WOMEN, MEN AND NEWS

In the twenty-teens, there are increasing numbers of women occupying executive positions in politics, business and the law but their words and actions rarely make the front page. In this article, we draw on data collected as part of the 2015 Global Media Monitoring Project (GMMP) and focus on England, Scotland, Wales and the Republic of Ireland. Since the first GMMP in 1995, there has been a slow but steady rise in the proportion of women who feature, report or present the news (now at 24 per cent), but that increase is a mere 7 per cent over 20 years. Not only is there a problem with visibility but our data also suggest that when women are present, their contributions are often confined to the realm of the private as they speak as citizens rather than experts and in stories about health but not politics. Just over a third of the media professionals we coded were women and older women are almost entirely missing from the media scene. Citizens and democracy more generally are poorly served by a news media which privileges men’s voices, actions and views over the other 51 per cent of the population: we surely deserve better.


Introduction
Since at least the 1970s, when the first studies of the relationship between gender and news were published, the under-representation of women as both news subjects and news producers has been regularly noted (Chambers, Steiner, and Fleming 2004;Tuchman, Daniels, and Benet 1978;Van Zoonen 1998). Although more women than men graduate from media and journalism programmes, although women enter media professions at more or less the same rate as men, they do not go as far or as fast or take up the same beats as men and leave the industry earlier (North 2009;Robinson 2005). What this means is that what passes for news is mostly defined and produced by men and is mostly about men and men's actions: where women do intervene in the news agenda is in their roles as wives, mothers and victims, and occasionally as politicians and professionals. Whilst any number of reasons have been posited to account for the enduring marginalisation of women across the news landscape, especially by media managers, including the impossibility of a successful work-life balance and women's disinterest in senior appointments, many have also indicated the particularities of local conditions and circumstances. The Global Media Monitoring Project (GMMP), on which this paper draws, is an international study of news media which monitors how women and men feature in, report and present the news around the world. Since the first report was published in 1995, GMMP has produced a one-day snapshot study every five years, the idea for which employ more women than men overall (Byerly 2013, 12)-what does remain consistent is the manifestation of "male superiority", even in situations where women have been promoted into decision-making roles (Byerly 2013, 18).
When we consider news content, research has documented the ways in which maledefined news selection criteria favour topics which privilege male voices and reach out to sources whose status position also favours men (Ross and Carter 2011). The persistence of this silencing of women's voices across time and space is indicated not only in the GMMP data upon which we draw in this paper, but also in two other recent studies with significantly different methodologies and scope. In the run-up to the local (and, in Scotland and Wales, national) elections in May 2016, the Women's Equality Party conducted snapshot analyses of photographs used on the business, politics and sports Web pages of seven major news outlets (Sky News, BBC, The Sun, The Mirror, ITV, The Guardian, The Telegraph) over four days in April. It found that women made up fewer than 10 per cent of the people pictured. 1 On a far larger scale, and using an innovative data-scraping methodology to analyse more than 2 million news articles from 950 online English-language news sites across six months in 2014-2015, Jia et al. (2016) found that men outnumbered women in both images and text. The findings presented in this paper-from data gathered from England, Scotland, Wales and the Republic of Ireland for GMMP 2015-contribute to, and confirm, these patterns, drawing comparisons with the Europe-wide GMMP analysis (WACC 2015a), whilst also identifying some intriguing national variations within and across the various nations. It is to this study that we now turn our attention.

GMMP 2015
The Who Makes the News? report of the 2015 GMMP (Macharia 2015) was launched on 23 November 2015 during a series of events around the world including at the United Nations and at a seminar, "Is News Failing Women?", hosted by the BBC in London as part of their "100 Women" season (Greenslade 2015). Capturing data from 114 countries, the Project monitored 22,136 stories published, broadcast or tweeted by 2030 distinct media outlets, written or presented by 26,010 journalists and included 45,402 people who were subjects or quoted in those stories. The report notes some improvements over the past 20 years: 24 per cent of people who feature in the news are women, which represents an increase of 7 per cent since 1995 although this percentage has remained static since 2010. As in previous studies, the biggest gender gap is in political news reporting where only 16 per cent of people who feature are women, a decrease of 3 per cent on 2010. The smallest gaps are in the areas of science and health, where 35 per cent of those who feature are women, although this is also the smallest category of news, occupying just 8 per cent of the total news agenda. Women are thus doubly marginalised, both in volume terms and in news category segregation, featuring more frequently in stories which are seen as less important or prestigious in news value terms (see Gans 1979).

