Layoffs, Inequity and COVID-19: A Longitudinal Study of the Journalism Jobs Crisis in Australia from 2012 to 2020
Nik Dawson, Sacha Molitorisz, Marian-Andrei Rizoiu, Peter Fray
LLayoffs, Inequity and COVID-19: ALongitudinal Study of the JournalismJobs Crisis in Australia from 2012 to2020
SAGE
Nik Dawson , Sacha Molitorisz , Marian-Andrei Rizoiu and Peter Fray Abstract
In Australia and beyond, journalism is reportedly an industry in crisis, a crisis exacerbated by COVID-19. However,the evidence revealing the crisis is often anecdotal or limited in scope. In this unprecedented longitudinal research,we draw on data from the Australian journalism jobs market from January 2012 until March 2020. Using Data Scienceand Machine Learning techniques, we analyse two distinct data sets: job advertisements (ads) data comprising 3,698journalist job ads from a corpus of over 6.7 million Australian job ads; and official employment data from the AustralianBureau of Statistics. Having matched and analysed both sources, we address both the demand for and supply ofjournalists in Australia over this critical period. The data show that the crisis is real, but there are also surprises.Counter-intuitively, the number of journalism job ads in Australia rose from 2012 until 2016, before falling into decline.Less surprisingly, for the entire period studied the figures reveal extreme volatility, characterised by large and erraticfluctuations. The data also clearly show that COVID-19 has significantly worsened the crisis. We can also teaseout more granular findings, including: that there are now more women than men journalists in Australia, but thatgender inequity is worsening, with women journalists getting younger and worse-paid just as men journalists are,on average, getting older and better-paid; that, despite the crisis besetting the industry, the demand for journalismskills has increased; and that the skills sought by journalism job ads increasingly include social media and generalistcommunications.
Keywords
Journalism, Jobs, Skills, COVID-19
Introduction
In Australia, the news about the news is not good.In early March 2020, newswire service the AustralianAssociated Press announced it would be shutting down itsoperations after 85 years. ‘Investors look to salvage partsof AAP as newswire faces closure’, reported the SydneyMorning Herald on March 2 (Samios 2020). The AustralianBroadcasting Corporation (ABC) predicted that 500 peoplewould lose their jobs as a result (Khadem and Pupazzoni2020). In the US, the news about the news is just as bad,if not worse. ‘On a rough day for American newspapers,investors arent buying Gannetts story and Tribunes not donechopping’ was the headline on a Nieman journalism Labstory published on February 27 (Benton 2020). Accordingto the report, layoffs looked likely at the countrys No. 1and No. 3 newspaper chains, while the countrys No. 2 chain(McClatchy) had already declared bankruptcy a fortnightearlier. As the Nieman Lab notes:The Internet has brought forth an unprece-dented flowering of news and information. Butit has also destabilized the old business mod-els that have supported quality journalism fordecades. Good journalists across the country arelosing their jobs or adjusting to a radically newnews environment online (Nieman-Lab 2020). Is journalism in crisis? A wealth of research in Australia,the US and comparable countries suggests yes. Profitsare hard, if not impossible, to come by; many firms arestruggling or collapsing; and layoffs and redundancies arethe norm. As Fenton (2011) wrote in a paper centred onthe UK, ‘News media are in crisis. The crisis is beingmanaged by closing papers or shedding staff [and] thesecuts are having a devastating effect on the quality of thenews.’ That was nearly a decade ago. Since then, the situationhas only worsened. In Australia, the commonly cited figurebased on research by the journalists union is that 3,000journalism positions have been lost since 2011 (Ricketsonet al. 2020). For instance, it is estimated that in 2011 newspublisher Fairfax Media employed about 1,000 editorial staffacross the
Sydney Morning Herald , The Age , The AustralianFinancial Review , and its Sunday papers,
The Sun Herald and
The Sunday Age . By mid-2017, however, half of thosejobs were gone (Zion et al. 2018), including the job of one Centre for Artificial Intelligence, University of Technology Sydney OECD Future of Work Research Fellow Centre for Media Transition, University of Technology Sydney UTS Data Science Institute, University of Technology Sydney Private Media
Corresponding author:
Nik DawsonEmail: [email protected]
Prepared using sagej.cls [Version: 2017/01/17 v1.20] a r X i v : . [ ec on . GN ] A ug Journal Title XX(X) of this papers authors. And then the coronavirus wielded itsscythe. As we discuss below, the impact of COVID-19 onjournalism jobs is proving devastating, with widespread joblosses, particularly in regional areas (Crerar 2020).This research aims to assess the extent of the claimed‘journalism crisis’ in Australia by analysing labour marketdata from 2012 to 2020. Our findings confirm that thereis a crisis in journalism; a crisis that is now in full bloomdue to the coronavirus pandemic. However, the data alsoyields more granular findings, including several surprises.One finding is that advertised journalism jobs only startedto decline from 2016, not before. A second is that asthe journalism jobs market becomes more volatile, genderinequity is worsening: women journalists who remain areyounger and worse paid than the men who remain. And athird is that according to our skill similarity calculations,generalist ‘Communications’, ‘Public Relations’, and ‘SocialMedia’ are skills that are becoming more important tojournalism, as opposed to traditionally specialist journalismskills such as ‘Reporting’, ‘Editing’, and ‘InvestigativeJournalism’. These findings, together with others, reveal thatthe crisis in journalism is not only real, but in some waysmore concerning than was previously understood.To fulfil this research aim, we analyse a range oflongitudinal data sources from job advertisements (ads)and official Australian employment statistics. The breadthand detail of these data provide us with the opportunityto comprehensively assess the journalism jobs market inAustralia and how it has changed. We apply Data Science andMachine Learning techniques to analyse how the underlyingskills of journalists in Australia have evolved. This allows usto build a data-driven methodology to determine which arethe top journalism skills per year and identify the occupationsin possession of these skills. Finally, we use these skill-levelresults to determine where people with journalism skills arelikely finding alternative career paths.
