Volume 4 Issue 1, Article
Tome 4 numéro 1, article

Mike Wise

Abstract

A 13-week study into a newsroom source diversity initiative promoted by the BBC looked at whether it was effective in getting student journalists to reach a goal of 50:50 male-to-female sources in their reporting. Findings show widespread support among student journalists to improve source diversity in their reporting, but mixed results into the effectiveness of the BBC approach in tackling the issue of balanced gender representation in interview sources.

Keywords: source diversity, gender balance, journalists, student newsrooms, content analysis, surveys

Résumé

Une étude de 13 semaines portant sur une initiative de diversité des sources en salle de rédaction promue par la BBC a évalué son efficacité à aider les étudiants journalistes à atteindre un objectif de 50:50 de sources masculines et féminines dans leurs reportages. Les conclusions montrent un large soutien des étudiants journalistes à l’amélioration de la diversité des sources dans leurs reportages, mais des résultats mitigés quant à l’efficacité de l’approche de la BBC pour aborder la question de l’équilibre des genres dans les sources d’interviews.

Mots-clés : diversité des sources, équilibre des genres, journalistes, salles de rédaction étudiantes, analyse de contenu, enquêtes

ARTICLE

---

A case study of the BBC 50:50 Equality Project: Can the source tracking system work in Canadian journalism programs?

Mike Wise

 

INTRODUCTION

In 2016, the BBC’s Ros Atkins heard something that bothered him while driving home listening to a news program on BBC Radio 1. In almost an hour, not a single woman had been interviewed about the day’s news (Chilazi et al., 2020). The next day, he set out to make sure his own TV news program wasn’t committing that kind of oversight. As the host of Outside Sources, a daily news analysis program that aired on several BBC TV News channels, he worked with his senior producers to develop a system to track the gender balance of the interviews they booked, in hopes of increasing the representation of female voices. After each broadcast, his show’s producer would count the gender of the guests they’d just put on the air and tabulate the numbers in a spreadsheet. At the next editorial meeting, the team would discuss the results. There were no quotas set and no judgements made about the results. Instead, chase producers were always told to seek out the most appropriate and qualified guests regardless of gender. Within four months of simply sharing the ongoing performance results, the show went from featuring 39% women to reaching a target of 50%. The accomplishment and its underlying methodology quickly spread to other programs with similar results. BBC management eventually embraced the 50:50 system, promoting it from an employee-led grassroots effort to a corporate-wide initiative for the BBC’s TV, radio, online and digital services (Wittenberg-Cox, 2020). By April 2019, 500 content units were part of what was known as the BBC 50:50 Equality Project, with 74% of them hitting that gender-balanced target (Rattan et al., 2019). The BBC then shared its methodology with other broadcasters, publishers and journalism schools worldwide.

This article presents a case study of the BBC 50:50 Equality Project and describes an experiment to evaluate whether the BBC’s methodology can be effectively applied in a Canadian journalism classroom setting. The study aims to see if regularly measuring and sharing source-diversity rates can change journalism students’ source selection practices. The study employed a 13-week content analysis of stories filed to a Canadian journalism school website, measuring male, female, and non-binary sources using the BBC’s methodology (BBC, 2021). Student reporters were also surveyed about their attitudes towards gender diversity in journalism, their efforts to seek interview subjects, and their response to the source-tracking measurements.

While many studies have conducted source audits of published stories, few have looked at how reporting staff react to learning about the results on an ongoing basis. This article will explore whether the BBC approach encouraged students to balance their interview subjects or change how they sought their sources. It will look at how student performance is affected by how a course replicates deadline pressure faced by professional journalists. It also looks at whether there’s a benefit for journalism schools to use the BBC methodology over other methods of measuring source diversity.

LITERATURE REVIEW

While the BBC uses a 50:50 ratio to describe the general breakdown of men to women in the population at large, statistically, women make up a slightly larger proportion of the population than men in Canada (50.9%), the United Kingdom (51%) and the United States (50.5%) (Office for National Statistics, 2023; Statistics Canada, 2022; U.S. Census Bureau, n.d.). Whatever the exact number, surveys of media content worldwide repeatedly show a tendency for reporters to over-represent male voices at the expense of females. Since 1995, the Global Media Monitoring Project has conducted international content analysis studies of women’s overall presence in the news media every five years. Its first study revealed that only 17% of news subjects were women. Subsequent studies saw that number rise to 23% in 2005 and 24% in 2015 (Macharia, 2015). The project’s most recent study found that in 2020, women still made up just 24% of news subjects and newspaper sources, although the numbers in Canada rose by 4% since the 2015 study to 31% (Global Media Monitoring Project, 2021).

Other international studies show a consistent under-representation of women. In Europe, the share of women as news sources cited in online stories ranged from 28% in the U.K., 32.9% in Norway, 33% in Ireland, and 36% in Switzerland (Global Media Monitoring Project, 2015; O’Brien & Suiter, 2017; Sjøvaag & Pedersen, 2019; Vogler & Schwaiger, 2021). Belgian TV reports cited women as sources in just 20% of stories (Swert & Hooghe, 2008).

