Entrepreneur: Start & Grow Your Business

Air mail: NPR sees "community" in letters from listeners.


by Reader, Bill
Journal of Broadcasting & Electronic Media • Dec, 2007 • National Public Radio

Nearly every week, listeners of National Public Radio's news programs hear a musical segue and then the announcement, "Today we read from your letters ... " What follows is NPR's version of newspapers' "letters to the editor"--a segment in which NPR presents excerpts from listeners' comments. Each "letter" is clearly just an excerpt of a longer submission, and the whole segment is just a snapshot of the dozens, perhaps hundreds, of the total submissions each show receives in a given week. The resulting collage of truncated commentary is one of the few regular NPR features that provides a glimpse of NPR's listeners. But what is the nature of that glimpse? What is NPR's purpose for providing it? What can the practice reveal about the intersection of journalism and the public?

This study explores those questions using a mixed-methods approach in the tradition of some important "gatekeeping" studies, such as David Manning White's foundational case study of "Mr. Gates" (White, 1950) and Dan Berkowitz's study of gatekeeping of local television news (Berkowitz, 1990). Although this study addresses a gap in journalism research--there is little published research of broadcast "letters to the editor"--the more substantive purpose is to explore the idea that "imagining community" influences journalism gatekeeping. To that end, the study combines two theoretical frameworks. The first is "imagined community," which suggests how large groups of people with similar interests can view themselves as parts of distinct communities (Anderson, 1991). The second framework is the "news making" branch of inquiry, which suggests that the process of making news is heavily influenced by news industry culture (Berkowitz, 1997, pp. 169-171). The research questions are meant to explore, via a study of NPR's letters from listeners segments, whether the process of imagining audiences as communities could be a process of news making, and, if so, whether the resulting imagined community could reflect the professional values of the journalists rather than the values of the "real" communities they serve.

Literature Review

Overview of Letters-to-the-Editor Research

The idea that journalists construct publics when they edit and publish letters to the editor is not new. Wahl-Jorgensen (2002) gave an excellent treatment of that concept in an ethnographic study of letters-selection procedures at newspapers in the San Francisco region. Nor is there any novelty to the theory that journalists' interpretation of "what's news" is largely informed by journalists' own professional values and rituals (Schudson, 2003; Tuchman, 1978, pp. 182-185). Several researchers have specifically investigated journalists' perceptions of their audiences, often finding that journalists know very little about their audiences, and in crafting the news place more emphasis on the interests of fellow journalists than on the interests of their audiences (Burgoon, Stacks, & Burch, 1982; Gans, 1979). Along those same lines, letters sections also are highly mediated and selective, and in the end perhaps more representative of journalists' ideals than the values of the audiences themselves (Wahl-Jorgensen, 2002). It is known that letters to the editor are not written by a representative sampling of society--most published letters are written by people who are middle-aged, upper-middle income, highly educated, and White (Reader, Stempel, & Daniel, 2004). Beyond matters of self-selection, however, news editors/producers invariably (and, perhaps, necessarily) inject their own personal and professional preferences into the selection and packaging process. For example, journalists' general distrust of anonymous speech explains why many editors hold anonymous letters in low regard and, as such, most of them reject unsigned letters regardless of their content (Kapoor, 1995; Reader, 2005). Although journalists clearly exert strong gatekeeping controls over the letters they publish, letters also can influence the journalists. For example, journalists can use letters forums to gauge public opinion (Herbst, 1990), to make editorial or news decisions (Davis & Rarick, 1964; Pritchard & Berkowitz, 1991), and to enhance newspaper readership (Kapoor, 1995; Ryon, 1992). Many journalists and scholars agree that a handy indicator of a publication's professionalism and credibility is the robustness of its letters-to-the-editor section (Aucoin, 1997; Hynds, 1992; Lauterer, 2006; Shaw, 1977). However, researchers have long found letters to be unreliable indicators of public opinion, and alignment with public opinion is largely coincidental (Forsythe, 1950; Foster & Friedrich, 1937; Grey & Brown, 1970; Hill, 1981; Hynds, 1992). Although all of the above mentioned articles focused on letters to the editor in newspapers, one might expect that similar limitations would apply to letters submitted to television and radio news outlets.

Brief History of Broadcast Letters to the Editor

Because so little has been written about broadcast letters, it is necessary for this study to include a brief history of the phenomenon.

A clear difference between the two media is the fact that, unlike the relatively unregulated print media, U.S. broadcasters are subject to government regulation, and the Federal Communications Commission often has regulated expressions of opinion over the airwaves. Prior to 1987, the FCC's so-called "Fairness Doctrine" required stations to facilitate dissenting viewpoints. Many scholars contend that because of the Fairness Doctrine, many radio stations did not air controversial opinions to avoid having to provide costly air time to opposing viewpoints (Aufderheide, 1990; Brennan, 1989; Cronauer, 1994). After the 1987 repeal of the Fairness Doctrine, the chilling effect seemed to dissipate, as many struggling AM stations became profitable by airing long-form opinion talk shows in which hosts could opine freely without having to accommodate opposing views (Albarran & Pitts, 2001, pp. 51-52). Although the FCC still requires stations to provide response opportunities if they broadcast personal attacks or station editorials, equal-opportunity claims are not required in most cases (Albarran & Pitts, 2001, p. 53). Therefore, broadcast media are in no way required to provide audience feedback segments on their newscasts; to do so is voluntary, just as it is for print media.