The main contributions of this research include: • providing a comprehensive and longitudinal assess-ment of journalism jobs in Australia from 2012 to2020 by analysing both job ads data and occupationalemployment statistics;• implementing a data-driven methodology to explorethe nature of the oft-cited ’crisis’ in journalism jobsin Australia, ;• applying this data-driven methodology to tease outmore granular and specific trends in journalismjobs , including the impact of the current coronaviruspandemic, the contrasting impacts on regional andurban journalism jobs, and the gendered nature ofongoing impacts; and• analysing the data to identify the skills sought injournalism jobs, and where people with journalismskills are likely finding alternate career paths.
Related Work & Background
Journalism jobs in crisis.
If there is a crisis, the simpleexplanation is the Internet. (Putting aside COVID-19, towhich we will return.) While digital channels have given journalism bigger audiences, they have also strangledincome. Once, advertising funded journalism, but nowadvertising has largely migrated online. As the AustralianCompetition and Consumer Commission (ACCC) found in2019, in the Final Report of its Digital Platforms Inquiry,‘The reduction in advertising revenue over the past 20 years,for reasons including the rise of online advertising, appearsto have reduced the ability of some media businesses tofund Australian news and journalism’. The ACCC citedCensus data showing that ‘from 2006 to 2016, the numberof Australians in journalism-related occupations fell by9% overall, and by 26% for traditional print journalists(including those journalists working for print/online newsmedia businesses)’. Further, the ACCC cited data providedby leading media companies showing that the number ofjournalists in traditional print media businesses fell by20% from 2014 to 2018 a time of growth for Australia’spopulation and economy (ACCC 2019).However, the pressures on news media are not spreadevenly. For instance, local news is bearing a particularbrunt. Between 2008 and 2018, 106 local and regionalnewspaper titles closed across Australia, representing a 15%decrease in the number of such publications. As a result, 21local government areas previously served by a newspaperwere now without coverage, including 16 local governmentareas in regional Australia (ACCC 2019). These figures aremirrored in the US. In 2018, Abernathy (2018) from theHussman School of Journalism and Media at UNC releaseda report, ‘
The Expanding News Desert ’, which found thatthe US had lost almost 1800 papers since 2004, with 7112remaining (1283 dailies and 5829 weeklies). This means thatthe US lost roughly 20% of its newspapers between 2004and 2018. These closures included large dailies such as the
Tampa Tribune and the
Rocky Mountain News , but also manynewspapers that had circulations of fewer than 5000 andserved small, impoverished communities.The big picture reveals that, in an era of misinformation,social media and news aggregators, news media companiesare under pressure, and journalism jobs are being cut. Thereis some hope in the shape of new players entering themarket and hiring journalists, including in the shape ofdigital natives such as Vice and Buzzfeed. However, in 2019these two companies were among the many that announcedsignificant staff layoffs (Goggin 2019). What’s more, asAustralia’s ACCC notes, these publications ‘tend to employrelatively few journalists’ (ACCC 2019). Even accountingfor new arrivals, the number of journalism jobs in Australiais falling (see
Jobs Data Analysis and Results ), and as aresult there are areas (including local government, localcourt, health and science issues) that journalism is no longercovering adequately (ACCC 2019).Further research is also revealing a clearer profile ofthe typical journalist, and also the typical journalist wholoses his/her job. Drawing on 2017 data, one study foundthat journalism jobs internationally are largely filled bya young, inexperienced and itinerant workforce (Josephiand Oller Alonso 2018). Meanwhile, research suggeststhat it is journalists with extensive experience who arelosing their jobs (at least in Australia) (Sherwood andO’Donnell 2018). And those who lose their jobs facedecidedly uncertain futures. In longitudinal research tracking
Prepared using sagej.cls awson et al. the post-journalism careers of Australian journalists whohad been made redundant, many of those surveyed revealedthey were experiencing job precarity. Further, a significantminority had moved into strategic communications or publicrelations (Zion et al. 2018). The impacts of COVID-19.
