Introduction

In just four decades, women have moved from virtual invisibility in television news to high visibility as anchors and reporters. Throughout the world, more women are in front of the camera than at any previous time and more females now occupy journalism programs than ever before (Janis 2001). This study reports an investigation into how women are featured in local newscasts. Specifically, we wanted to determine whether gender made a difference in what stories are reported by anchors and reporters, and whether experts quoted in news stories tended to be predominantly male or female. We also wanted to know whether gender made a difference with respect to who presented the lead story in local newscasts.

Rationale

Women’s involvement in news is an important area of investigation for several reasons. First, if there are systematic gender differences in news stories reported, viewers may come to expect that only certain domains of news reporting are appropriate for men as opposed to their female colleagues. If young women do not see themselves in roles of importance as broadcast journalists, they will not be inclined to enter a profession where they can employ all of their intellectual talents. One path through which media portrayals can influence viewers is identification with media personalities. Viewers who observe recurring media characters such as newscasters, for example, may come to like them, compare them to themselves and want to be like them (Cohen 2001). Research under the rubric of social learning theory has long established that identification is a necessary condition for imitation of character behavior (Bandura 1986). Identification with female newscasters among young women may reproduce the belief that certain stories are the type reported by women, and limit the expectations and foci of what constitutes professionalism among news professionals, according to their gender. Perceived gender differences in what reporters do on newscasts may determine how young viewers accept and imitate the behaviors of newscasters they see on the evening news.

As a large body of cultivation theory and research suggests, when attitudes and behaviors of population sub-groups, including members of various professions, are represented in television programs as typical, viewers’ expectations of them in the real world are created and reinforced. This process of shaping viewers’ world view is designated as mainstreaming (Gerbner et al. (1980). Mainstreaming means that heavy television viewing may blur the differences in people’s perspectives which come from individual experiences and influences resulting in the perception that television mirrors the “real world.” Research has demonstrated that heavy viewers of television give estimates of jobs occupied by males as opposed to females that reflect the artificial world of television rather than the percentages from surveys of real populations of workers (Seels et al. 2004).

Women are over half of the audience for local newscasts; viewer surveys document that women outnumber male news viewers in nearly every one of the top 100 television markets (Pew Research Center 2008). There is evidence from audience surveys that women prefer to see women on news programs as reporters and anchors (Cramer 2002). The dual processes of identification and mainstreaming can underline young viewer’s expectations that certain topics are appropriate for each gender, and limit their expectations of what constitutes professionalism in broadcast news.

A secondary and pragmatic rationale for this study is that as women take their place among news professionals, they gain confidence in their career choice when they view other females in highly visible positions in news programs. If female news professionals are viewed as only possessing talent for certain stereotyped topics, viewers will be likely to accept this limitation as the norm.

Finally, the news profession is not well served if there is gender bias in story assignments. If stereotypical topics are assigned by gender, rather than by expertise or talent, the quality of news suffers. Women reporters have become powerful role models for their female audience (Mack 2003). When Katie Couric was recently appointed as anchor for the C.B.S. Nightly News, critics lauded that decision as a first in the history of network news. Beyond mere presence, however, is the issue of news content. If news viewers perceive that women are predominantly interested in fashion, home maintenance, child care and other stereotypically “feminine” topics, the extent to which they see female news professionals report these stories may limit their desire to enter the news profession.

Women’s visibility has also increased in investigative news documentaries. Nearly two decades ago, network management exhibited reluctance to place women as reporters on “hard news” programs. Linda Ellerbee, who anchored two network magazine shows, said, “You don’t see women doing old-fashioned Edward R. Murrow-style documentaries. The networks are always afraid to make women look abrasive, and hard news looks abrasive.” (Carter 1990, p. 37).

While the visibility of women has risen at the network level, there is very little information about their presence and significance on local newscasts. Defined by audience concentration and size, 210 Dominant Market Areas comprise the U.S. television audience, where the majority of viewers turn for most of their local news (Imber and Toffler 2006). The present study sought to extend our knowledge of the news-gender relationship about network news to these important local markets. Local newscasts in late afternoon/early evening news shows are 60 to 90 min or more in length, while network news is usually only 30 min in length. If female viewers are looking to reporters as role models, it is local news that provides the majority of them.