An ongoing study of Canadian media outlets shows similar results, with men quoted three times more frequently than women on the national newscasts on CBC, CTV and Global (Asr et al., 2021). That research comes from an ongoing monitoring project run by researchers at Simon Fraser University and the Ottawa-based advocacy group Informed Opinions. The group is trying to increase the representation of female experts in Canadian media by examining barriers preventing women from speaking freely when contacted by reporters, such as childcare responsibilities or women not wanting to seek the spotlight (McKeon, 2011). The joint study further suggests some women may fear abuse or harassment when discussing controversial topics (Asr et al., 2021).

The same study cited journalists’ deadline pressures as a possible reason they do not quote women as often as men. When it comes to choosing whom to interview, reporters may rely on resumes, credentials, and subject matter expertise (Newsome, 2021) as well as a source’s willingness to speak, timeliness in returning phone calls, and their relationship with the journalist (Meer et al., 2016). Faced with tight deadlines, “journalists tend to use sources they are comfortable with—that is, sources they get along well with and sources they have used in the past” (Martindale, 2006), and this can privilege male voices (Cukier et al., 2019; Everbach et al., 2010).

Robert Entman suspects that since many journalists strive to hold those in power accountable, a “watchdog bias” may be responsible (Entman, 2010). This may result in media outlets focusing on events and people within “legitimate institutions” such as education, finance, police, and the courts (Tuchman, 1980). Carolyn Byerly argues those leaders may still be predominantly male (Byerly, 2021). A content analysis of Canada’s 2011 Federal election found that men made up more than 80% of the clips on TV newscasts, which the authors argued was due to men being in positions of power where they would be considered experts (Barber & Levitan, 2013).

The literature suggests that many newsrooms associate traditional female gender roles with issues they consider ‘feminine’ and thus more likely to appeal to their female audience (Swert & Hooghe, 2008). Several content analysis studies find that when many women do appear in the news, they are depicted in so-called “soft” stories, dealing with traditional female topics like arts and entertainment, health and lifestyle, and family and education (Rao et al., 2021; Sjøvaag & Pedersen, 2019; Swert & Hooghe, 2008; Vogler & Schwaiger, 2021). Men, by comparison, are more likely to appear in so-called “hard” stories about politics, business and sports (Macharia, 2015; Rao et al., 2021).

There also is debate in the literature about whether the percentage of women working in a newsroom can address gender stereotyping and the under-representation of female voices. The Canadian Association of Journalists’ 2023 Newsroom Diversity survey of 273 Canadian newsrooms found 51.4% of staff identify as female, 48.3% identify as male, and just 0.3% identify as non-binary (Canadian Association of Journalists, 2023). Some studies suggest women are more inclined than their male colleagues to include female sources in their stories (Zoch & Turk, 1998) and may have more access to female voices than men (Zeldes et al., 2012). Other studies have found that due to similar organizational routines and policies in their newsrooms, there are no significant differences between female and male reporters when interviewing female sources (Liebler & Smith, 1997; Swert & Hooghe, 2008). In addition, at least two studies found little difference between male and female reporters presenting so-called “hard” or “soft” news (Craig, 2017; D’Heer et al., 2019), suggesting the gender of the journalist may have limited influence on the type of news they present. The BBC admits its methodology is about setting benchmarks and measuring performance (BBC, 2020), which differs from approaches used in academic studies. The Global Media Monitoring Project collects data as a snapshot in time, conducting worldwide content surveys on the same day. Others collect their data sets over months or even years (Swert & Hooghe, 2008). Many studies identify expert and non-expert voices and categorize stories by type and genre (Cann & Mohr, 2001; Vogler & Schwaiger, 2021), while others code for differences in format and market demographics (Humprecht & Esser, 2017). Some studies count the number of sources and their weight or importance in an article by measuring the number of lines attributed to the source (Voakes et al., 1996). Other papers performed audits of newsroom content management systems to examine how stories and information were passed through the gatekeeping process (Lyons, 2002). Natural language processing is used in some studies (including Informed Opinion’s project) to automate the real-time identification of sources (Asr et al., 2021; Fu, 2021; Sjøvaag & Pedersen, 2019) by comparing them to a database of names or previously identified interviewees. Most academic content studies are done retroactively, with researchers pouring through data and then presenting their findings (Everbach et al., 2010).

The BBC is not the only media organization that tracks source diversity. Both TVOntario (Graydon, 2017) and the CBC have staff measuring content diversity rates and how sources are used (Fenlon, 2020). The Toronto Star and The Globe and Mail participated in pilot projects in 2019 to increase the gender diversity of their sources (Cabrera et al., 2020). The Canadian Press and the Chicago Sun-Times both rely on local journalism students to perform content audits of their published material (Carleton University, 2022; Lu, 2021). These approaches share a common weakness: they rely on researchers, producers, or designated staff to identify and classify interview subjects. This may introduce inaccuracies or misidentifications as those recording the data may not be directly involved in the reporting. To mitigate this, NPR encourages its journalists to directly inquire about the gender and demographic identities of their interviewees, which is then fed into an automated system for tabulation (Fu, 2021). Other outlets ask subjects to enter this information directly into Google Forms after being interviewed (Davenport & Grimm, 2021). Both strategies are aimed at ensuring the accuracy of demographic data.