One of the more notable proponents of reading audience letters on the air was Dick Salant, the president of CBS News during the 1960s and 1970s and a member of NPR's board of directors in the late 1980s. In his memoirs, Salant explained: "Over the years, I struggled with ... a letters to CBS News broadcast series--the broadcast equivalent to print's letters to the editor" (Salant, 1999, p. 232). Part of the problem was the sheer volume of letters CBS News received, a number Salant estimated to be 150,000 letters annually (Salant, 1999, p. 232). Salant said he would respond personally to some of those letters, writing: "Because broadcasting is such an extraordinarily one-way street with not even a regular letters-to-the-editor opportunity, it was important, as a safety valve at least, to try to respond to mail" (Salant, 1999, p. 232). Salant eventually implemented an on-air letters segment for the 60 Minutes news show, a feature that media historian Richard Campbell wrote "celebrates viewer diversity through its semi-regular 'Letters' segment.... It is perhaps the only program in the history of prime-time American television that, within its limited and carefully edited forum, explicitly encourages diverse and alternative readings of its own narrative interpretations" (Campbell, 1991, p. 176).

Research of NPR's Audience

But who, exactly, is writing those comments? This author could find no published research into the demographics of NPR letter-writers, but there is some research about NPR's audience that might help give a sense of who might be writing such letters. A 2002 study found that NPR had an estimated 27 million listeners, and the average NPR listener spent at least half of his or her radio-listening time listening to NPR (Bailey, 2004). NPR's Morning Edition and All Things Considered, both of which have regular "letters" segments, accounted for 23% of all listening to U.S. public radio (Giovannoni, Peters, & Youngclause, 1999). NPR's own audience research in 2004 showed NPR's audience skews male (54%); that 85% of listeners are White and 10% are Black; that 66% are between 25 and 54 years of age (48% are 35 to 54); that 69% have household incomes above $50,000 per year; and that the majority have lifestyles that include public involvement (61% claim to vote), patronage of performing arts (51% attend theatre, concerts, or dance performances), travel (72% traveled domestically in the previous year, 38% overseas), and computer use (85% own computers, 71% use online services) (Corporation for Public Broadcasting, 2005). They consistently rate learning about the world around them as being "very important" (Corporation for Public Broadcasting, 2005, p. 23). NPR's audience also is predominantly "iconoclastic" in that they are much less likely than most Americans to make sense of the world through traditional or religious views (Corporation for Public Broadcasting, 2005, p. 24). NPR's audience also is highly educated. In 2004, 68% of NPR news listeners had college degrees, and 32% had attended graduate school (Corporation for Public Broadcasting, 2005, p. 4).

But middle-class intellectuals were not the target audience of NPR pioneer William Siemering, who wrote the original NPR mission statement and served as the network's first program director. Siemering wanted NPR to be a vehicle for diverse views from all of America's cultural corners, not just college towns where affiliated stations would be based (Engelman, 1996, p. 116). An overarching goal was to have NPR serve as a forum for multiple and differing points of view--Siemering envisioned "a structured approach for direct people-to-people communication, eliminating some of the middle information brokers. We would ... facilitate connections of ideas and form networks of unions of common interest...." (Siemering, 1979, p. 35-37). Whereas some saw that communitarian ideal come to life via NPR's programming, during policy debates in Washington many critics accused NPR of skewing toward certain demographics, with self-proclaimed liberals accusing NPR of appealing too much to White, middle-class men, and conservatives accusing NPR of being too deferential to "Great Society forces that had been seen as undermining traditional U.S. economic and spiritual values" (Rowland, 1986, p. 262). Rowland (1986) noted that political pressures coupled with technological advances pushed public broadcasting toward trying to reach wider audiences with programming that would have national appeal. Over the years, NPR shifted from primarily broadcasting locally produced programs to the nation to producing many popular national programs to be broadcast through local stations--as such, NPR has changed its approach from connecting geographically dispersed communities to attempting to serve a nationwide community based on shared interests and values of NPR listeners (Stavitsky, 1994). In selecting and reading letters from listeners, NPR could be applying those same principles.

Theory: News Work and Imagined Community

The theoretical framework of this study melds two well-studied frameworks-imagined community and news work--to make predictions about why and how NPR presents and constructs its letters from listeners segments. Simply put, this study predicts, first, that NPR's goals for the letters segments are to create a sense that NPR is a community of listeners, and, second, that NPR ends up constructing an imagined community that reflects the journalists' own professional values rather than any shared goals or values of those who submit the letters. Past research supports those predictions.