Research on News and Gender

In terms of newsroom management, Janis (2001) found that in 1972, women held only .5% of news director positions but by 2001 the percentage had risen to 20.2%. While increased appearance of women on news is seen by many as progress, there are indications that women are treated differently than are their male counterparts in the newsroom. In a content analysis of essays by female journalists, Kim (2006) found that women in both print and broadcast media perceived themselves as marginalized in their newsrooms, and that they attributed this marginalization to less access to news due to traditional gender-related roles in their jobs. This perception results in less satisfaction on the job.

A survey of female news correspondents at national networks found that women were less satisfied than men about their work environment, and also less satisfied with their jobs overall (Price and Wulff 2005). Williams (1997) found that gender altered the nature of decision making for female journalists. In his study simulating a newsroom female participants had significantly less interest in peer advice than did male participants. He argued that the reason for this difference is that women are a clear minority in the news organization and may not seek advice from peers because of the pressure to compete against males. He suggests that, “female journalists in television may feel pressure from males to prove their toughness and intelligence to be accepted as more than window dressing” (p. 157).

There is research indicating that although women are increasingly visible as news professionals they are not used as expert sources in newscasts as often as are men. Hermano and Turley (2001) found that women were just 18% of the news subjects, although women made up 41% of announcers and reporters of news. Liebler and Smith (1997) conducted a content analysis of ABC, CBS, CNN and NBC news coverage during the first hundred days of the Clinton administration in 1997, in order to determine whether gender-related differences in reporting were apparent in news content. They found no gender differences in how economic policy and governmental staffing issues were reported, but they did find differences in source selection. Women were rarely used as news sources and were most likely to appear in traditional female stories, while male sources were used more often than female sources and were more likely to be shown in a professional capacity regardless of the reporter gender or policy covered. They also found that women were just as likely as men to use a greater number of male sources.

Another content analysis found that women are sources for less than 20% of news stories but they accounted for 40% of “ordinary citizens” quoted, indicating that they were less likely to be contacted for expert opinion (Media Report to Women 2002). Larson and Bailey (1998) analyzed five years of “ABC World News Tonight’s Person of the Week” segments to identify the attributes of individuals who were selected. They found that 72% of the individuals chosen were male, and 23% were females (5% were cited off camera so that gender identification could not be made).

There are also gender differences in the content of stories that women report. An analysis of television news in Canada concluded that female anchors and reporters were more likely than male reporters to report softer, more general news (Soderlund et al. 1989). Most western (and many non-western) nations’ style of newscasts are comparable, because the historical primacy of U.S. news formats elicited imitation by many budding operations outside of the U.S. (Smith 1990). Australian news, for example, began in the early 1970’s and styles tended to emulate U.S. news formats more than the more reserved style of the B.B.C. (Morecroft 2002). Therefore, conclusions drawn from research on Canadian and Australian news are quite likely to apply to examinations of U.S. news.

Content analysis of U.S. network crisis coverage found relative similarity of male and female reporters’ coverage of international crises, but also found that female reporters were more likely to use female sources (Sutcliffe et al. 2005). Story placement has also been demonstrated to be male-dominated. Analyzing Australian television news, Cann (2001) found that two thirds of all news stories were reported by men, and that males dominated reporting of the earlier stories in the newscast. He argued that male predominance in the lead stories was an indication of male association with the most important stories. Cann (2001) also found that men were over-represented as anchors, reporters, and expert sources. Female reporters dominated in only in “low frequency, lower ranked subjects” (p.162). Male expert sources were in the majority in every area except female sport, while in some areas (disaster, technology, and male sport) female experts were entirely absent. The author reasons that the way men and women report news has become similar but that gender-based divisions are still occurring at a subtle level, such as topic differences of reported news stories.

Although during the past decade the increase in the number of women journalists has precipitated many changes in network news, questions remains whether this increase has resulted in a greater inequality in terms of news story topics and gender of anchors, reporters and sources in the total of viewer experience. Local news precedes evening newscasts, with large numbers of viewers abandoning news for other programs or activities. It is a general rule in the news business that the station that leads in the audience for local news is the highest in revenue in the local market.