The BBC’s methodology also differs from other measurement systems because it does not count every interview subject. It excludes sources deemed essential to a story’s narrative, referring to them as “players” or “newsmakers.” It instructs journalists to not count a politician making a policy announcement if that politician is an essential part of the story and the story can’t be told without them. Instead, the focus should be on counting the sources the reporter has more editorial discretion over, such as people who react to the politician’s announcement. Likewise, a significant eyewitness to an accident would not be counted if their voice is essential for telling the story (BBC, 2021). The goal is to correct journalists’ reliance on routine newsgathering channels, which can predispose them to news from official sources like government proceedings, news releases, press conferences, and non-spontaneous events like ceremonies and speeches (Shoemaker & Vos, 2009, p. 54). However, this can also erase those voices from the overall analysis and potentially offset “the gender imbalance effects of journalism’s attraction to power” where, generally, men hold power in society (Sjøvaag & Pedersen, 2019). It is also to guard against producers trying to influence the measurement data by assigning stories featuring more male or female voices. This additional layer of subjective judgement may also result in intercoder reliability issues, as producers may have differing opinions on whether a particular subject should be included or not.

Finally, when assessing how well this can work in a journalism school setting, there is limited research comparing the differences in decision-making between professional newsrooms and journalism classrooms (Farquhar & Carey, 2019). Professional journalists have the advantage of well-connected networks of official sources (Zoch & Turk, 1998), something student journalists have not had the time to develop. Students’ openness to seeking more diverse sources may result from being unable to rely on the same established names repeatedly (Haney & Paskey, 2020). While student publications can try to replicate the dynamic of a working newspaper, few can match the same deadline pressures (Josephi, 1999, p. 79). Larger papers have more resources and may foster competition among reporters to publish the best-sourced articles: smaller papers rarely discard news stories (Lyons, 2002). Classrooms must balance theoretical and ethical instruction with practical skills demanded by industry, such as the ability to write and work under pressure and tight deadlines (Skinner et al., 2001; Wenger et al., 2018). However, source diversity tracking reported at the University of Minnesota, Concordia University, and Michigan State University suggests students may be closer than their professional colleagues in reaching a 50:50 gender balance (Cabrera et al., 2020; Content Diversity Board, 2020; Davenport & Grimm, 2021).

METHODS

During the 2021-22 school year, I set out to study the effectiveness of using the BBC 50:50 methodology in my journalism program. With this study, I wanted to explore the following questions:

RQ1: Did creating awareness of ongoing source diversity measurements change the ratio of male to female and non-binary sources in student articles?

RQ2: Do students change their methods for finding interview sources when they’re aware of ongoing source-diversity tracking measurements?

RQ3: To what degree do the results differ when simply counting all sources versus using the BBC methodology?

The classroom experiment included both a quantitative content analysis and a qualitative survey of student responses to evaluate the effectiveness of using the BBC methodology. The content analysis looked at 13 weeks’ worth of student journalism articles published on my college news site from January to April 2022. These stories were written as part of class assignments by students in the second and third year of a journalism diploma program and by post-graduate students in a graduate certificate program. Of the three grade levels, the third-year course most closely replicated traditional reporting deadlines: students were expected to pitch a story or receive an assignment in the morning and file their story in the afternoon. As one of three professors attached to that course, it was the only one where I directly interacted with students.

The content analysis used a coding scheme based on the BBC 50:50 methodology, which meant I excluded sources that I defined as “players” or “newsmakers” (although I did count those sources separately). As the sole person responsible for recording the gender identity of sources, I used social media accounts and professional biographies to inform my decisions when someone’s gender identity wasn’t apparent in the story’s context. This followed the BBC’s advice to its producers to use publicly available data whenever possible to help classify sources. Measurements were done every Saturday morning and shared with students at the start of each week. This approach allowed me to analyze whether any change in the ratio of male/female/non-binary sources correlated to a feedback loop caused by students being made aware of the measurements. While I did not track the performance of individual students, I did track the reporter’s gender to see if it played a role in the results.

I also conducted two qualitative surveys of student reporters using SurveyMonkey. They consisted of Likert-scale questions with sections devoted to their awareness of gender balance in the media they consumed, as well as questions about their reporting routines. An initial survey of 63 students was done during the fall semester, before they were aware I would be tracking the gender identity of their sources, and before they were informed how the BBC methodology worked. It was also before all the students were reporting on a regular basis. A second survey of 63 students was administered in the winter term, and after all three grade levels had been reporting for a few weeks and were being made aware of the ongoing measurements. This second survey included additional questions about the effectiveness of the source tracking system. Students were not required to participate in the surveys. Since no personally identifiable information was collected, I could not establish a direct correlation between individual student results nor determine whether the same students participated in both surveys. Additionally, at least two students left the program between the two surveys, which may have resulted in a slight difference in the composition of the respondent groups. This limitation should be considered when interpreting the survey results. The research received ethics and institutional research approvals.