Crafting audience forums is part of the news work, or "news making," process studied by Gaye Tuchman (1978), Michael Schudson (2003), Dan Berkowitz (1997), and others. News-work research can help explain the subjective nature of gathering, packaging, and disseminating information to "enable geographically dispersed individuals to know something about one another, one another's ethnic and neighborhood groups, and events in group life" (Tuchman, 1978, p. 4). Although letters forums differ from news sections, their production is still a journalistic process subject to common constraints on news media, including professional standards used in gatekeeping and agenda setting, limitations of time and space, and organizational norms used to recruit, train, and manage journalists (Donohue, Olien, & Tichenor, 1989). All of those constraints apply to letters to the editor for example, editors value letters that are short and to the point, that are well-written, that reference current news, and that are signed by the writers (Kapoor, 1995; Reader, 2005; Wahl-Jorgensen, 2002).

Journalists sometimes interpret letters to be representations of shared views within their broader audiences (Hynds, 1992; Pritchard & Berkowitz, 1991). As such, letters segments also can be seen as exemplars of imagined communities, which Benedict Anderson defined as collectives in which "the members of even the smallest nation will never know most of their fellow members, meet them, or even hear of them, yet in the minds of each lives the image of their communion" (Anderson, 1991, p. 6). Anderson's focus was on the creation of national identities, but he noted that a daily collection of news reports in a newspaper also exemplifies his theory, as a newspaper takes varied and dispersed independent events with nothing in common (except for when they occurred), and "the arbitrariness of their inclusion and juxtaposition ... shows that the linkage between them is imagined" (Anderson, 1991, p. 33). Several scholars have built upon Anderson's theories to view news media as both members of and facilitators of imagined communities, from the biases they reveal as they cover immigration issues (Vukov, 2003)to their assumptions about race and culture in coverage of the O. J. Simpson murder trial (Silberstein, 2003) to how journalists, through shared discourse, come to see themselves as a coherent community of like-minded professionals (Zelizer, 1993). The concept also has been applied specifically to the idea that radio is used in developing nations as a tool for community building by providing a forum for public discourse (Hartley, 2000).

Taken together, these two frameworks can suggest that imagined community isn't just an outcome of a news-making process--such as a letters segment for NPR--but is actually an integral part of that very process. For example, NPR producers might see letters segments as recognizing the NPR "community" of listeners and as ways to connect to their audience. But in constructing those segments, NPR producers likely employ their own professional values much as they would when making other editorial decisions, such that the community depicted in those segments is based more on the journalists' professional values than on the goals or values of the listeners who write. By extension, the community the journalists construct would likely appear to be more like the journalists themselves than the "true" audience.

Research Questions

To explore the above theory using NPR's letters segments, the study employed the following research questions, to be explored via interviews with producers:

[RQ.sub.1]: How do NPR producers manage their letters segments?

[RQ.sub.2]: What are the goals NPR producers have for those segments?

The analysis applied to those questions were then used to formulate two additional questions, to be answered via textual analysis:

[RQ.sub.3]: To what extent do NPR's letters segments express the journalistic values of balance, accountability, and accuracy expressed by the producers?

[RQ.sub.4]: To what extent do NPR's letters segments use inclusive language to express a sense that NPR listeners are members of the NPR community?

Method

The mixed-methods approach of this study is similar to that used in several notable gatekeeping studies, particularly the foundational study by White (1950) of a single wire editor, "Mr. Gates." In that study, White analyzed the wire content of the newspaper in question, and then compared those findings to statements made by the editor regarding why and how he made his selections. White found that the choices made by "Mr. Gates" were largely subjective and were based on his personal preferences and his assumptions about his readers rather than on any real understanding of the goals and values of those readers (White, 1950). White's method was directly replicated by Snider (1967) and Bleske (1991), who discovered similar findings--when making editorial choices, journalists rely more on their own (individual or shared) values than on any serious effort to really understand the values of their audiences. Berkowitz (1990) used a more robust combination of content analysis and qualitative interviews to study the gatekeeping procedures of a single television news station in Indianapolis, and found that the gatekeeping process was susceptible to many other factors than an individual's preferences, such as group dynamics, aesthetic considerations, and resource limitations--but, again, the process was influenced more by journalists' own considerations than on any real attempt to understand audiences.

Methodologically, those studies first analyzed the content quantitatively, then used qualitative interviews to further explain the results of the content analysis. This study uses qualitative methods only and reverses the process, using interviews to identify and explain journalists' "values" which were then applied to a textual analysis of the letters segments those journalists constructed. The approach seemed appropriate for this study, as the goal was to, first, explore the ways NPR producers consider their imagined communities in their gatekeeping roles, and, second, to examine how those considerations of community are articulated via the selection, packaging, and presentation of letters from their audiences.