Methodological Approach
The key research question of the GMMP is: how are women and men represented in news media, as subjects and producers of news? Aligned to this fundamental question, the the extent to which they function as journalists and anchors, as reporters and presenters. The data-gathering instruments for the 2015 exercise were developed by the core team within GMMP and comprised a set of coding sheets for each type of medium sampled (TV, radio, newspapers, online news sites and Twitter), together with a code book. 2 Each country was given a media density score on the basis of which national teams were asked to collect data from a specified number of each type of media, with a preference for media which had the largest reach and audience share. The density score was determined by the number of media, the reach and the population in each partipating country. The inclusion of online news sites and Twitter feeds was at the discretion of national teams. For our four-nation study, only the Twitter feeds of traditional news organisations (e.g. BBC, Daily Mail, The Guardian and Independent.ie) had sufficient daily volume to be monitored according to GMMP criteria. Because of the danger of over-reporting where we monitored two platforms (online and offline) for the same news outlet, we have included basic statistical data for all platforms but the focus of the substantive analysis which follows is on data for TV, radio and newspapers only. For our study, we looked at 76 media comprising 16 TV and 13 radio programmes, 22 newspapers, 11 Twitter news feeds and 14 online news sites. We coded 979 stories (672 stories across TV/radio/newspapers; 307 across online and Twitter), 1960 sources and 431 announcers and reporters. We coded the top stories in each medium, with a minimum of 10 for newspapers and a maximum of 15 for all other media. Appendix A lists the media monitored in England, Scotland, Wales and the Republic of Ireland, while Appendix B lists all the coders involved in the study, whose contributions we are pleased to acknowledge here.
In addition to the collection of quantitative data, national teams were also asked to undertake some qualitative analysis and the Gender and Media (GEM) classification system developed by Gender Links for the Southern African Gender and Media Baseline Study was suggested as a useful framework which categorised stories as: (1) blatant stereotype; (2) subtle stereotype; (3) missed opportunity/gender-blind; or (4) gender-aware. Although such categories are subjective and their discursive boundaries rather porous, we identified four newspaper stories for qualitiative analysis which fit broadly into the first three categories and are discussed in more detail below.

Quantitative Findings: A Day in the Gendered Life of the News
There is always a danger with a methodology which samples a single newsday that, for completely unforeseen reasons, the newsday turns out to be atypical. On the GMMP monitoring day, 25 March 2015, the Germanwings Airbus 320 plane crash accounted for around 16 per cent of all news stories we coded. However, this tragic event aside, there were a sufficient number of other stories reported on this day to give us confidence that what we discuss below does describe a "regular" newsday. Although there are some variations in terms of topic across the five media types we monitored, there is also a large degree of consistency, suggesting that there is a shared understanding of what counts as news. With the exception of Table 1, the analysis focuses on the legacy media of TV, radio and newspapers only, and we include details of the online and Twitter stories in Table 1 only to show the broad shape of the offline and digital media landscapes.
Given the importance and volume of political news stories, we were interested to investigate whether the number of women occupying seats in the different Parliaments/ Assemblies was related to their media presence. One change to the political landscape WOMEN, MEN AND NEWS 827 since the 2010 GMMP report is that Scotland now has a female First Minister (since November 2014), so we were particularly interested to see if this was influential, but the data turned out to be rather ambiguous ( Table 2).