The advertising crisis forjournalism has been described not as a single black swan, butas a flock of black swans (Doctor 2020). According to oneestimate, from 2006 to 2020, US newspapers lost more than70 percent of their ad dollars (Doctor 2020). And then cameCOVID-19. Just as the coronavirus has been claiming lives,it has also been claiming journalism jobs, with particularlydevastating impacts on regional and local news outlets.This is true in many countries, including the US. In March,layoffs were announced at the
Detroit Metro Times andits six sibling mastheads, with remaining staff told theirpay would be cut (Flynn 2020). With concerts cancelledand restaurants shuttered, promoters and restaurateurs hadnothing to advertise. On March 25, 2020,
The Atlantic rana story under the headline, ’The coronavirus is killing localnews’ (Waldman and Sennott 2020). The story called forgovernment and philanthropic intervention, and for peopleto subscribe: ‘Among the important steps you should takeduring this crisis: Wash your hands. Don’t touch your face.And buy a subscription to your local newspaper.’ In amatter of weeks, many American news websites’ advertisingrevenues are said to have fallen by as much as 50%. As onemedia expert noted in late March, ’Advertising, which hasbeen doing a slow disappearing act since 2008, has beencut in half in the space of two weeks’ (Doctor 2020). Flynn(2020) reported in March, ‘At least 100 people have lost theirjobs in media over the past two weeks, with most outletsciting coronavirus as the direct cause.’In the UK in April,
The Guardian reported thatnewspapers were set to lose 57million if the outbreak lastedfor another three months (Sweney 2020). This was partlybecause advertisers were refusing to place their ads nextto stories about the pandemic, which they deemed to beinappropriate content.In Australia too, as we have seen, there were widespreadclosures and job losses before coronavirus, but COVID-19compounded the problem. In late March, Rupert Murdoch’spublishing business News Corp warned of ’inevitable’ jobcuts and the closure of regional titles (Meade 2020b). Soonafterwards, News Corp Australia’s biggest publisher -suspended the print editions of 60 Australian newspapers,including the
Manly Daily and
Wentworth Courier inSydney, the
Brisbane News and the
Mornington PeninsulaLeader in Victoria (Meade 2020b). The cuts came inthe wake of a dramatic drop in advertising from theentertainment, restaurant and real industries, the titles’ mainrevenue sources.In many countries, governments have announced assis-tance packages. On April 6, the Australian governmentannounced it would bring forward the release of $5mil-lion from its Regional and Small Publishers InnovationFund to support public interest journalism during COVID-19 (Fletcher 2020). In April 2020, the Danish governmentallocated approximately 24m to save local media. ‘Thescheme can compensate for the lost advertising revenue,’ saidculture minister Joy Mogensen (Zalan 2020). By contrast, however, some governments are making thecoverage of coronavirus harder. In China, authorities havecracked down on doctors and reporters who exposed theoutbreak (Kuo 2020); in the US, journalists are being barredfrom talking to staff at public hospitals (Carville et al.2020); and in countries including Venezuela, Niger andIndia, journalists have been arrested and intimidated (CPJ2020).
Job ads as a proxy for labour demand.
Job ads provide‘leading’ indicators of shifting labour demands as they occur,as opposed to the ‘lagging’ indicators from labour marketsurveys. Consequently, job ads are increasingly used as adata source for analysing labour market dynamics (Markowet al. 2017; Blake 2019). For instance, job ads data havealso been used to assess labour shortages. Dawson et al.(2019) defined a range of indicators to evaluate the presenceand extent of shortages, such as posting frequency, salarylevels, educational requirements, and experience demands.They also built a metric based on the forecasting errorfrom Machine Learning models trained to predict postingfrequency. Intuitively, occupations experiencing high postingvolatility are difficult to predict. Subsequent work showedthese indicators to be predictive of labour shortages in theAustralian Labour Market (Dawson et al. 2020). In thepresent research, in
Jobs Data Analysis and Results , weuse a similar set of indicators to analyse labour demand forjournalists. Further details on job ads data are provided in theSupplemental Material.