RESULTS

After analyzing 492 published stories and counting 1,266 sources, 603 were identified as female, 653 were identified as male, and 10 were identified as non-binary. This works out to 47.6% female, 51.6% male and .08% non-binary. I also broke down the content analysis by grade level (Table 1).

Table 1

Content Analysis by Grade Level

Grade	Female sources	Male Sources	Non-binary sources
2nd year	45.1%	53.3%	1.59%
3rd year	46.7%	52.6%	0.66%
Post Grad	51.6%	48.1%	0.29%

Regarding the gender of reporters, there were more female reporters than male reporters in this sample and none identified as non-binary. Female reporters filed 56.91% of the stories in my content analysis, while male reporters filed 43.09%. As Table 2 shows, female reporters were considerably more likely to include female sources in their stories than their male counterparts.

Table 2

Content Analysis by Student Gender

Reporter Gender	Story percentage	Female Sources	Male Sources	Non-Binary sources
Female (n=39)	56.91%	51.9%	47.2%	.09%
Male (n=27)	43.09%	39.1%	60.3%	0.6%

The analysis did not track story type or article length.

The qualitative student surveys found overwhelming support for the idea that a media outlet’s reporting should strive to achieve gender balance and diversity of sources. 82.4% of respondents felt it was “important” or “extremely important” for media outlets to reflect an overall 50:50 balance in their reporting. However, when it came to seeing that reflected in the media they consume, just 38.2% thought general news stories they read were gender balanced. Students thought arts or entertainment stories (53.3%) did the best job at providing an equal balance of male-to-female voices, while business stories (13.3%) did the worst.

As for the effectiveness of the BBC’s 50:50 methodology, the results were mixed. 73.5% of respondents said they understood how the content measurement system worked (this question was only asked in the second survey); 91.2% understood they were to pursue the best interviews they could, regardless of gender or background, and a majority thought the overall website was achieving gender balance. However, despite the results being publicized every week on our learning management system, just 41.2% said they were aware of the weekly measurements (awareness was highest among third-year students).

Students were also asked about their techniques for finding interview sources. Only a slight change was found between the first and second surveys. Table 3 shows the percentage of students answering “likely,” or “very likely” in response to the options for researching potential interview subjects.

Table 3

Selected Survey Questions Related to Finding Sources

Question	First survey	Second survey
Check existing stories by other media outlets for potential interviews	82.4%	88.2%
Google names for potential interviews	91.2%	94.1%
Check Twitter for potential interviews	67.6%	75.3%
Check Facebook/Instagram for potential interviews	58.8%	61.8%
Check LinkedIn for potential names	52.9%	61.8%
Call someone and ask for help in identifying potential interviews	52.9%	61.8%
Check with a professor for advice in identifying potential interviews	64.7%	73.5%
Check with a classmate for advice in identifying potential interviews	79.4%	64.7%
Consult an online expert guide to identify potential interviews	52.9%	70.6%

 

While my initial findings were calculated using the BBC’s methodology, which excludes “players” or “newsmakers,” I also tabulated those categories separately to see how the results would shift if they were counted. Including all sources added 21 female and 62 male sources and changed the ratio (see Table 4). This had a statistically significant altering of the results, with a decrease in the percentage of female sources and a corresponding increase in male voices.

Table 4

Comparison of BBC Results With All Sources Counted.

	Total sources	Female percent	Male Percent	Non-binary percent
BBC Methodology	1,266	47.6%	51.6%	0.8%
All sources (including “players” or “newsmakers”)	1,349	46.3%	53%	0.74%

DISCUSSION

This paper evaluates the experimental use of the BBC’s source diversity measurement system in a journalism school setting. It aimed to see if the process of regularly measuring and sharing source-diversity rates could influence student behaviour. Over the course of the study’s analysis, the sample of students came close but did not reach the BBC’s goal of 50:50 gender representation. The percentage of women quoted surpassed the 50% mark in just four of the 13 weeks analyzed here. However, the overall results produced by this study (47.6% women, 51.6% male, 0.8% non-binary) were still considerably higher than the 31% Canadian industry average for women recorded by the Global Media Monitoring Project (Global Media Monitoring Project, 2021).

RQ1 sought to determine whether creating awareness of ongoing measurements changed this ratio of source genders. Answering this required analyzing the quantitative results over time as students learned about the results and potentially changed their behaviour. Overall, there was a statistically significant change in the number of male sources quoted at the start of the term versus the end of the term. Analyzing the results by grade level saw no statistically significant change for 2nd-year students, while the results for post-graduate and 3rd-year students produced opposite results. The post-grads, which featured a class with only four male reporters, started the term with a gender balance ratio of 73% female and 26.7% male and gradually reached a more balanced representation by the end of the term (although this change was not statistically significant). Third-year students started with a gender balance of 38.6% female, 61.4% male, and 0% non-binary, finishing the term with a statistically significant change in representation: 47.5% female, 50.8% male, and 1.64% non-binary. This provides quantitative evidence that measurements could be effective in some classes.