Questionnaires and Telephone Interviews

Questionnaires and telephone interviews were used to gather information from producers responsible for managing letters at the three news shows. This part of the project focused on [RQ.sub.1] (how letters segments are managed) and [RQ.sub.2] (NPR's goals for letters segments). In the summer of 2004, the three producers were first sent questionnaires asking for general information about how letters are managed, then each was interviewed during the winter of 2004-05 via telephone to allow them to clarify or expand their written responses. The questions related to this study were intentionally broad to allow for unprompted expressions, and none of the research concepts were mentioned in the questionnaire. Follow-up questions during telephone interviews were derived only from responses to the questionnaire.

Textual Analysis

Textual analysis was used to determine a large-scale overview of the ways letters segments represented the producers' stated journalistic goals of "balance," "accountability," and "accuracy" ([RQ.sub.3]) and the degree to which inclusive language is used to suggest the producers' idea that the NPR audience is a community ([RQ.sub.4]). This portion of the study used a census of transcripts for all the letters segments broadcast in 2003 (the most recent complete year in early 2004, when this study was started) for three popular NPR news programs: All Things Considered (51 letters segments, 289 total letters), Morning Edition (40 segments, 162 letters), and Weekend Edition Sunday (29 segments, 91 letters). In all, the analysis involved 542 discrete letters in 120 segments.

Each letter was analyzed for "balance," "accountability," and "accuracy." "Balance" was a value the producers expressed as the need to present different viewpoints in letters, with deference to opinions expressed by multiple listeners or opinions that add new information to the earlier reports. In the textual analysis, "balance" was assessed in terms of letters that clearly agreed with, clearly disagreed with, or "added to" (neither agreed nor disagreed with) statements in those reports. "Accountability" was expressed by the producers as the presentation of letters that commented on the quality of NPR's journalism; in the textual analysis, "accountability" was assessed in terms of letters that either praised or criticized NPR and its work. "Accuracy" was expressed by the producers as the need to correct factual errors in previous reports; this concept was analyzed at the "letter" level (individual letters used to correct factual errors in news reports), and at the "segment" level (segments referring to "non-letter corrections," or corrections to earlier news reports that were not attributed to letter writers).

Each segment was then analyzed for "inclusiveness," which is the concept derived from producers' expressions that NPR journalists value and routinely consider feedback from listeners. Expressions of inclusiveness were only analyzed when a host suggested multiple people had written on a particular topic. Drawing from psychological research that tests the influence of "inclusive pronouns" on individual perceptions of being included in groups (Crisp, Hewstone, Richards, & Paolini, 2003; Perdue, Dovidio, Gurtman, & Tyler, 1990), this study grouped "multiple letter" references by NPR into four categories: "representative second-person" (example: "John Smith spoke for many of you by writing ..."); "generic second-person" (example: "Many of you wrote regarding ..."); "third-person" (example: "Many listeners wrote regarding. .."); and "non-personal" (example: "We received many letters regarding ..."). In the analysis, the two second-person categories were considered indicative of "high inclusiveness," as they might position listeners to be among the "in group," and the third-person and non-personal categories were considered to indicate "low inclusiveness," as they suggested listeners to be among the "out group."

The textual analysis considered the frequencies of the above characteristics not as generalizable predictors, but rather as more qualitative indicators of "how a particular issue is being presented in a large number of texts" and to consider "the sense-making practices of audiences" (McKee, 2004, p. 128). Specifically, this analysis looked for ways that the selection and packaging of letters segments could be representations of the producers' journalistic values and, in turn, interpreted by NPR's audience as representative of the values of the NPR community.

Findings and Analysis

Questionnaires and Interviews

[RQ.sub.1]: How do NPR producers manage their letters segments? The three producers averaged 26 years each with NPR. None had handled audience feedback letters in previous jobs, and none had received formal training as to how to manage letters. None had formal letter-handling guidelines. All three indicated that managing letters segments is important work: "This is one of the most important jobs I do at NPR," one stated.

The producers estimated that they each reviewed 100-200 responses per broadcast, mostly e-mail (NPR stopped soliciting paper letters in 2003 after several other offices in Washington, DC, received envelopes containing hazardous powders; by the time this manuscript was written, NPR had switched to soliciting feedback via a form on the NPR Web site). (1) None kept detailed accounts of the total letters received. The producers estimated that only 2% to 5% of letters get selected to read on the air, but, again, they had no data to support those estimates.

The selection and handling of letters varied somewhat by show. One producer had editorial assistants "'weed' the large volume of letters and forward (to) me about 15-30 a day," whereas another noted: "1 read all the e-mail that comes in." The third producer reviewed letters "for general appropriateness: Is it for our show? Is it accurate? ... "All three said acceptable letters should be short and engaging: one producer looked for letters that were "articulate" and "to the point"; another sought the "best-written" and "pithiest" letters; "thoughtful," "critical," "clarity," and "brevity" were attributes sought by the third. Contrary to standard practice at most newspapers (Kapoor, 1995), letter writers usually were not contacted if their comments were chosen for the segments. Two producers said they usually only contact a writer to clarify the pronunciation of his or her name. One producer will occasionally contact a person who writes "a fervent letter lamenting that we won't respond. It upsets me to think that people think we don't read their letters."