Based on the dual rationale of identification and cultivation in forming expectations of role differences in news topics appropriate for men vs. women, and the trends reported in the review of literature, the following hypotheses regarding local news were posed. In light of the results of Soderlund, for example, (1989) and Cann’s (2001) observation of gender differences in stories reported, we expected to find systematic differences between stories reported by men and those reported by women:

  1. H1:

    There is a significant association between reporter gender and the type of stories reported on local TV newscasts.

Since news anchors traditionally come from the ranks of reporters, we had reason to believe that stories reported by anchors would exhibit systematic gender differences in topic.

  1. H2:

    There is a significant association between the gender of anchors of local TV newscasts, and the type of story he or she is assigned.

Sutcliffe, for example (2005) found gender differences in expert sources in American network newscasts (women used female experts more frequently than males) and Liebler & Smith found that females were rarely used as sources in network news. Cann’s (2001) study of Australian national news also found that more males were used as expert news sources. Therefore, we expected to find more male experts cited in local news than their female counterparts.

  1. H3:

    Male expert sources will be cited more often than female expert sources in local TV newscasts.

While we wanted to discover any gender differences in expert sources cited in newscasts, we also wanted to determine whether there were gender differences in nonexpert sources cited. Often, “person on the street” sources are cited in newscasts in the role of witnesses to events. We therefore predicted a similar trend for nonexperts.

  1. H4:

    Male nonexpert sources will be cited more often than female nonexpert sources in local TV newscasts.

Cann (2001) found that male-reported stories outnumbered female-reported stories in the early portion of newscasts. Story placement is a significant issue because the first few stories of a newscast parallel the front page of a daily newspaper, where stories are placed in descending order according to their perceived importance. We predicted that this trend would be in evidence for both reporters and anchors, when there were male and female anchors for newscasts.

  1. H5:

    More stories will be reported by men than by women in the early segments of a local TV newscast.

  2. H6:

    More stories will be reported by male anchors than by female anchors in the early segments of local TV newscasts.

Method

Sample of Newscasts

Three local television news programs in a “top 30” market in southern New England during the weeks of November 8th through November 21st, 2004, were used for content analysis. A market in the top 30 of audience size was chosen because their news production practices and presentation styles are frequently imitated in smaller markets whereas a top 5 market station is typically owned by a network and the same management structures for local and national news render them less representative of local news across the U.S. On advice from station management, early November was chosen as an appropriate time for study because the three month audience rating data were in for the previous quarter, and there would be less potential for “stunting” or providing special news events to boost ratings than for the middle of a sweep period. The three stations selected for study included a CBS affiliate, an NBC affiliate and an ABC affiliate. The CBS station had an 8.5 rating and 18% share for the 11 pm newscast. The ABC affiliate had a 5.3 rating and an 11% share. ABC’s outlet is also the oldest station in the market and the only one with a nonstop news format. NBC had a 7.0 rating and a 15% share in the market. Although there is a Fox Network affiliate in the market it was not included because its news airs at 10 pm (rather than at 11) and its ratings are not comparable in terms of audience size (overall rating less than 3 in this time placement). Also, breaking news, an important component of local news, is not comparable in this regard because of the one-hour difference. These three stations were selected based on their overall high rating, breaking news content and similar evening newscast times at 11 pm.

The newscasts were also chosen according to information provided by phone interviews with personnel from each of the three newscasts including the Executive Producer of the CBS station, the ABC Assistant News Director, the NBC Director of Programming and the Sales Director for Fox Television. Sources from the stations explained that the 11 pm newscast was geared towards an adultaudience compared to other times in the day when children are watching television. According to the news Director of the C.B.S. affiliate, each newscast time has a unique audience with the 11 pm news cast being the most watched. Fox News competes with network entertainment programming in contrast to the other network affiliates (that compete with each other.) In the market profile the 11 pm newscast was the highest rated time period combined for the three stations analyzed.

A total of 11 h were recorded on videotape yielding 580 stories including stories read by the reporter or the anchor (Table 1). Within these 580 stories, 147 news stories were included in which a reporter could be identified in the newscast. Male reporters covered 67 stories and women reported 80. The entire sample of news anchors included three men and three women. A total of 448 stories were included when categorizing what type of stories anchors reported. A total of 202 stories were anchored by men while 197 news stories were anchored by women and 49 stories were anchored by both. On some days stations maintained a single female anchor. Since our hypotheses addressed gender differences in types of stories reported, 131 single anchored stories were not included when interpreting results for gender of reporters and anchors. One story was not included where recording equipment difficulty made it impossible to determine gender of reporter.