RQ2 asked whether students might change their reporting strategies as they learned about the source-diversity measurements. This was addressed through two qualitative surveys. In questions only administered in the second survey, 73.5% of students said they were aware of ongoing efforts to track gender balance; when asked directly, just 41.2% said they changed how they searched for sources based on the latest measurement results. While the two surveys showed modest changes in students adopting new measures for finding interview sources over the year, this is inconclusive evidence that the desire to seek more balanced sources drove them to try new methods for finding voices. It may just be a sign that, as students gain more experience and more instructor feedback, they learn new reporting techniques as they are socialized into newsroom norms (Shoemaker & Reese, 2014, p. 154). The only possible exception is when the answers from third-year students are analyzed. These students were more likely than other grades to indicate they relied more on social media, expert guides, or assistance from faculty or colleagues in finding potential interview subjects. This could be attributed to their attempt to broaden their avenues for finding sources, or it could reflect the additional reporting opportunities they had by filing daily news reports.

Third-year students were the only students who faced daily deadline pressures. Their initial lower rates of female representation may have resulted from the deadline pressures they faced in this class. Trying to replicate real-world reporting pressures is one of the most significant challenges journalism faculty face (Josephi, 1999, p. 79). Deadlines force journalists to stop seeking new information and file their reports (Shoemaker & Vos, 2009, p. 196). The pressure of meeting a deadline means journalists privilege experts who return emails or phone calls and are willing to speak on short notice (Meer et al., 2016).

As one of the instructors for this group, my presence and regular reporting of data might have influenced their behaviour. To minimize this effect, I followed a consistent protocol for all grade levels, ensuring the results were posted to our program’s learning management system without engaging in additional discussions or interventions with the students about the data. The surveys showed that third-year students were more engaged in the source-tracking system and more aware of the weekly results. Furthermore, the content analysis of third-year stories showed a clear improvement over the 13 weeks. This suggests that as third-year students learned about their initially poor gender balance ratios, enough individuals changed their reporting routines to affect the overall results for their grade level.

By comparison, students in the second year or post-graduate courses had higher percentages of female voices in their stories, but they also had several days or even weeks to file their stories. This may have given them extra time to find sources they wanted and time to wait for a response before arranging for an interview. That is a luxury few journalists working to deadline enjoy. The literature suggests student journalists are more likely to include diverse sources in their stories than professional journalists (Cabrera et al., 2020; Content Diversity Board, 2020; Davenport & Grimm, 2021; Smith, 2008), but the research doesn’t explore whether that is a due to more generous deadlines, student perspectives on inclusion, or the demographic makeup of their campuses.

RQ3 asked to what degree the results differed when using the BBC’s methodology instead of just counting all sources. In analyzing the research data, this study found the BBC methodology under-represented male voices. The BBC methodology excludes “players” or “newsmakers:” interview subjects deemed editorially vital to a particular story. This differs from other approaches to content analysis where anyone who speaks is noted (Macharia, 2015), or sources are categorized by expert or non-expert status (Cann & Mohr, 2001). While the BBC approach does not penalize students for covering stories featuring official sources, it can end up underplaying the gender imbalance that comes from covering those who hold power in society (Sjøvaag & Pedersen, 2019) or in institutional stories about politics, policing, or business (Byerly, 2021). While I did not officially track story types, I did make note of a notable shift in the percentage of male “players” or “newsmakers” during weeks when major news events dominated the headlines. When the trucker convoy protest arrived in Ottawa (Week 3), much of the student coverage included quotes from newsmakers such as Prime Minister Trudeau, Ottawa’s police chief, and several provincial premiers. Likewise, the start of the Russian invasion of Ukraine (Week 5) saw the inclusion of quotes by President Joe Biden and Ukrainian President Volodymyr Zelenskyy. The BBC encourages partners to collect data over time and to expect numbers to fluctuate as news agendas can change week-to-week and affect measurement results. This was apparent in my data.

When the BBC methodology is used for academic purposes, excluding “players” or “newsmakers” poses significant limitations due to its inherent subjectivity, forcing researchers to judge what sources are considered essential to a story. This also raises pedagogical issues if students use this approach to measure their own work: it requires all participants to understand and agree upon the definitions of players and newsmakers. This can raise questions of stability (Krippendorff, 1980, p. 72) and whether coders’ “later judgements match their earlier judgements” (Potter & Levine-Donnerstein, 1999, p. 271).

CONCLUSION

Journalism has a gender representation problem. Although women slightly outnumber men in Canada, the United Kingdom, and the United States, it is rare for the average consumer of mainstream media to see or hear female experts quoted at that rate. The BBC’s 50:50 initiative successfully shifted the culture at the public broadcaster by regularly measuring source diversity rates and sharing them with staff, creating a real-time feedback loop. This study conducted an experiment to evaluate the BBC methodology within an educational context, and it produced mixed results.

An analysis of 13 weeks’ worth of student journalism content found students came close to meeting a 50:50 gender balance goal. The experiment successfully raised awareness about gender balance, but the evidence attributing that achievement to using the BBC 50:50 methodology is mixed. Research supports the idea that in courses where students must file stories under daily deadline pressures, routine reminders of the progress towards a newsroom goal might encourage them to adapt their reporting techniques to diversify their sources. Students who did not face this deadline pressure and thus had more time to file their stories did not appear to be as influenced by the regular feedback.