Rather than being concerned about holding writers accountable for their letters, the NPR producers rather applied the principle of "accountability" to letters that held NPR responsible for its own reporting. For example, all three noted they will only select letters that respond specifically to NPR reports (as opposed to other topics of interest to the letter writers). Also, all three noted that they try to edit letters down to the most important points, removing such things as "scolding" unsupported claims that NPR shows a liberal/conservative bias. One producer rejected "nasty" comments about NPR or its staff; another rejected "sappy praise" for NPR and its on-air journalists.

Letters responding to other letters are not considered. Noted one producer, "Why perpetuate the dialogue? That's not the job of [letters segments]." Another stated: "We rarely will broadcast a listener comment on another listener's comment ... [that] could lead to an unending cycle of responses. But we do it occasionally." The process thus clearly limits the discussion to direct feedback from listeners about the programming--that is, NPR wants the segments to be about NPR and its reports, rather than as segments about what the broader audience might want to discuss.

The producers all said they strive for balance in the selection process, and do so by considering the volume of mail received on specific topics to determine which letters are most representative. One producer specifically looked for letters that "represent a point of view shared by other listener[s]." Another noted:

With news stories, I also look for letters from both sides of an

issue. However, that also depends to some extent on the volume

coming in and the approximate ratio of letters. For instance, if we

deal with a controversial subject and get 20 letters critical of

what we did and only one supportive, I probably wouldn't air the

supportive letter.... Sometimes I will air more than one letter

that is either supportive or critical, but then they must highlight

a different aspect of an issue.

The producers also described how the letters segments are used to ensure accuracy. Corrections are an integral part of NPR's letters segments, and all three noted that most corrections are initiated by listeners. Said one, "For the most part, listeners are the first to point out an error to us, so [the letters segment] seemed like a natural place to put [corrections]." Two noted that corrective e-mails often arrive almost immediately after an error is broadcast, and the report in question can be corrected for later and archived feeds of the show. All three noted that minor or obvious errors are usually not attributed to letter writers; one noted that "I think I can write a correction better than most listeners. People, when correcting us, can be circuitous.... In the time it takes to say, 'Joe Schmoe corrected us for saying such-and-such,' we can just do the correction."

Thus, to answer [RQ.sub.1], NPR's letters segments are managed with high regard for three common journalistic values--balance, accountability, and accuracy. The producers strive for balance by giving preferences to letters about topics that generate the most mail or that add substantive information to earlier reports. By selecting only letters that respond to the reports on specific shows, the producers limit the forum to one of audience feedback, rather than a general forum in which listeners can help to set agendas. And they use letters to promote accuracy by using the letters segments to issue corrections.

[RQ.sub.2]: What are the goals NPR producers have for the letters segments? All three noted that their staffs find letters useful. One producer noted that the mail is "a method of evaluating listener response to the program." Another stated: "We often talk about comments during editorial meetings, and when we get a flood of mail criticizing the balance or approach to a particular story--particularly a political one--we sometimes tailor upcoming coverage or story assignments accordingly." The third stated:

Informally, we will sometimes judge a segment as successful if it

receives a large number of complimentary e-mails. Of course, the

opposite is true.... And we are all aware that the amount of mail

that comes in does not necessarily reflect the opinion of all our

listeners.

The themes of "including" and "connecting" were repeated often by the producers, in both the questionnaires and the telephone interviews, and without any prompting. One producer stated: "Radio is an intimate medium, and many listeners feel a personal bond with a show or its host ... airing their letters is a way to strengthen that bond." Another responded:

Broadcasting these e-mails on a regular basis is one of the most

important dialogues we can have with our listeners. It is, quite

literally, the "public" in public radio. When people write, it

means we have reached them in one way or another, and that is

important to me and to the show.

The third offered: "It lets listeners know we're listening to them, not just broadcasting at them.... It keeps us honest." That producer added:

In every producer's mind, there's an "idealized" listener who may

or may not reflect the real people who hear their work. Keeping

tabs on the listener mail reminds me I'm not producing this show

for myself or people "like" me. Listener comments are a constant

reality check, which I believe helps me do my job.

Stated another: "Listeners get a sense that in addition to them listening to us, we listen to them." That producer gave an example: When a popular host of one show was reassigned (and, in short order, resigned from NPR), "We were 'blizzarded' with very emotional responses, and we had a mass mailing of comments about what had happened." The producers noted other instances that attracted considerable e-mail--specifically the terrorist attacks of September 11,2001, and the explosion of the U.S. Space Shuttle Columbia in February 2003. "When that happens, I really think, 'Y'know, I'm proud of this place,'" one noted, adding, "It just shows that NPR is a part of people's lives, and that's really great." Another producer volunteered, "Making our listeners feel part of a community is perhaps the greatest benefit of airing comments" (emphasis added).