Wales fared best in terms of the overall presence of women in news about politics and government. As the Welsh Assembly is the legislature with the highest percentage of female members across the four nations, this is not entirely surprising. However, this does not explain why Scotland-which is not far behind Wales in this regard and has a female First Minister-does comparatively poorly. That a controversial interview with the former (male) First Minister of Scotland was widely reported on monitoring day may offer a partial explanation. Even so, the enduring use of male news sources in stories on politics and government and the notable lack of stories in this category in which women are a central focus in all parts of the United Kingdom, suggest that the news media are mostly lagging behind the (still fairly glacial) pace of change in the presence of women  in governing structures. These findings indicate little change, therefore, since Ross' (2007) study which showed that men are more than twice as likely as women to be quoted as sources. The exception to this otherwise gloomy scenario is the Irish data which reflect the nation's general performance in relation to women's news visibility across the fournation data-set, with the highest overall presence of women in all news categories (at 33 per cent) despite having the lowest level of representation of women in political decision-marking structures. This may be partly attributable to the high number of women newsreaders and current affairs show hosts in RTÉ as well as to the fact that the commercial station TV3 is heavily dominated by "intimized" (Van Zoonen 1994) news and factual programmes featuring female presenters (see Ross and Padovani 2016). It may also be partly due to the fact that the Irish national broadcaster, RTÉ, has recently engaged in training initiatives for small groups of women in an effort to promote increased participation. Although the BBC has undertaken similar initiatives across the United Kingdom (in 2013), the lack of sustained investment and follow-through in this initiative could be contributing to the UK figures, although it is possble that a new phase of training will be launched in 2016. 3

Women as Sources in News
Of the 1960 sources coded across the five media, 28 per cent were women (31 per cent across newspapers, 24 per cent on radio, 29 per cent on TV, and 25 per cent for both internet and Twitter). Half of all sources were included in stories of national significance, however, women fared less well in these stories (a quarter of sources) than in local and international news (around one-third). We can see from Table 3 that women's visibility is differentiated in terms of main topic type across media formats, clustering around the broad categories of science/health and crime/violence which are the categories of news which had the lowest frequency.
Comparing the 2015 data with 2010 (Table 3), the number of women sources as a proportion of all has decreased by 3 per cent, even though the total number of sources has increased by around 5 per cent. This finding is even worse than the European data, which shows a 1 per cent decrease in women as sources for print, TV and radio, despite their increasing prominence in political and public life. The most significant decrease is evident across broadcast media-radio is down 7 per cent and TV is down 6 per cent. There are also significant shifts in the frequency with which women appear across the range of stories. In two summary categories (social/legal and celebrity/arts), there has been a signficant increase in the number of sources overall but a reduction in the number of women sourced. Conversely, the number of sources in the crime and violence category has decreased considerably, whereas women's visibility has gone up, unlike their frequency as sources in politics and government stories, which has gone down over the past five years. These findings are not unique to our four nations or to Europe but our data do suggest a rather troubling narrowing of the news agenda and what passes for newsworthy stories, so that the experiences and voices of women are simply given less attention. As Carter (2011, 1150) have commented in relation to the 2010 GMMP study, "Feminist scholars who have examined journalist-source relationships have argued that journalists tend to rely on a narrow range of sources, most of whom are white, middle-class and middle-aged professional males". This is particularly true of sources whose views are solicited in order to yield expert opinions (Armstrong 2004;Ross 2007Ross , 2010.

WOMEN, MEN AND NEWS 829
So far, we have looked at the summary news topics, but we can also drill down to the occupations of individual sources which were composited to produce the summary categories shown in Table 4.
There were five occupational categories where women comprised more than 50 per cent of sources, and nine in which they comprised less than 20 per cent of particular occupational groups, as shown in Table 4. In addition, women comprised 22 per cent of the combined occupational group "government, politician, minister, spokesperson" (404 individual sources) which was the largest occupational group identified, comprising 27 per cent of all sources. This finding echoes that of many other studies, demonstrating the media's propensity to use "official" sources, a strategy which compounds women's marginalisation. The other two most popular source occupations for women were "celebrity, artist, actor", constituting 8 per cent of all source occupations and where women comprised 32 per cent of that category, followed by "business person, executive, manager, stockbroker" (6 per cent of all) where women constituted 13 per cent of such occupations. Women are significantly under-represented in the authoritative, professional and elite occupational categories and, conversely, are significantly over-represented as voices of the general public (homemaker, parent, student, child) and in the occupational groups most associated with "women's work" such as health, social and childcare worker, office or service industry worker. These findings are consistent with the 2010 GMMP study, which showed that women were significantly more likely to be identified as health or social workers, as teachers, activists and office workers than men (Ross and Carter 2011).