This study suggests the combination of realistic deadlines and weekly feedback generated from source diversity measurements might offer a robust pedagogical combination for journalism schools. Educators should also consider students’ overwhelming support of the goal that the media they consume and produce should reflect their audience’s diversity and gender balance. Journalism schools are already finding ways to integrate discussions of diversity across the curriculum using analysis/critical thinking assignments and community-orientated projects (Biswas et al., 2017). Source diversity tracking initiatives could provide a practical vehicle to encourage students to think about their interview choices, especially if instructors recreate the time pressures contributing to the under-representation of female sources in the broader news industry (Martindale, 2006).

Journalism  schools looking to introduce a source-diversity tracking system may want to consider what effect the BBC methodologies may have on their results. Excluding “players” or “newsmakers” introduces subjective decision-making and may affect intercoder reliability. In an early attempt to use this methodology with just one class, I asked a student editor to conduct his own analysis to compare against mine for accuracy. While he successfully tallied male, female and non-binary sources, we had consistent disagreements about who to exclude as a “player” or “newsmaker.” This was only reconciled through a detailed review of our application of the BBC’s rules. I conducted the content analysis for this study independently, as the student editor was unavailable, and the workload of analyzing three grade levels was substantial. To check the reliability of my sample and the reproducibility of my findings, I conducted a test-retest reliability exercise by randomly choosing 20 stories from my dataset to re-analyze.

In subsequent years, I experimented with a self-reporting tracking system, where student reporters were required to count their use of male, female and non-binary sources and record that data using a Microsoft Form that automatically populated an Excel spreadsheet. While this helped address workload issues involved in conducting a content analysis, it raised further problems of reliability, accuracy, and consistency problems. Given my previous concerns about the subjective elements of the BBC methodology, I decided to eliminate the “players” and “newsmakers” distinctions. Instead, I focused on tracking overall sources quoted through interviews, press conferences, social media embeds, and media statements. I am still reviewing that data. While the initial results were promising, this self-reporting system revealed new challenges with students’ accuracy and timeliness in recording data. This underscores the importance of assessing students’ understanding of whatever methodology is chosen and evaluating their accurate use before it is deployed. It also highlights the necessity of creating a system to verify the results once the source-tracking system is in place. Ensuring both comprehension and reliability is crucial for the effectiveness of any student or journalist-led source-diversity tracking initiative.

Finally, this study’s focus on gender diversity precluded a deeper look at ethnic or other intersectional diversities. Content analysis coding schemes can be expanded to allow for an intersectional analysis by comparing a source’s gender, ethnicity, or professional background, and the subject matter or length of the story in which they’re cited. Future research might consider focus groups or student interviews to measure attitudinal changes. While this project measured source diversity in online content, future research could measure and compare other student media, such as audio, video or social media. More effort could also be made to address a potential predisposition bias by seeking out views from students or reporters who are critical of source diversity efforts.

---

Mike Wise is a professor and program coordinator with the journalism program at Humber Polytechnic in Toronto. Email: mike.wise@humber.ca

---

REFERENCES

Asr, F. T., Mazraeh, M., Lopes, A., Gautam, V., Gonzales, J., Rao, P., & Taboada, M. (2021). The gender gap tracker: Using natural language processing to measure gender bias in media. PLos ONE, 16(1), 1–28. https://doi.org/10.1371/journal.pone.0245533

Barber, M., & Levitan, J. (2013). Not seen, not heard: Gender representation on canadian television news during the 2011 federal election campaign. The International Journal of Diverse Identities, 12(2). https://www.researchgate.net/publication/287047611_Not_seen_not_heard_Gender_representation_on_canadian_television_news_during_the_2011_federal_election_campaign

BBC. (2020). How to measure your content. BBC. http://bbc.com/5050

BBC. (2021). How it works. BBC. https://www.bbc.co.uk/5050/methodology

Biswas, M., Izard, R., & Roshan, S. (2017). What is taught about diversity and how is it taught? A 2015 update of diversity teaching at us journalism and mass communication programs. Teaching Journalism & Mass Communication, 7(1), 1–13. https://www.proquest.com/scholarly-journals/what-is-taught-about-diversity-how-2015-update/docview/1877372575/se-2

Byerly, C. (2021, October 8). Why white male dominance of news media is so persistent. Women’s Media Centre. https://womensmediacenter.com/news-features/why-white-male-dominance-of-news-media-is-so-persistent

Cabrera, H., Chiadmi, M., Laumonier, L., Payette, M., & Daoust-Boisvert, A. (2020, October 30). How tracking can improve gender representation in sourcing from j-school to the newsroom. J-Source. https://j-source.ca/how-tracking-can-improve-gender-representation-in-sourcing-from-j-school-to-the-newsroom/

Canadian Association of Journalists. (2023). Canadian newsroom diversity survey 2023. https://caj.ca/wp-content/uploads/Diversity_Survey_Report_2023_EN.pdf

Cann, D. J., & Mohr, P. B. (2001). Journalist and source gender in Australian television news. Journal of Broadcasting & Electronic Media, 45(1), 162–174. https://doi.org/10.1207/s15506878jobem4501_10