Overall, this part of the study suggests two general findings. First, consistent with findings of past gatekeeping studies, the process of selecting and editing letters is one in which the producers apply their own journalistic principles (here "balance," "accountability," and "accuracy") to the process of "making" the letters segments. Second, the producers' statements about the goals for the segments clearly (and, sometimes, literally) suggest that the producers want the segments to facilitate the community of NPR listeners by "including" them, via their letters, in both behind-the-scenes and on-air aspects of the news-making process. Thus imagining community is not just a goal, but a deliberate process used by these producers. And in that process, the producers largely rely on their own professional values and preferences as they "create" an image of that community, rather than truly attempting to understand the values and preferences of that community.

On its own, however, the above analysis can only present the procedures and goals NPR producers apply to their on-air letters segments. The next part of the study explores the extent to which those procedures and goals become evident in the content of the segments that get aired.

Textual Analysis

The analysis began with a tabulation (from transcripts obtained via LexisNexis) of the letters and segments for each show in 2003. In that year, the three news programs presented 542 letters, with a median of four letters per segment. The hosts made 233 announcements that "multiple writers" had responded, with a median of two such references per segment. Letters addressed 376 discrete topics (median of three topics per segment). Of the total letters, 104 agreed with comments made during the reports, 191 disagreed, and 231 added information that could not be classified as agreeing or disagreeing. Regarding NPR itself, 101 letters praised NPR and 133 criticized NPR. Seventy-eight letters offered corrections or clarifications to information in reports. In four letters segments, the announcers did not read or identify any individual letters, but only issued corrections attributed to "many listeners." Twenty people were allowed to read their own letters on the air.

Table 1 compares frequencies (and means) across the three shows in regard to the concepts of "balance," "accountability," and "accuracy." The data show considerable differences among the three programs, supporting the idea that there is an idiosyncratic nature to letters selection even within a single news operation such as NPR.

The dominance of "disagree" letters (n = 191) over "agree" letters (n = 104) suggests that the NPR producers might have applied their value of "balance" as providing more time for negative comments about news reports than for positive comments, although the largest group of letters (n = 231) simply added information in the form of personal observations or historical perspectives from listeners. Likewise, the dominance of "bad NPR" letters (n = 133) over "good NPR" letters (n = 101) suggests that NPR producers might express "accountability" by providing more opportunities for comment to NPR's critics than to NPR's supporters. "Accuracy" was represented in the 78 individual letters of correction, as well as in the four segments devoted entirely to corrections attributed only to "many listeners."

In response to [RQ.sub.3], the above first-level analysis shows that all three programs had letters segments that were constructed to strongly reflect the producers' own values for accountability, balance, and accuracy, even though those values were expressed somewhat differently by each producer. Much like White (1950) found with his study of "Mr. Gates," the selection of letters for NPR's news programs was perhaps more a reflection of the values of the individual gatekeepers at each show than on the values of NPR as a whole.

In response to [RQ.sub.4] (use of inclusive language), the three NPR programs regularly used "multiple writer" statements to indicate that they received more letters than they could acknowledge on the air (n = 233, with a mean of 2.04 such references per segment). (It should be noted that a relatively small number of letters (n = 20) were read on-air by the individual writers themselves, and such a presentation could be viewed as highly inclusive.) But in terms of "multiple writer" statements, the largest category (69%) was "low inclusiveness" statements, mostly third-person (n = 104, or 45%, were akin to "Many listeners wrote ... ") and "non-personal" (n = 56, or 24%, were akin to "We received many letters regarding ... ") The remainder represented highly inclusive language: 46 (20%) were generic second-person ("Many of you wrote regarding ..."); and 27 (11%) were representative second-person ("John Smith wrote for many of you, stating ...").

Because the largest group of "multiple writer" statements were third-person, and the second largest group were non-personal, one might argue that when constructing these segments, NPR's producers were more prone to use out-group rhetoric toward listeners than in-group rhetoric (i.e., "we" the journalists are getting feedback from "them" the listeners), perhaps contradicting their stated goals of using the letters segments to, in the words of one producer, include "the public in public radio." This finding could be explained by the commonly understood divisions that exist between large media organizations and their audiences.

This study stopped short of a more in-depth study of the rhetoric of the individual letters that were selected to be on air to assess other "values" identified by the producers, such as the quality of the writing or the tone of comments (i.e., some letters were presented as "humorous" or "hostile" or "somber"). Such an inquiry would be worthwhile for future research, but is beyond the intended goals of this study.

