stories, however, were riveting. We retold favorites and made copies to pass around. This one, about a fire in Milwaukee in which nine firefighters died, lists each casualty in a separate paragraph and describes events in detail: It was a few minutes after 5 and the firemen were pouring great streams of water on the burning roof and the water was beginning to trickle through into the theater below, when the entire roof over the theater part of the building suddenly began to sink and in an instant dropped to the floor, fifty feet below. The ill-fated firemen had not a moment’s time to think of the possibilities of escape. A cry of horror went up 50 American Journalism from the firemen who saw the awful catastrophe. The people from the hotel, who had left their rooms, knew nothing of the frightful tragedy that was taking place in the theater. The crowd of onlookers who saw it knew nothing of the disaster until the firemen who had escaped with blanched faces ran to the street and tremulously told of the sad fate which had befallen their brave brothers, who were buried beneath the burning girders, roof, and floors in the theater. (Chicago Tribune, April 10, 1894, pp. 2, 7) Politicians and civic leaders talked about Chicago’s fire “problem,” historians knew its background, but we learn none of that here. Instead, a story tumbles out without a pause. Reading passages like this, we all shook our heads in disbelief—could the story be true? Professional journalism now defines such stories as poorly attributed, sensational, fabricated in the details, and lacking contextual balance. Instead of molding and shaping stories, the new long journalism fabricates other things: abstract themes, expert analyses, and discouraging problems. Readers make the final judgment. At the newsstand, and when ordering and renewing subscriptions, fewer of them choose newspapers, especially the young (Barnhurst & Wartella, 1991). Among the many consequences of the shift to interpretive and thematic news, one may be the most ironic: The new long journalism makes news about ordinary people more appealing: The front page of the New York Times in 1894 contains a story of two Long Island residents who “were speeding their trotters on the Oyster Bay Turnpike Saturday evening. As they tried to pass, their vehicles became locked together” (April 16, p. 1). The drivers jump out, argue, throw punches and begin a “tussel.” We see a grist mill and pond nearby, and suddenly the fight splashes into the canal. As the loser is about to drown, a passer-by intervenes. The case goes “this morning” before a justice, who fines the aggressor $50. All this in about four column inches. No matter how accessible or full of interest, such a narrative could not appear in the New York Times today, least of all on the front page. Its timeless themes, human conflict and the value of individual action, lack the cargo of current issues such as traffic congestion, bridge safety, or menacing drivers. Whatever its origins and consequences, the new long journalism represents a significant modification in the conception of news. References Barnhurst, K. G. (1995). Newspapers as twentieth century texture. In L. Mirsky & S. Tropea (Eds.), The news aesthetic (pp. 22–35). New York: Cooper Union/Princeton Architectural Press. Barnhurst, K. G. (1994). Seeing the newspaper. New York: St. Martin’s Press. Barnhurst, K. G., & Nerone, J. C. (1991). Design trends in U.S. front pages, 1885–1985. Journalism Quarterly, 68, 796–804. Barnhurst, K. G., & Wartella, E. (1991). Newspapers and citizenship: Young adults’ subjective experience of newspapers. Critical Studies in Mass Communication, 8, 195–209. Burke, K. (1973). Semantic and poetic meaning. In K. Burke, The philosophy of literary form: Studies in symbolic action (3rd ed., pp. 138–167). Berkeley: University of California Press. Calhoun, C. (1988). Populist politics, communications media, and large scale societal integration. 51 Journal of Communication, Autumn 1997 Sociological Theory, 6, 219–241. Commission on Freedom of the Press. (1947). A free and responsible press. Chicago: University of Chicago Press. Donovan, R. J., & Scherer, R. (1992). Unsilent revolution: Television news and American public life. New York: Cambridge. Graber, D. A. (1980). Crime news and the public. New York: Praeger. Graber, D. A. (1993). Mass media and American politics (4th ed.). Washington, DC: Congressional Quarterly Press. Guba, E. G. (1978). Investigative journalism. Unpublished report, School of Journalism, Indiana University, Bloomington. Hallin, D. C. (1994). We keep America on top of the world. London: Routledge. Herbst, S. (1993). Numbered voices: How opinion polling has shaped American politics. Chicago: University of Chicago Press. Iyengar, S. (1991). Is anyone responsible? How television frames political issues. Chicago: University of Chicago Press. Katz, J. (1994). On line or not, newspapers suck. Wired, 2 (9), 50–58. MacDougall, C. D. (1977). Interpretative reporting (7th ed.). New York: Macmillan. Meyer, P. (1973). Precision journalism: A reporter’s introduction to social science methods. Bloomington: Indiana University Press. Nerone, J. C., & Barnhurst, K. G. (1995). Design changes in U.S. newspapers, 1920–1940. Journal of Communication, 45(2), 9–43. Patterson, T. E. (1993). Out of order. New York: Knopf. Schudson, M. (1978). Discovering the news: A social history of American newspapers. New York: Basic Books. Schudson, M. (1982). The politics of narrative form: The emergence of news conventions in print and television. Daedalus, 3(4), 97–112. Steele, C. A., & Barnhurst, K. G. (1996). The journalism of opinion: Network coverage in U.S. presidential campaigns, 1968–1988. Critical Studies in Mass Communication, 13, 187–209. Weaver, D. H., & McCombs, M. E. (1980). Journalism and social science: A new relationship? Public Opinion Quarterly, 44, 477–494. Wills, G. (1983). Lead time: A journalist’s education. Garden City, NJ: Doubleday. Wood, G. S. (1991). The radicalism of the American revolution. New York: Vintage. 52 American Journalism 53 ...
View Full Document

This note was uploaded on 02/02/2012 for the course 790 345 taught by Professor Markmajor during the Fall '10 term at Rutgers.

Ask a homework question - tutors are online