Analysing journalism jobs with job ads.
Journalismjobs have also previously been analysed using job ads.Young and Carson (2018) collected and assessed howAustralian media outlets defined journalism job positionswhen hiring journalists from November 2009 to November2010. The authors used a content analysis methodology andmanually labelled data fields, such as employer, educationalqualifications, job responsibilities, experience requirements,location, work hours, media platform, skill demands, jobtitle, and any other miscellaneous information. The authorsfound that journalism was not a high priority during thisperiod; instead employers advertised four times as many jobads for sales, marketing, and advertising positions.More recently, Guo and Volz (2019) conducted contentanalysis on 669 journalist job announcements from USmedia organisations, as posted on
Indeed.com from 1July to 31 December 2017. The authors’ objective wasto define, compare, and analyse the journalists’ expertiserequirements as expressed through job ads. To achieve thisobjective, the authors manually reviewed and codified jobvacancies. This research found that ‘multi-skilled’ journalistsare experiencing higher levels of demand. The authors alsofound that journalists’ ability to flexibly adapt to changingsituations was a characteristic of growing importance. Thesestudies, while significant, are relatively limited in scope. Inthis paper, we analyse a nine-year long dataset which allowsus to uncover longitudinal dynamics of journalism jobs.
Limitations of job ads data.
Job ads data arean incomplete representation of labour demand. Someemployers use traditional forms of advertising for vacancies,such as newspaper classifieds, their own hiring platforms,or recruitment agency procurement. Furthermore, anecdotalevidence reveals that some vacancies are filled informally,
Prepared using sagej.cls
Journal Title XX(X) using channels such as word of mouth, professionalnetworks and social media. Job ads data also over-represent occupations with higher-skill requirements andhigher wages, colloquially referred to as ‘white collar’jobs (Carnevale et al. 2014). Finally, just because a job isadvertised, does not mean that the position will be, or hasbeen, filled.
Employment statistics and occupational standards.
Employment statistics provide data on populations employedin standardised occupational classes. Occupations in Aus-tralia correspond to their respective occupational classesaccording to the Australian and New Zealand Standard Clas-sification of Occupations (ANZSCO) (Australian Bureau ofStatistics 2013).There are significant shortcomings to analysing occu-pations within ANZSCO categories. Official occupationaltaxonomies (like ANZSCO) are often static and are rarelyupdated, therefore failing to capture emerging skills, whichcan misrepresent the true labour dynamics of particular jobs.For example, the occupational class of ‘Print Journalist’ hasbeen a constant in Australian occupational statistics. Yet,the underlying skills of a ‘Print Journalist’ have changeddramatically in recent decades.To overcome the above-stated limitations, in our dataconstruction, we leverage the BGT occupational ontologytogether with the ANZSCO ontology. We also use the richskill-level information from job ads that are missing fromoccupational employment statistics to build an encompassingjournalism job ads dataset.
Data & Methods
Data Sources
This research uses both labour demand and labour supplydata to analyse journalism jobs. On the labour demandside, we use a detailed dataset of over 6.7 millionAustralian job ads, spanning from January 2012 to March2020. These data were generously provided by BurningGlass Technologies * (BGT). For labour supply data, weleverage official employment statistics (Australian Bureauof Statistics 2019a) and salary levels (Australian Bureauof Statistics 2019b) provided by the Australian Bureau ofStatistics (ABS) over the same period. These data sourcesprovide longitudinal employment and salary informationthat have been disaggregated by gender, location, and typesof employment (full-time and part-time). Further detailsof data sources and data construction are provided in theSupplemental Material. Skill Similarity
To analyse the underlying journalism skills within occupa-tions, we implement a skill similarity methodology adaptedfrom Alabdulkareem et al. (2018) and then by Dawson et al.(2019) to calculate the pairwise similarities between skillsfrom job ads.
Skill similarity.