Having noted the occupation of news sources, we also looked at the function they perform in stories. Table 5 shows even more clearly the ways in which women's voices as experts or spokespeople are considerably under-valued.
Women were quoted as experts in only 20 per cent of stories which mirrors the global data where women comprise 19 per cent of all experts appearing across all types of news stories. There were some interesting differences in terms of medium, with radio and internet news including expert women at nearly three times the rate of TV and newspapers.
But generally, women are mostly relied upon to provide popular opinion or personal experience and constituted 80 per cent of such sources in the Republic of Ireland, 45 per cent in Wales, 43 per cent in Scotland and 38 per cent in England. As Ross and Carter (2011) have pointed out elsewhere, these findings suggest that women's voices are used to provide personal testimony and anecdote rather than authoritative or expert perspectives, thus reinforcing professional-domestic and public-private dichotomies. Men were also much more likely to be quoted (72 per cent) in stories where they were the subject, than women (59 per cent), further adding to women's silencing. On the other hand, stories about women were more likely to feature an accompanying photograph (34 per cent) than stories about men (23 per cent) which is a pattern replicated in the global data (Macharia 2015, 45;also Jia et al. 2016). In general terms, given both the frequency of and credence given to expert sources and spokespeople, once again we see that women's views on stories in which they are not in some way involved is marginal, further confining them  to the sphere of the private, emotional and subjective, with men continuing to dominate the sphere of the public, rational and objective.
Comparisons with the 2010 data show that although there were more sources quoted in 2015, women's visibility has actually decreased, although the spread of women's voices is broadly similar across the two sample years. Our data are broadly in agreement with both the global and European data in this regard, with the exception of the proportion of women in the popular opinion category (54 per cent). This is likely bolstered by the exceptionally high percentage for the Irish data. The figures for the Republic of Ireland for eyewitness (62 per cent), personal experience (76 per cent) and popular opinion (80 per cent) are also extremely high, perhaps indicating that the tendency to confine women to the personal/domestic/ private realm is particularly exaggerated in the Irish media.
We also considered the age of sources and coded for age where it was either explicitly mentioned in newspaper or TV stories or where the coders were able to ascertain age. Although this is obviously not an exact science, as part of the testing phase with coders, we tested this aspect for consistency and determined a high degree of intercoder reliability. Broadly in line with the global data, our findings also show that there is an inverse relationship between sex, age and visibility so that as a woman's age increases, her visibility in the news decreases: a mere 20 per cent of all sources/subjects who were perceived as being over the age of 50 years were women. This is in line with the European data which show that women comprise only 19 per cent of everyone in the 50-64 year age group. It also echoes other recent research which shows the disappearance and/or stereotyping of older women in the news (Lepianka 2015;Wylie 2013). Given that the world's population continues to be an ageing one, and that women live longer than men, one irresistible conclusion to draw from this otherwise counter-intuitive result is to suggest that although women's stories may well resonate with news consumers, they seem rather less interesting to the-mainly male-journalists who decide what we see, hear and read (North 2016a).

Women as Media Professionals
In relation to who writes, announces and reports the news, women comprise 39 per cent of media professionals which is slightly higher than the global GMMP average of 37 per cent. In fact, this global statistic has not changed since GMMP 2005, indicating that female reporters and broadcasters appear to have hit a global glass ceiling. In Europe, the statistic is similar: women comprise 37 per cent of all reporters and are disproportionally represented in the categories of science and health (44 per cent) and celebrity, arts, media and sport (42 per cent). These findings, again, reflect much of the research already undertaken on women's employment in news media which demonstrates various kinds of horizontal segregation and gender-based beats (De Bruin and Ross 2004;O'Brien 2015).