Table 1 Types of stories within the newscasts.
Full size table

Coding Procedures

Each story in the newscast was coded for the following variables: order in the newscast, gender of reporter, gender of anchor, number of sources interviewed, gender, source expertise (expert or nonexpert) and story topic. Coders were two graduate students blind to the study hypotheses who were compensated for their work.

Order in the newscast was used as a measure of importance within the newscast. Each commercial break determined a new block of time within the newscast. Therefore, block one was a news story appearing before commercial break one. Block two was a news story appearing after commercial break one but before commercial break two. Block three referred to a news story appearing after commercial break two and before commercial break three. Block four was the news stories appearing after commercial break three.

Sources within the study were coded as expert or nonexpert. We used Liebler and Smith’s (1997) operational definition of source, defined as “any person in the report who provided an interview or sound bite and who was identified either through fonted or verbal identification” (p. 61). Cann’s (2001) definition of source expertise was used: An expert is described as “the source’s occupation was given and appeared relevant to the story” and nonexpert as “the source is interviewed as a bystander, witness, relative or victim or otherwise signified as a general public response”

Story topic was analyzed by two coders using a coding system similar to Cann’s (2001). For the purposes of this study the following 20 topics were included: Disaster, Traffic Accident, Accident, International, Politics, Crime, Judicial, Economics, Environmental, Technology, Industry, Human Interest, Social Welfare, Weather, Health, Arts, Military, Male Sport, Mixed Sport and Female Sport. The categories excluded from coding included: daily weather reports, traffic reports, sports scores, and statistics. Following Cann’s (2001) rationale they were excluded because they are not “news” in the formal sense, and there would be no expectation of a gender bias in their assignment to reporters. When a story appeared to fit more than one category the dominant topic of the story was used to assign it.

In order to determine reliability, twenty-five percent (2.5 hours) of the sample was used to calculate percentage of agreement between raters. The intercoder percent of agreement was 83% for overall coding of the story topics within the newscast. The intercoder percent of agreement for expertise was 94%. The source of the expertise coding error was due to vagueness of attribution of a nonexpert and expert sources within the newscast. The percentage of agreement was considered adequate and therefore the remaining 75% (8.5 h of news) were divided between the two coders and coded individually. While these percentages are relatively high it is possible that reliability is overestimated using this measure. The high percentage of agreement in the sample of attributions, however, for two raters reinforced our decision to use this measure.

Results

The first hypothesis predicted a significant association between gender of reporter and type of story reported. Chi-square analysis revealed a significant association between gender of reporter and type of story he or she was assigned, χ2 (19, N = 147) = 82.53, p < .001 (Table 2). Category frequencies revealed that females reported certain topics more often than males, specifically human interest and health stories. Males reported more political stories than did females. The first hypothesis was supported.

Table 2 Reference to gender of reporter and type of story in the newscast.
Full size table

The second hypothesis predicted an association between gender of anchor and the type of story he or she reported. No significant association was found between gender of anchor and type of story reported. While Chi-square showed a significant association, χ2 (38, N = 448) = 96.45, p < .025 (Table 3) examination of the categories reveals that most topics were almost equally shared among between males and females. Because of the low cell frequencies for the third category for gender (“both” when both a male and a female anchor reported the same story) cell frequencies for that value were too low in comparison with “male” vs. “female” to support any inferences regarding gender differences in story topics.

Table 3 Reference to gender of anchor and type of story newscast.
Full size table

The third hypothesis predicted that men would be used as expert sources within the news stories more frequently than would women. Because this hypothesis predicted a difference between two groups, t-tests were used for comparison. A t-test revealed a significant difference in frequency between male and female experts used in newscasts, t (579) = 4.72, p < .005. Male expert sources (M = .20, SD = .117) were cited significantly more often than female expert sources (M = .11, SD = .62) in the news stories, supporting hypothesis 3.

Hypothesis 4 predicted that men would also be used as nonexpert sources in local newscast more than would women. There were no significant differences between male nonexperts and female nonexperts cited t (579) = .504.

The fifth hypothesis predicted that stories reported by men would outnumber stories reported by females in the early segments of local newscasts. Chi square revealed no significant relationship; χ2 (4, N = 148) = 7.67, p > .10. There were no significant gender differences in stories placed early in newscasts; the hypothesis was not supported.