Two skills are similar when the twoare related and complementary, i.e. the two skills in askills-pair support each other. For example, ‘Journalism’and ‘Editing’ have a high pairwise similarity score becausetogether they enable higher productivity for a journalist; whereas ‘Journalism’ and ‘Oncology’ have a low similaritybecause they are generally seldom used jointly. We measurethe similarity of skill-pairs based on their co-occurrencepatterns in job ads, while accounting for skill ubiquityand specialisation. To capture how journalism skills havechanged over time, we measure skill similarity duringcalendar years.Formally, given J as the set of job ads posted during aspecific calendar year, we measure the similarity betweentwo skills s and s (cid:48) as: θ ( s, s (cid:48) ) = (cid:80) j (cid:48) ∈ J e ( j, s ) e ( j, s (cid:48) ) max (cid:32) (cid:80) j (cid:48) ∈ J e ( j, s ) , (cid:80) j (cid:48) ∈ J e ( j, s (cid:48) ) (cid:33) (1)where j and j (cid:48) are individuals jobs ads from the set J , and e ( s, j ) ∈ { , } measures the importance of skills s for job j using theory from Trade Economics (Hidalgo et al. 2007).Skills s and s (cid:48) are considered as highly complementaryif they commonly co-occur and are both ‘important’ forthe same job ads. Finally, θ ( s, s (cid:48) ) ∈ [0 , , a larger valueindicates that s and s (cid:48) are more similar, and it reaches themaximum value when s and s (cid:48) always co-occur (i.e. theynever appear separately).We build the top yearly lists of journalism skills bycomputing θ ( Journalism, s ) – i.e. the similarity betweenthe skill ‘Journalism’ and each unique skill that occursfor each year from 2014-2018. The yearly top 50 mostsimilar skills to ‘Journalism’ are shown in the SupplementalMaterial together with the full details of the θ measure. Journalism skill intensity.
Finally, we determine theoccupations that most require the top journalism skillsuncovered from the above. We propose η , the ‘JournalismSkill Intensity’ , for each standardised BGT occupation,defined as percentage of journalism skills relative to thetotal skill count for the job ads related to an occupation o .Formally: η ( o, D ) = (cid:80) j ∈O ,s ∈D x ( j, s ) (cid:80) j ∈O ,s (cid:48) ∈ S x ( j, s (cid:48) ) (2)where D is the set of journalism skills, and O is the set ofjob ads associated with the occupation o . This method allowsus to adaptively select occupations based on their journalismskill intensity. Jobs Data Analysis and Results
In this section, we perform a data-driven analysis ofjournalism jobs in Australia based on job ads data and officialoccupational statistics. First, we longitudinally examine keyfeatures of jobs data, such as employment levels, job adsposting frequency, salaries, and posting frequency growthand predictability level. We also analyse how the underlyingskills of journalists have changed over time, and which skillsand occupations are growing in similarity to journalism.
Prepared using sagej.cls awson et al. Journalism Jobs AdsJournalists Employed (Unit Level) J ou r n a li s m Q ua r t e r l y J ob A d s J ou r n a li s t s E m p l o ye d Q ua r t e r l y ( ' s ) Figure 1.
Quarterly posting frequency of journalism job ads(see Sec.
Data & Methods ) and employment levels of‘Journalists & Other Writers’ at the ANZSCO Unit level (000’s)from Jan 2012 to Dec 2019.
Posting Frequency & Employment levels
In Australian journalism, 2012 is considered a watershedyear. An estimated 1,500 journalists were made redundant,the majority of those from Australia’s two largest printcompanies, Fairfax Media (now Nine Entertainment) andNews Limited (now News Corp Australia) (Zion et al.2016). The severity of this industrial shock can be observedin Fig. 1. Against the left y-axis, the blue line showsquarterly job ads posting frequency for journalism jobs. Asthe graph depicts, posting frequency for journalism job adsexperienced extremely low levels in 2012 until 2013, whenthey began to increase. The volume of vacancies increaseduntil mid-2014, before plummeting in late-2014 to the levelslast seen in 2012. From 2015, journalism job ads experiencedstrong growth, reaching a peak in mid-2016. Since then,journalism job ads have trended downward until the end of2019, albeit with volatile peaks and troughs. In summary, thedata shows that journalism job ads have not been in freefallsince 2012. Rather, there was erratic growth in journalismjob ads until a peak in 2016, followed by erratic decline.Similarly, employment levels underwent immense volatil-ity from 2012 to 2013. Against the right y-axis of Fig. 1,the orange line shows the number of quarterly employedfor ‘Journalists & Other Writers’ at the ANZSCO Unitlevel. Employment levels peaked in mid 2012, before dra-matically dropping in early 2013. This is an effect of thethe mass journalist redundancies made in 2012, given thatemployment statistics are ‘lagging indicators’ and it takestime for labour markets to reflect changes in occupationalstatistics. Early 2013 marked the lowest point of journalistemployment seen in this time-series. As also observed injob ads data, journalist employment levels grew until 2016-2017 and has since trended downwards, exhibiting volatilequarterly changes through to the end of 2019.
COVID-19 and journalism jobs.