Discussion

The primary goal of this study was to explore the concept that, as part of their work in news making, journalists might imagine their audiences as communities, and then represent those communities such that they appear to share the same professional values as the journalists themselves. At the imagining stage, NPR producers clearly selected comments for their letters segments based on the journalistic values of balance, accountability, and accuracy, rather than any consideration of the actual values/desires of those who sent the letters (for example, they rejected letters that didn't refer to NPR reports, and they also rejected letters responding to other letters, thus essentially ignoring and omitting from the community those who had other interests or who wanted to engage in a debate with other listeners). Second, they described the goals of such segments not in terms of "customer service," "regulatory compliance," or "company policy," but rather in unambiguous and consistent terms of "including" listeners in the news-making process and creating a sense that NPR is a community in which the journalists and audience members work together. Those beliefs, not surprisingly, were reflected somewhat in the text of the resulting letters segments.

The segments also included suggestions that letters came not just from the few individuals whose comments were read on the air, but from many other listeners, sometimes referred to informally and intimately as "you," but more often referred to with detached out group pronouns. Absent from the segments were more comprehensive summaries of all of the news reports that attracted letters (averaging just three topics per segment, the letters segments clearly omitted comments on the dozens of other topics covered each week by the three shows). The segments were inclusive only of those who wrote about reports that the NPR producers viewed as worthy of commentary, or whose comments appealed to the journalistic standards and whims of the producers. As such, and despite their best intentions to produce letters forums that represented the real community of NPR's massive audience, in the end NPR's producers were much like other gatekeepers in that they ended up creating letters segments that most strongly reflected their own professional values.

That is by no means meant to criticize NPR or its producers for the limitations of its letters segments--as one of the few broadcast news outlets that provides such forums, and as one that obviously does not use such forums purely for self-promotion, NPR should be lauded for providing such a service. The earnestness expressed by all three producers during the interviews (not to mention their willingness to devote several hours over a period of several months to this research project) should indicate that they have strong positive attitudes and interest toward their letters segments. But like their newspaper counterparts, NPR's letters segments may create an image of a community that is not representative of the whole audience, perhaps even of all of those who send letters. That may be an unavoidable, and perhaps even necessary, limitation of producing such features. But it is a limitation nonetheless, and one that is worthy of consideration by all who participate in such forums, from the writers of the letters, to the journalists who handle them, to those who read or hear them.

From a theoretical standpoint, this study goes beyond the ideas that journalism can facilitate community (see Park, 1929; Stamm, Emig, & Hesse, 1997), that community can influence journalism (Donohue et al., 1989), and that journalism itself represents an interpretive community (Zelizer, 1993). Rather, it suggests that imagining community is as much a process of journalism as a product of it, a process that is largely defined by the sociology of the news profession. Journalists may try to imagine community just as they try to be balanced, to promote accountability, and to strive for accuracy. The mechanisms journalists use for achieving and evaluating those ideals may likely be based more on the shared discourse within the journalism community than the discourse of the broader public.

For example, journalists often demonstrate "balance" by reporting "both sides" of a story, when there might be six or eight or hundreds of different sides; or a demonstration of "accuracy" could be a statement such as "almost six-million dollars" rather than the awkward "five million, eight hundred and sixty-seven thousand, three-hundred and forty-two dollars and seventy-six cents." Of course, the fulfillment of ideals is limited in all professions--just as physicians cannot cure all ailments or repair all injuries, journalists cannot present all of the facts or document all possible viewpoints. So if the most commonly identified journalistic values are necessarily limited, how, then, might the journalistic goal of imagining community be limited? That's a line of inquiry ripe for future study.

One obvious limitation to that process may be that the very nature of a national news outlet such as NPR requires journalists to imagine their communities to a much greater degree than, say, the editor of a weekly newspaper in a small, rural town. With an estimated 27 million listeners, it would be impossible for NPR's news staff to personally meet even a representative sample of its entire national (even international) audience. NPR is by its very nature disconnected from its audience in nearly all ways that generally are attributed to community (interpersonal communication, shared world views, shared meeting places real or virtual, and so on). To compensate for such disconnect, NPR may attempt to construct community via its letters segments, and then incorporate that constructed community into other aspects of its news work, such as reading letters to identify news topics or to judge the "success" of previous reports by the volume and nature of mail received. As evidenced by the producers' own statements, their reading of audience mail had an influence on more than just the letters segments--NPR staff used those letters to assess their own work, to gauge public opinion, even to modify their consideration of what to cover and how to cover it.

A challenge to the above conclusion might be that NPR (and perhaps all journalists) make use of audience feedback simply as a means of attracting and satisfying their customers. This study provides evidence to dispute such arguments. The producers' rhetoric demonstrated a desire to establish strong connections between themselves and their audiences, saying such things as "many listeners feel a personal bond with a show" and "broadcasting these e-mails.., is one of the most important dialogues we can have with our listeners" and "[it] lets listeners know we're listening to them, and not just broadcasting at them." The journalistic process of imagining community is not necessarily a crass market consideration, but could be a genuine attempt by some (many?) journalists to not just serve a community, but rather to be a part of that community, and to have their work reflect the values of that community.