In terms of women reporting and speaking in broadcast news, we recorded 431 individual announcers and reporters, of whom 32 per cent (136) were women, although there were wide variations between both role and medium, with radio being less inclusive of women media professionals. That women are mostly visible as announcers on TV (although their presence has gone down by 6 per cent since 2010) reflects the findings of many other studies and is associated with what has been termed the "feminisation" or "intimisation" of news, where news is fronted by an attractive, often young, woman whose friendly persona is regarded as more appropriate for delivering news-as-infotainment (Bliss 2015;Gibson 2009). There were no differences between the type of journalism women and men write in terms of local, national, sub-regional or foreign/international news, although there were differences in terms of story focus and beat. For example, the stories covered by women journalists were more or less evenly spread across the primary topic groups of science and health (15 per cent), politics and government (13 per cent), economy (13 per cent), crime and violence (12 per cent), with a larger proportion of stories on celebrity, arts, media and sport (20 per cent), and social and legal issues (23 per cent). The spread of stories was largely similar for men, but because women comprise a smaller proportion of all journalists coded, they were disproportionately bylined on stories about science and health (58 per cent) and crime and violence (52 per cent), so we could say that they are both conforming and confounding gender stereotypes around "soft" and "hard" news stories (see North 2016b).
Gender stereotypes were much in evidence in relation to age, so that women announcers and presenters were mostly coded within the age category 35-49 years (89 per cent) and only 4 per cent of women reporters were coded as 50 years or over (compared to around 33 per cent of male reporters). Male presenters were more evenly spread across the 35-49 and 50-65 years age groups (few in the youngest age category but more or less similar numbers across the other two). Again, this finding echoes research elsewhere which has noted the popular pairing of the attractive younger presenter with the avuncular older male, the latter reporting the main news stories with the female side-kick rounding off with the second-tier human-interest stories (Carter-Olsen 2012;Holland 1998). In recent years, we have seen a number of cases of women journalists and presenters taking their erstwhile employer to an industrial tribunal on grounds of both sex and age discrimination. In one of the most high-profile recent cases, in 2011 the Irish TV presenter Miriam O'Reilly won her case of unfair dismissal from the BBC's programme Countryfile on the grounds of ageism: she was 52 years old at the time and her co-presenter, John Craven, who remained on the show, was 68 years old (Martinson 2013).
Comparing the findings for 2015 with those of 2010, in the earlier study, 31 per cent of stories in the press were written by women, 36 per cent of reporters on radio and 30 per cent of reporters on TV were women. Given that in 2010 women were more present in radio than the other two media, it is interesting to question why they are so much less present now. The number of stories with an identified reporter on radio has more than halved in the past five years, from 115 to 56, which could go some way to account for the difference. Even more dramatically, the percentage of women coded as announcers on radio has also gone down over the same period from 53 per cent to 23 per cent. Given that most of the radio programmes we monitored were almost the same in both sample years, there is no obvious reason why women's voices have become less frequent on the radio airwaves, although other research has produced similar findings. In a recent report (Walsh, Suiter, and O'Connor 2015) on women's under-representation in current affairs programming on radio in the Republic of Ireland, the overall breakdown of voices (female and male) and the proportion of time each is heard over the three radio stations examined shows an overall distribution of 28 per cent women and 72 per cent men.

Qualitative Analysis
Guidelines provided by WACC (2015b) for the qualitative, textual analysis of selected news stories from the GMMP data-set provided "generalised criteria" drawn from a classification system developed by the Southern Africa Gender and Media Baseline Study (WACC 2015b). Researchers were asked to identify stories that fit the following categories:

WOMEN, MEN AND NEWS 833
(1) blatant stereotyping; (2) more subtle stereotyping; (3) missed opportunity/gender blind; or (4) gender aware; and were instructed to include analysis of sourcing, language, gender awareness or bias, and presence or absence of stereotypes. The purpose of the qualitative analysis was to provide case studies to support findings from the quantitative data and "build gender and media literacy within the public (media consumers), media advocates and among those who work within the media" (MISA and Gender Links 2003).