Carleton University. (2022, January 5). CP, Carleton journalism school awarded grant to study diversity in news sources | School of Journalism and Communication. https://carleton.ca/sjc/2022/cp-carleton-journalism-school-awarded-grant-to-study-diversity-in-news-sources/

Chilazi, S., Rattan, A., & Georgeac, O. (2020). Ros Atkins and the 50:50 project at the BBC. London Business School. https://publishing.london.edu/cases/ros-atkins-and-the-5050-project-at-the-bbc-a/

Content Diversity Board. (2020). Minnesota daily content diversity board. https://mndaily.com/wp-content/uploads/2020/04/Minnesota-Daily-Content-Diversity-Board-End-Of-Year-Report-2019-2020-9.pdf

Craig, K. (2017). Gender in local television news presentation: An analysis of TV news markets in the US Northwest [Masters Thesis].

Cukier, W., Elmi, M., & Gagnon, S. (2019). Content analysis: Examining representation in broadcast news. SAGE Research Methods Cases in Business and Management. https://doi.org/10.4135/9781526492425

Davenport, L. D., & Grimm, J. (2021). Diversity sourcing tool: Intentions, self-observation and learning. Annual Meeting of the AEMJC, 27. http://citation.allacademic.com/meta/p1870578_index.html

D’Heer, J., Vergotte, J., Vuyst, S. D., & Leuven, S. V. (2019). The bits and bytes of gender bias in online news: A quantitative content analysis of the representation of women in Vice.com. Feminist Media Studies, 20(2), 1–17. https://doi.org/10.1080/14680777.2019.1574858

Entman, R. M. (2010). Media framing biases and political power: explaining slant in news of campaign 2008.Journalism, 11(4), 389–408. https://doi.org/10.1177/1464884910367587

Everbac h, T., Batsell, J., Champlin, S., & Nisbett, G. S. (2010). Does a more diverse newspaper staff reflect its community? A print and digital content analysis of the Dallas Morning News. Southwestern Mass Communication Journal, 35(1), 1–17. https://journals.tdl.org/swec-jmc/index.php/swecjmc/article/view/41

Farquhar, L. K., & Carey, M. C. (2019). Self-censorship among student journalists based on perceived threats and risks. Journalism & Mass Communication Educator, 74(3), 318–335. https://doi.org/10.1177/1077695818815276

Fenlon, B. (2020, November 20). What CBC’s newsrooms are doing to deliver on pledge to improve diversity, inclusivity | CBC News. Editor’s Blog. https://www.cbc.ca/news/editorsblog/editors-blog-inclusivity-update-1.5796872

Fu, A. (2021, August 12). New tool allows NPR to track source diversity in real time. Poynter. https://www.poynter.org/reporting-editing/2021/new-tool-allows-npr-to-track-source-diversity-in-real-time/

Global Media Monitoring Project. (2015). Who makes the news? England, Scotland, Wales and the Republic of Ireland global media monitoring project 2015 four-nation report. Global Media Monitoring Project. https://whomakesthenews.org/wp-content/uploads/who-makes-the-news/Imported/reports_2015/national/UK.pdf

Global Media Monitoring Project. (2021). Who makes the news? Canada national report. World Association for Christian Communication. https://whomakesthenews.org/wp-content/uploads/2021/08/CANADA-GMMP-Country-Report-FINAL.pdf

Graydon, S. (2017, October 18). What gets measured gets done – #WGMGD. InformedOpinions.Org.  https://informedopinions.org/what-gets-measured-gets-done/

Haney, S. K., & Paskey, J. (2020). Canadian journalism programs respond to local news poverty: The challenge of doing more. Canadian Journal of Communications 45(4). https://doi.org/10.22230/cjc.2020v45n4a3769

Humprecht, E., & Esser, F. (2017). A glass ceiling in the online age? Explaining the underrepresentation of women in online political news. European Journal of Communication, 32(5), 439–456. https://doi.org/10.1177/0267323117720343

Josephi, B. (1999). From journalism school to newsroom: What rite of passage? Asia Pacific Media Educator, 7, 74–85. https://ro.uow.edu.au/apme/vol1/iss7/5

Krippendorff, K. (1980). Validity in content analysis. In Computerstrategien für die Kommunikationsanalyse (pp. 69–112). Campus. https://repository.upenn.edu/entities/publication/67c47fe1-87b9-417a-b1e4-018c011a770a

Liebler, C. M., & Smith, S. J. (1997). Tracking gender differences: A comparative analysis of network correspondents and their sources. Journal of Broadcasting & Electronic Media, 41(1), 58. https://doi.org/10.1080/08838159709364390

Lu, K. (2021, December 7). What inclusion work looks like in 4 Chicago newsrooms. Poynter. https://www.poynter.org/from-the-institute/2021/what-inclusion-work-looks-like-in-4-chicago-newsrooms/

Lyons, J. H. (2002). Audit trails: A new tool in gatekeeping research (University Of Florida) [Masters Thesis]. https://www.researchgate.net/publication/228895313_Audit_Trails_A_New_Tool_in_Gatekeeping_Research

Macharia, S. (2015). Global media monitoring project 2015 (Global Media Monitoring Project). World Association for Christian Communication. https://whomakesthenews.org/gmmp-2015-reports/

Martindale, S. (2006). Source diversity in journalism. University of Southern California.