Although this study does not assess the degree to which NPR's imagined community might correlate with the true characteristics of its audience, or whether that audience in turn views itself as a community of its own, it does set the stage for future research into how imagining community can influence the journalistic process, and whether the size of those communities (and the size of the news organizations within those communities) might be a factor in assessing the journalist-community relationship. Future research along those lines might do well to look beyond mere questions of audience desires and satisfaction, and focus more precisely on ways to measure, analyze, and investigate the aspects by which audiences might also conceptualize community, perhaps through letters to the editor--be they in print, online, or over the airwaves.

References

Albarran, A. B., & Pitts, G. G. (2001). The radio broadcasting industry. Boston: Allyn and Bacon.

Anderson, B. (1991). Imagined communities. New York: Verso.

Aucoin, J. (1997). Does newspaper call-in line expand public conversation? Newspaper Research Journal, 18(3, 4), 122-140.

Aufclerheide, P. (1990). After the Fairness Doctrine: Controversial broadcast programming and the public interest. Journal of Communication, 40, 47-72.

Bailey, G. (2004). Mega Trends: The public radio tracking study, special report. The Public Radio Tracking Study. Retrieved May 31,2007, from http://www.ARAnet.com

Berkowitz, D. (1990). Refining the gatekeeping metaphor for local television news. Journal of Broadcasting & Electronic Media, 34, 55-68.

Berkowitz, D. (1997). Social meaning of news. Thousand Oaks, CA: Sage.

Bleske, G. (1991). Ms. Gates takes over. Newspaper Research Journal, 12(4), 88-97.

Brennan, T. J. (1989). The fairness doctrine as public policy. Journal of Broadcasting & Electronic Media, 33, 419-440.

Burgoon, J. K., Stacks, D. W., & Burch, S. A. (1982). The role of interpersonal rewards and violations of distancing expectations in achieving influence in small groups. Communication, 11(1), 114-129.

Campbell, R. (1991). 60 Minutes and the news: A mythology for middle America. Urbana: University of Illinois Press.

Corporation for Public Broadcasting (2005). Profile 2005: National Public Radio station audiences. Washington, DC.: Public Radio Satellite Service. Retrieved March 7, 2005, from http://www.prss.org/resources/profile.cfm

Crisp, R. J., Hewstone, M., Richards, Z., & Paolini, S. (2003). Inclusiveness and crossed categorization: Effects on co-joined category evaluations of in-group and out-group primes. British Journal of Social Psychology, 42, 25-38.

Cronauer, A. (1994). The fairness doctrine: A solution in search of a problem. Federal Communications Law Journal, 47(1), 51-77.

Davis, H., & Rarick, G. (1964). Functions of editorials and letters to the editor. Journalism Quarterly, 41(1), 108-109.

Donohue, G. A., Olien, C., & Tichenor, P. J. (1989). Structure and constraints on community newspaper gatekeepers. Journalism Quarterly, 66(4), 807-812.

Engelman, R. (1996). Public radio and television in America: A political history. Thousand Oaks, CA: Sage.

Forsythe, S. A. (1950). An exploratory study of letters to the editor and their contributors. Public Opinion Quarterly, 14(1), 143-144.

Foster, H. S. Jr., & Friedrich, C. J. (1937). Letters to the editor as a means of measuring the effectiveness of propaganda. The American Political Science Review, 31(1), 71-79.

Gans, H. J. (1979). Deciding what's news: A study of CBS Evening News, NBC Nightly News, Newsweek, and Time. New York: Pantheon.

Giovannoni, D., Peters, L., & Youngclause, J. (1999). Audience 98: Public service, public support. Corporation for Public Broadcasting. Retrieved May 31, 2007, from http://www.ARAnet.com

Grey, D., & Brown, T. (1970). Letters to the editor: Hazy reflections of public opinion. Journalism Quarterly, 47(3), 450-456, 471.

Hartley, J. (2000). Radiocracy. International Journal of Cultural Studies 3, 153-159.

Herbst, S. (1990). Assessing public opinion in the 1930s-1940s: Retrospective views of journalists. Journalism Quarterly, 67(4), 943-949.

Hill, D. 13. (1981). Letter opinion on ERA: A test of the newspaper bias hypothesis. Public Opinion Quarterly, 45(3), 384-392.

Hynds, E. (1992). Editorial page editors discuss use of letters. Newspaper Research Journal, 13(1,2), 124-136.

Kapoor, S. (1995). Most papers receive more letters. The Masthead, 17(2), 18-22.

Lauterer, J. (2006). Community journalism: Relentlessly local. Chapel Hill: University of North Carolina Press.

McKee, A. (2004). Textual analysis. London: Sage.

Park, R. E. (1929). Urbanization'as measured by newspaper circulation. American Journal of Sociology, 25(1), 60-79.

Perdue, C.W., Dovidio, J. F., Gurtman, M. 13., & Tyler, R. B. (1990). "Us" and "Them": Social categorization and the process of intergroup bias. Journal of Personality and Social Psychology, 59, 475-486.

Pritchard, D., & Berkowitz, D. (1991). How readers' letters may influence edit