Using these criteria, what follows is a textual analysis of four news stories, two from The Sun, one from the Daily Express and one from the Daily Telegraph, each of which offered the most ideologically rich examples of stories with an explicitly gendered aspect. 4 Unfortunately, we did not find any stories which could be coded as "positive" examples of gender-sensitive reporting and we are being consciously selective in our four case studies to illustrate the enduring and complex nature of stereotypes. Given the relative infrequency of women's presence in the news, it is discouraging to see the continuation of gender stereotypes when they do appear.
The first story is an example of "blatant stereotype" (women as sex objects) and was headlined "Booby Wonderland" (Wilson 2015; The Sun, UK, Scotland) (Figure 1). Although the last official "Page 3" model appeared in the UK edition of The Sun on 22 January 2015, semi-naked pictures of women continue to appear in the newspaper. Most of the page was devoted to a "cracking cleavage competition" as part of the Sun's "Cleavage Week". Noting that many of the newspaper's "busty readers' had already 'taken the plunge' and entered the competition, the story is primarily composed of photographs of five women in their 20s, one aged 34 years (a "young bride" standing next to her smiling husband) and a youthful-looking 42-year-old showing off her "assets'. All are identified only by their first name, age and home town, the same formula used for the Page 3 "model" in the 44-year history of this infamous newspaper feature to construct these topless glamour models as the "girl next door". Using bits of women's body parts as a salacious end in itself as well as advertising everything from perfume to shoes, is a familiar media practice which feminist and critical media scholars have described as yet another way in which women are objectified and dehumanised (Cortese 2016;Gill 2003;Goffman 1979;Kilbourne 2010;Rocha 2013).
Our second article was the lead (of three) on page 7 of The Sun (UK, Republic of Ireland, Scotland), headlined as "I'm Fit and I'm All Woman … Kellie Recovers After Sex Op" (Wootton and Pink 2015), offering an example of more subtle stereotyping and a missed opportunity (Figure 2). It begins by referring to Kellie Maloney, "the transsexual boxing manager, 61" who has been preparing for her "sex op" with a "gruelling fitness regime" to lose four pounds and reduce her Body Mass Index by 2 per cent. A head and shoulders photo in the top corner of the article features an older man in a pinstripe blazer with a deeply lined and tired looking face, captioned as "Before … as promoter Frank". This photo is set against four featuring Kellie exercising in the gym, focusing on her body. She wears a grey, black and pink track suit, sporty headband holding back shoulder-length blonde hair, tight facial skin and makeup. The mention of a "long road to recovery" captioning one of the photos relates to Kellie's journey back to health after invasive surgery. Discursively, this "recovery" is associated with the loss of one constructed gender identity of (masculine) physicality, replaced by a new (feminine) but still physically active other. Kellie's recent participation in Celebrity Big Brother allowed her to "find herself", representing the emotional aspect of her recovery. In the post-operative period, readers are told, Kellie will redefine herself as a mother of two young daughters, a boxing manager and a transgender community campaigner.
While there are politically progressive features in the story, it fails to challenge normative assumptions around gender identity. Photos and words emphasise the recurring importance of standardised ideals of feminine beauty-being slim, youthful and sexually attractive. What remains potentially subversive, however, is that this ideal is thrown into sharp relief precisely because it is represented as a construct that may be obtained through hard physical and emotional work and, ultimately, surgery. Significantly, however, the accompanying pieces by an "equality activist" and the "Sun doctor" both address an imagined male reader by appealing to the putative emotional response of the "average bloke" (naturally masculine) to "having your todger sliced up and made into a vagina" as being one of abject terror and the loss of masculine privilege.