Similarly the sixth hypothesis predicted that male news anchors would outnumber female anchors reporting stories in early segments of the newscast. Chi-square showed no significant association χ2 = (9, N = 448) = 7.67, p = .496. Frequencies showed that anchor-reported stories were almost equally divided between male and female anchors in all time blocks offering no support for Hypothesis 6.

Discussion

The purpose of this content analysis was to determine whether gender bias exists in local TV news in terms story topics reported, placement of stories in newscasts and sources cited in the news. These results reveal that the gender of reporter makes a difference in the topics of stories reported by men and women. We also found that when experts are cited in news stories those experts are more likely to be males.

This content analysis revealed that women reporters reported “softer” news stories such as health and human interest while men reported politics. Like Cann’s (2001) analysis, our data revealed gender related differences in topics of news stories reported. Females reported all six health stories in two weeks, as well as 12 of 18 human interest stories. Cann also found that men were over-represented as anchors, reporters, and expert sources in Australian news. Female reporters dominated in only “low frequency, lower-ranked subjects.”( p.162) . Like Cann, we found that males predominated as expert sources. Our results for local news extend Cann’s conclusions to local market formats in the U.S. Similar to our findings Liebler and Smith (1997) found that male sources were used more often than female sources and were more likely to be shown in a professional capacity, regardless of the story topic. Again, we have extended their conclusions from national network news to the local market.

We did not find a gender trend in the use of nonexperts as sources in news stories, which we expected in light of Cann’s (2001) results for the dominance of males as expert sources. Perhaps the reason for gender equality in “people on the street” interviews is that typically, news producers do not select nonexpert sources, reporters and production assistants do. In their rush to get news footage, it is likely that the exigencies of who is available to talk to reporters override any possibility of gender bias. We also did not find any primacy effects of gender of either reporters or anchors in reporting lead stories early in the newscast as did Cann. There is evidence from recent surveys of television station managers and news consultants that the perceived credibility of female anchors and reporters among news audiences has increased in the past five years, and the sheer number of females on-air has also increased (Diaz 2009). These perceived gains in female anchor and reporter credibility and increased numbers may offer a possible explanation for our failure to find gender differences in the reporting of lead stories. Lead stories are now more likely to be assigned without regard to gender of anchors and reporters.

A limitation of these conclusions lies in the methods used in this investigation; because this was a content analysis, we were not present for each story assignment. There is a possibility that women chose, or negotiated for, favored topics to report on. While this question is relevant, future research should address it with other methods. It is also possible that editors assign certain types of stories to female reporters. We spoke with one news director who said that he always assigned stories based on who was be best qualified to write it. He also said that assignments were discussed, but that the final decision was always his. In light of the fact that all news directors in this market were male, we believe that gender bias in story assignment is plausible.

Future research is needed to determine how the process of story assignment operates. Do female editors assign stories differently to men and women? Is the gender inequality of expert sources in evidence when female reporters write and edit news stories? In light of the importance of local news across the U.S., finer-grained analysis of the construction of news will become more necessary.

While this study found that male expert sources dominated in the local news stories, female sources were not entirely absent. We also found that lead stories were not determined by the gender of anchors or reporters, while Cann did. He reasoned that male reporters were given the lead story because it is the strongest most newsworthy item within the newscast (p.169). In our results females reported the nearly the same number of lead stories as did males in all time periods measured. As more women assume the position of news anchor, research is needed to shed light on this and many other aspects of the news gender relationship. In our results, when only one anchor was shown in the newscast, in every instance it was a female. Our results suggest that television news has evolved to include a more equal gender balance among anchors, reporters and experts, but inequalities still exist.

In terms of our theoretical rationale, we offer some evidence that a mainstreaming effect in perceptions of newscasts has the potential to lead female viewers to accept certain patterns of stories assigned to either men versus women, and accept these differences as “normal” If women see that they will be assigned “feminine” topics, they will come to accept these limitations as part of the job. With respect to identification, it will be difficult for a female viewer who contemplates a TV news career to see herself as an expert in, for example, the business beat. If she is interested in business, then non-visibility of women presenting business stories or cited as experts are real blocks to identification. The loss of any potential expertise in journalism is a genuine detriment to the profession and to the viewers.