McKeon, L. (2011). Five questions for Shari Graydon. JSource. https://j-source.ca/article/five-questions-for-shari-graydon/

Meer, T. G. L. A. van der, Verhoeven, P., Beentjes, J. W. J., & Vliegenthart,(2016). Disrupting gatekeeping practices: Journalists’ source selection in times of crisis. Journalism, 18(9), 1107–1124. https://doi.org/10.1177/1464884916648095

Newsome, M. (2021, April 1). Diverse sources, inclusive reporting. Neiman Reports, 75(2), 32–33.

O’Brien, A., & Suiter, J. (2017). Best and worst practice: A case study of qualitative gender balance in Irish broadcasting. Media, Culture & Society, 39(2), 259–269. https://doi.org/10.1177/0163443716686942

Office for National Statistics. (2023, August 2). Male and female populations. GOV. UK Ethnicity Facts and Figures. https://www.ethnicity-facts-figures.service.gov.uk/uk-population-by-ethnicity/demographics/male-and-female-populations/latest/

Potter, W. J., & Levine-Donnerstein, D. (1999). Rethinking validity and reliability in content analysis. Journal of Applied Communication Research, 27(3), 258–284. https://doi.org/10.1080/00909889909365539

Rao, P., Taboada, M., & Graydon, S. (2021, October 28). What we can learn from three years of data on the gender gap in news reporting. Poynter. https://www.poynter.org/reporting-editing/2021/what-we-can-learn-from-three-years-of-data-on-the-gender-gap-in-news-reporting/

Rattan, A., Chilazi, S., Georgeac, O., & Bohnet, I. (2019, June 6). Tackling the underrepresentation of women in media. Harvard Business Review. https://hbr.org/2019/06/tackling-the-underrepresentation-of-women-in-media

Shoemaker, P. J., & Reese, S. D. (2014). Mediating the message in the 21st century. A media sociology perspective. 3rd edition. Routledge. https://doi.org/10.4324/9780203930434

Shoemaker, P. J., & Vos, T. P. (2009). Gatekeeping theory. Routledge. Sjøvaag, H., & Pedersen, T. A. (2019). Female Voices in the News:

Structural conditions of gender representations in Norwegian newspapers. Journalism & Mass Communication Quarterly, 96(1), 215–238. https://doi.org/10.1177/1077699018789885

Skinner, D., Gasher, M. J., & Compton, J. (2001). Putting theory to practice. Journalism, 2(3), 341–360. https://doi.org/10.1177/146488490100200304

Smith, L. K. (2008). Race, ethnicity, and student sources: Minority newsmakers in student-produced versus professional TV news stories. Howard Journal of Communications, 19(2), 182–199. https://doi.org/10.1080/10646170801990995

Statistics Canada. (2022, November 23). Women – Canada at a glance. Government of Canada. https://www150.statcan.gc.ca/n1/pub/12-581-x/2022001/sec7-eng.htm

Swert, K. D., & Hooghe, M. (2008). When do women get a voice? Explaining the presence of female news sources in Belgian TV news. International Communication Association, 1–29. https://doi.org/10.1177/0267323109354229

Tuchman, G. (1980). Making news. Free Press.

US Census Bureau. (n.d.). Quickfacts: United States. Retrieved June 28, 2024, from https://www.census.gov/quickfacts/fact/table/US/SEX255223#SEX255223

Voakes, P. S., Kapfer, J., Kurpius, D., & Chen, D. S.-Y. (1996). Diversity in the news: A conceptual and methodological framework. Journalism & Mass Communication Quarterly, 73(3), 582–593. https://doi.org/10.1177/107769909607300306

Vogler, D., & Schwaiger, L. (2021). Situational effects of journalistic resources on gender imbalances in the coverage of Swiss news media: A longitudinal analysis from 2011 to 2019. Journalism, 146488492110363. https://doi.org/10.1177/14648849211036309

Wenger, D. H., Owens, L. C., & Cain, J. (2018). Help wanted: Realigning journalism education to meet the needs of top US news companies. Journalism & Mass Communication Educator, 73(1), 18–36. https://doi.org/10.1177/1077695817745464

Wittenberg-Cox, A. (2020, July 10). How the BBC gender balanced… everything, everywhere, fast. https://www.forbes.com/sites/avivahwittenbergcox/2020/07/10/how-the-bbc-gender-balanced-everything-everywhere-fast

Zeldes, G. A., Fico, F., & Diddi, A. (2012). Differences in the way broadcast, cable and public TV reporters used women and non-white sources to cover the 2008 presidential race. Mass Communication and Society, 15(6), 831–851. https://doi.org/10.1080/15205436.2011.634084

Zoch, L. M., & Turk, J. V. (1998). Women making news: Gender as a variable in source selection and use. Journalism and  Mass Communication Quarterly, 75(4), 762–775. https://doi.org/10.1177/107769909807500410