The third story, found on page 9 of the Daily Express, is another example of subtle stereotyping headlined, "The Petite Armed Guards Whose Guns Were Too Big" (Twomey 2015) (Figure 3). The story concerns a Central London Employment Tribunal hearing involving two female firearms officers, Victoria Wheatley and Rachel Giles, who argued that they faced humiliation from colleagues upon returning to work after winning their sex discrimination case against the Civil Nuclear Constabulary for whom they worked as security guards at nuclear sites. The hearing considered the claim that the women's low shooting scores and resulting dismissal from their jobs was attributable to having been issued firearms which were too heavy for them to handle. Both had asked for smaller weapons but their requests were ignored. On returning to work, they were given menial jobs which left them feeling insulted and disparaged. As a result, both were signed off on extended stress-related sick leave. Whilst at first glance it would appear that the story is broadly sympathetic to the women, describing their size as "petite" and the tools of the job too big for their small hands discursively positions a "big man" norm against which they are judged as inadequate. Discussion of their stress-related sick leave again characterised them as unsuited to the job of armed guard which, by its very nature, is a stressful occupation. Overall, the story constructs the women as incapable (too small to do a big man's job, too emotional, too easily stressed), petulant (refused to retake firearms tests with smaller weapons) and weak (lacking in the toughness required for the job). When men are regarded as the default standard for any occupation, "women are discouraged from adapting the environment in ways that are most suitable for them [and are often] afraid to ask for special treatment such as smaller tools' (Messing 1998, 24), although in this case, they did ask but were ignored. In the male-dominated workplace, differences between colleagues have usually been resolved privately, so calling public attention to blatant discrimination not once but twice, constructs the women as outsiders to security guard "culture". In addition, both women are referred to as "Ms' which could be regarded as progressive since they are not identified by their marital status. However, this form of address is also employed to slyly identify a woman as feminist and therefore potentially unreasonable in making demands (of men).
Our final article, from page 13 of the Daily Telegraph, is headlined "PM Meets His Match as He is Heckled by Elderly Over NHS" (Holehouse 2015), describing (former) Prime Minister (PM) David Cameron's speech at a conference organised by Age UK, an advocacy NGO run by and for older people (Figure 4). The strapline "Pre-election event turns into rabble of protest with pensioners unhappy about exodus of nurses" and headline together discursively construct older people as an homogenous group, connoting simultaneously that they are physically decrepit ("elderly"), on low incomes and therefore needy ("pensioners") and irrational ("rabble") in their zealous challenge to the PM ("jeering" and "heckling") to provide more National Health Service (NHS) nurses. There are two photographs, one of an open-mouthed, wide-eyed, clearly empassioned older woman holding a microphone asking a question and the other from the back of the room, so only the backs of people's heads are visible, although a person holding up a walking stick is clearly a man (wearing a suit jacket). The front-facing woman is not named but in the body of the story an older man is named, his previous occupation given and he is quoted at four separate points. This is a good example of a gender-blind and missed opportunity story, given that women live longer than men and thus are likely to use the NHS more frequently and have therefore even more interest in the continuation of a good service. Why is the woman not named? As she clearly asked a question, why use her picture and then not interview her as happened with the male source? Feminist news research has consistently demonstrated that women are seen rather more than they are heard (Jia et al. 2016), tending to be valued as visual pleasure over expression of their views (Holland 1998). As Carter (2011, 1150) have argued, "by privileging issues seen to be in both the interest and purview of men and privileging their views and WOMEN, MEN AND NEWS 837 voices over those of women, news discourse contributes to the ongoing marginalization of women's participation as citizens".
These four examples illustrate the normalised persistence of stereotypical representational practices in some elements of mainstream news media as well as alerting us to the subtle ways in which certain stories and topics are pitched to an implied male audience.

Conclusion
The results reported in this paper demonstrate that the relationship of women to media continues to be complex and complicated, sometimes showing advances over time and sometimes a retrenchment, identifying differences in women's experiences of news media between genre and geography, between medium and mode. While our data largely echo the European and global GMMP trends in demonstrating that the presence of women in news media overall has increased since the first GMMP in 1995, our study suggests that women are now less present in certain categories of news (e.g. politics) and in certain media (e.g. radio) than in 2010. We therefore argue that the very modest progress which has been made is not the result of a cultural shift in news values or cultures over the past 20 years but rather describes a slow trend which is hard to interpret: the decrease in the number of female sources in stories about politics and government is a good case in point. The national variations within and across the four nations in this news category-variations which do not map neatly against the presence of women in the respective national Parliaments and Assemblies in any consistent way-suggest that, despite the increasing number of elected women politicians, the lobby still reaches for the usual (male) suspects. The slight caveat to this gloomy conclusion, however, is that women reporters were nearly twice as likely as men to write stories which had a central female focus and more likely to feature women sources more generally. This marks a significant departure from Ross' (2007) earlier study, which showed that there were no significant differences in this practice if the journalist writing the story was male or female, and gives some cause for optimism if more women are assigned the politics beat.