The early effects ofCOVID-19 are apparent in the posting frequency of job adsin Australia. This is the case for most occupations, includingjournalists. At the time of writing, official Australianemployment statistics were not yet available, making itdifficult to determine the extent of job losses caused by thepandemic. However, job ads provide a leading indicator oflabour demand (Dawson et al. 2019). Higher vacancy rates typically mean higher levels of labour demand by employers,which is a critical component of healthy labour markets.As Fig. 2 highlights, vacancy volumes have declined forboth journalism jobs and at aggregate levels in Australia.Since mid-February, weekly posting frequency has decreasedacross all Australia job ads, as seen in Fig. 2a. Such adecline this early in the year is atypical. As Dawson andRizoiu (2020) show, the frequency of job ad postings follow ayearly seasonal pattern, with late February and early Marchtypically being a period of upward trend growth. However,late February and early March 2020 coincided with theInternational outbreak of COVID-19. During this period, theAustralian government instituted widespread quarantine andsocial distancing measures, which significantly constrainedeconomic activity (Boseley and Knaus 2020). The impactsof these COVID-19 containment laws are starkly apparentin Fig. 2b. Posting frequency for journalism jobs are down63% when comparing March 2019 volumes to March 2020.This is significantly higher than the aggregate market of allAustralian job ads, which is down 37% over the same period.Fig. 2b shows that Melbourne appears to be the city hardesthit, recording no journalism job ads in March 2020 and only3 posts for the first quarter of 2020. Clearly the pandemicis having a highly damaging effect on the journalism jobsmarket.
Salaries
We compare salaries extracted from job ads with ABSreported wage data for ‘Journalists and Other Writers’ † .Fig. 3 reveals two main findings regarding journalistsalaries. First, according to job ads data, journalists attractconsiderably lower annual wage levels (solid blue line)than the market average (dashed blue line). As of 2018,job ads indicate that journalists earn approximately $10,000less than the market average. These findings, however, aresomewhat contrary to the wage earnings data collected bythe ABS (Australian Bureau of Statistics 2019b), accordingto which ‘Journalists and Other Writers’ (solid orange line)have been earning a growing wage premium over the marketaverage (dashed orange line) since 2014. This discrepancycan be explained by the fact that job ads data tend to over-represent occupations in the ‘Professional’ and ‘Manager’classes (Carnevale et al. 2014), which typically attract higherwages. As a result, the average salary levels from job adsdata (dashed blue line) are about $20,000 higher than averagesalary levels from ABS data (dashed orange line), from 2014to 2018. However, the salary levels for journalists are verysimilar when comparing across the two data sources.Fig. 3 yields a second observation: journalist salary levelsincreased in both absolute and relative terms compared toaverage market levels, between 2012 to 2018 in both datasources. More importantly, the relative salary growth ofjournalists has exceeded the market averages, during theperiod studied. ∗ BGT is a leading vendor of online job ads data. † ABS wage data is reported biennially, with the latest reporting year being2018. Therefore, wage values in the ‘odd’ years in between the reportingperiods were interpolated, calculated as the mean of the previous and thesubsequent years.
Prepared using sagej.cls
Journal Title XX(X)
All AUS Journalist ads Sydney Melbourne Brisbane All AUS Job Ads
May 2019 Jul 2019 Sep 2019 Nov 2019 Jan 2020 Mar 20200510152025 10k15k20k25k30k J ou r n a li s m W ee k l y J ob A d s A ll AU S W ee k l y J ob A d s Jan 2019 Mar 2019 May 2019 Jul 2019 Sep 2019 Nov 2019 Jan 2020 Mar 20200102030405060 70k75k80k85k90k95k100k105k J ou r n a li s m M on t h l y J ob A d s A ll AU S M on t h l y J ob A d s (a) (b) COVID-19 and journalism jobs posting
Figure 2. Posting frequency for journalism jobs during the early stages of the COVID-19 crisis in Australia and its majorcities : (a) Weekly posting frequency volumes for journalists and all Australian job ads between April 2019 and March 2020. Bothdecrease as the early stages of the COVID-19 crisis hit; (b)
Monthly posting frequency for journalists are down 63% whencomparing March 2019 to March 2020. This is significantly higher than all Australian job vacancies, which is down 37% over thesame period.
Journalist salaries job ads averageAll AUS job ads average Journalists Unit Group salaries averageUnit Group salaries average S a l a r i e s ( A UD ) J ou r na li s t s a l a r i e s Figure 3.
Journalist salaries (solid blue line) have increasedaccording to job ads data, but still remain below market averagelevels (dashed blue line). However, according to ABS data,‘Journalists & Other Writers’ (ANZSCO Unit level, solid orangeline) earn a growing wage premium above the market average(dashed orange line).
Trend Analysis & Predictability
Posting trends.