We can only speculate on the circumstances which have produced these data but we do not believe that, for example, women's declining presence in political news is the result of random journalistic practices, particularly when women's real-world representation has actually increased. 5 Moreover, the findings tally broadly with other large-scale studies (e.g. EIGE 2013) which indicate that women have hit a glass ceiling and are now struggling to move beyond the "one-third" rule. The mere 7 per cent improvement in women's presence in the news (as both subjects and reporters) across the global GMMP data in the past 20 years suggests something troubling in the ways in which the media consider the relative importance of women and men in society. What is revealed by the data is the extent to which there appears to be a shared understanding of what constitutes news, whose voices are important and whose actions should be represented, not just in our four nations but everywhere. That understanding seems universal and privileges men's domination in a spectacularly consistent display of hegemonic reproduction which maintains the patriarchal status quo. Studies such as the GMMP stand as public records and timely reminders of the ways in which gender stereotypes are discursively normalised, recycled and recirculated in a tightly controlled and globalised information and communication system. As an indicator of the persistence of an overwhelmingly male-ordered news focus which ignores or marginalises the lives and interests of 51 per cent of the population, the GMMP is, in itself, a form of media activism and a rallying cry.
Gender-based advocacy for equality draws on findings such as those produced by the GMMP to provide evidence of inequalities and stereotyping, and to raise awareness and campaign for change. The Project's website 6 has numerous examples of good practice, including how findings have informed media and, indeed, government policy in relation to gender and media. How findings can be leveraged locally and nationally is a challenge not only for media organisations but also for feminist campaigners. Initiatives to broaden the pool of women experts available to be interviewed through training days have been set up by both RTÉ and the BBC. "The Women's Room", established by Caroline Criado-Perez, is an online database which encourages any woman who considers herself to be an expert to create an entry, aiming to make it easier for journalists to supplement their usual network of contacts and bring more diverse voices into the news. 7 But it is not about numbers alone. Studies like the GMMP are, of course, only snapshots and, in any case, news discourse is as open to interpretation as many other forms of media content and the polysemy of stories means that findings cannot stand as straightforward and unambiguous indicators of women or men's improving or worsening status in the world. We cannot draw too many inferences from a methodological approach which is mostly quantitative, since simply counting frequencies cannot answer questions of tone, orientation and point of view, all of which are equally if not more important in the WOMEN, MEN AND NEWS 839 process of meaning-making. This is not at all to undermine our own analysis, which is broadly supported by other national and international studies, but merely to signal that a mixed methodology combining both quantitative and qualitative elements will move us towards a more nuanced and comprehensive understanding of news phenomena. As a very modest nod in this direction, our qualitative analysis also suggests both the persistence of gendered stereotyping and some of the ways in which this is changing over time. For example, The Sun's Kellie Maloney story echoes Jon Stewart's important discussion of the US media's coverage of Caitlyn Jenner's debut in Vanity Fair 8 which suggests that transwomen are rendered both visible and intelligible in mainstream news through the exercise of routine tropes of gendered objectification.
The recent report of the House of Lords Select Committee on Communications (2015) makes clear that the British media industry is still failing women, both in terms of promotion and employment as well as in media content and sex-based stereotyping. The change we need to make in order to fundamentally shift the news agenda away from its androcentric gaze is not simply to "add women and stir" but, rather, to mount a more comprehensive challenge to the normative nature of news routines which masquerade sexism as objective practice. It is anything but. Breaking News