We constructed an auto-regressive MachineLearning model to predict posting frequency of journalismjob ads in Australia (Dawson et al. 2019). The modelaccounts for long term trends, seasonality patterns andexternal events (see the Supplemental Material for technicaldetails). We isolated the posting frequency trend componentand, in Fig. 4, plotted it comparatively for ‘Journalists’against two occupations experiencing high levels of labourdemand, ‘Data Scientists’ and ‘Data Analysts’, as well asagainst the aggregated market trend. Visibly, journalismjobs experienced varying degrees of growth until mid 2016,at which point growth plateaued, and started to decline.From the end of 2017 until the beginning of 2019, thetrend for journalism job ads has heavily decreased, evenwhen compared to the aggregate market, which also showsa more modest decrease during the same period. ‘DataScientists’ and ‘Data Analysts’ have been consistentlygrowing throughout the the entire period.
Journalists
Data Scientists
Data AnalystsAll Australian Jobs
Figure 4. Trend lines of posting frequency for ‘Journalists’,‘Data Scientists’, ‘Data Analysts’, and ‘All Australian job ads’.Posting frequency for ‘Journalists’ have trended downwardssince 2016.
Quantify labour demand volatility.
When constructingMachine Learning models, it is standard procedure touse error metrics to evaluate the prediction accuracy.Volatility in posting volumes inherently lead to loweredprediction performance. Here we use the prediction errormeasured using the ‘Symmetric Mean Absolute PercentageError’ (Scott Armstrong 1985; Makridakis 1993) as aproxy for the the volatility of labour demand for differentoccupations (see the technical section in the SupplementalMaterial for more details).Fig. 5 shows the prediction performance for three occupa-tions (‘Journalists’, ‘Data Scientists’, ‘Data Analysts’) and
Prepared using sagej.cls awson et al. S M APE
Journalists DataScientists DataAnalysts All Australian job postings
Error in forecasting posting frequency
Figure 5. (a)
Predictability comparison of temporal postingfrequency highlighting the difficulties of predicting journalism jobads. for the volume of ‘All Australian job postings’. We use asliding window approach to obtain multiple predictions (seethe Supplemental Material) that we aggregate as boxplots.The higher the error score on the vertical axis, the lowerthe predictive abilities for that occupation. As Fig. 5 reveals,predicting the daily posting frequency of journalism jobs isconsistently more difficult than for the other occupations,and the market as whole. ‘Data Scientists’, an occupationundergoing strong relative growth, is also showing a highprediction error compared to the market as a whole, indica-tive of experiencing a degree of volatility. However, it isnot nearly commensurate to the predictive difficulties, andvolatility, of journalists. This was true from 2012 to 2019,and has become worse in 2020 with the spread of COVID-19.
Gender
There have been growing gender differences of employedjournalists in Australia since 2014. Fig. 6a shows that theratio of female employed journalists has increased relativeto male journalists (ANZSCO Unit Level) (AustralianBureau of Statistics 2019a). In 2014, the female-to-maleemployment ratio was 0.7. In 2018, the proportion more thandoubled, with almost 1.8 female journalists employed forevery male journalist. It has since declined in 2019 to 1.35,but this proportion is still almost double that of 2014.Fig. 6b also shows that wage inequality between femaleand male journalists has worsened (Australian Bureauof Statistics 2019b). Since 2014, the annual salaries forfemale journalists increased by only AU$3,000, whereasannual salaries for male journalists increased by morethan AU$30,000. Male journalists thus experienced anaverage wage growth that was ten times greater than femalejournalists from 2014 to 2018.There are also changing age demographics of employedjournalists during the studied period. The markers on Fig. 6bhighlight the average age of journalists by gender, peryear. Male journalists have been getting older, their averageage increasing by two years from 2014 to 2018. Femalejournalists, however, have been steadily getting younger. Theaverage age for female journalists decreased by more thanfour years from 2014 to 2018.
Location
Fig. 7 plots the location and volume of employed journalistsin Australia. Fig. 7a shows the absolute and relative numberof job ads posted for each of the capital cities, andoutside them, and Fig. 7b shows the location of employedjournalists per state. Unsurprisingly, Sydney and Melbourne,the respective capital cities of New South Wales (NSW) andVictoria (VIC), consistently have the highest job ad postingfrequencies. However, the relative share of job ad postingfrequency in Australian capital cities has shrunk in recentyears, with Fig. 7a showing an increase outside of majorcities, both in relative and absolute terms. This trend reacheda peak in 2017, when less than 50% of all journalist jobads were for positions inside capital cities. A small reboundfollowed, and in 2019 Sydney commanded approximatelyone-third of all journalism job ads.
Education & Experience
Figs. 8a and 8b show respectively the number of years offormal education required for journalists, and the experiencerequirements (both per year, extracted from job ads data)The education requirements consistently remained at marketaverage levels, with journalists required to possess aBachelor-level degree (approximately 16 years of education).By contrast, the experience requirements have been morevariable. Since 2012, employers have required fewer yearsof experience from journalists than is required in the marketgenerally. However, the gap is narrowing. In 2018, employersdemanded of journalists, on average, one additional yearof experience compared to 2014. This counters the generalmarket trend of employers demanding less experience ofprospective employees.
Employment Type
Casual and temporary work have become more common-place in Australia (Gilfillan 2018), and we study if this isalso the case for Australian journalism jobs. In Fig. 9 we plotthe number of permanent and temporary journalism jobs, percalendar year. The number of ‘Temporary’ journalism jobshas increased in absolute terms since 2012, and they havemade up the majority of all journalism ads in every year. Itis noteworthy too that the share of ‘Permanent’ journalismvacancies has also increased since 2012. However, this trendshould be interpreted with a degree of scepticism as only ∼
50% of all journalism job ads specify whether the rolesadvertised are permanent or temporary.
Journalism Skills
Growing demand for journalism skills.
Here we performa detailed analysis of the top 50 journalism skills that weidentify for each year from 2014 to 2018 (see
Data &Methods for details, and the Supplemental Material for thetop 50 skills for each year). We calculate (and show asstacked bar charts in Fig. 10a) the posting frequency ofthree of the fundamental journalism skills within job ads: (1)‘Journalism’, (2) ‘Editing’, and (3) ‘Writing’. These skillsare counted across all job ads in Australia, regardless of theiroccupational class. While Fig. 4 shows that labour demandfor journalists has decreased since 2016, Fig. 10a presents
Prepared using sagej.cls
Journal Title XX(X)
Male Journalists Employed Female Journalists Employed Female-to-Male Ratio N u m be r E m p l o y ed F e m a l e - t o - M a l e R a t i o Gender: employment levels (a)
Male Journalists Avg SalariesFemale Journalists Avg SalariesMale Journalists Avg AgeFemale Journalists Avg Age A v e r age A nnua l S a l a r i e s ( A UD ) A v e r age A ge ( Y ea r s ) Gender: salaries (b)
Figure 6. Journalist employment levels and salaries by Gender : (a) Since 2015, the employment ratio of female-to-malejournalists has increased; (b)
Wage inequality appears to be increasing between males and females in the ‘Journalists & OtherWriters’ Unit group. This is at the same time that the average age of journalists has been decreasing for females and increasing formales since 2014.
Sydney Melbourne Brisbane Canberra & ACT Perth Other (a)
NSWQLD ACTTASVICWA SANT © 2020 Mapbox © OpenStreetMap
Journalists Employed 2019 (000s) (b)
Figure 7. Location of journalists in Australia : (a) Postingfrequency for journalism jobs decreased in major Australiancities, in relative terms; (b)
As of 2019, the majority ofjournalists in Australia are employed in New South Wales,Victoria, and Queensland, respectively. the more nuanced story that the posting frequency for eachof these core journalism skills has increased every year from2012 to 2018.The relative rankings of these three skills have alsoincreased. For each year, we count the posting frequencyof each unique skill that appears in job ads. We then rank these skills by posting frequency as a proxy for labourdemand. Fig. 10b shows that the rankings of all three of thesefundamental journalism skills have improved from 2012 to2018. In other words, not only has the posting frequency ofthese three journalism skills increased in job ads, but theirimportance relative to all other skills has also increased.
Changing importance of journalism skills.
Here, weaim to determine whether a change occurred in the relativeimportance of the core journalism skills over time. Giventhe dynamics of skill requirements in job ads, skills canbecome increasingly more (or less) similar over time. We usethe similarity measures in Eq. (1) to identify the skills thatare becoming more relevant to being a journalist. Fig. 11ashows the changes in similarity scores between the skill‘Journalism’ and each of the eight other top journalism skills(as per the top yearly journalism skills lists). The greaterthe area covered in the radar chart, the greater the similarityscore, with the blue area representing 2014 and the redarea 2018. Visibly in Fig. 11a, ‘Social Media’ related skillsare becoming increasingly relevant for journalists, with therelative ratio of more traditional skills such as ‘Editing’ and‘Copy Writing’ diminishing with respect to ‘Social Media’,from 2014 to 2018.
Occupations that require journalism skills.
Here westudy which are the occupations that require most journalismskills, and their dynamics over time. Given the yearly listsof top journalism skills (described in
Skill Similarity ),we use Eq. (2) to determine the occupations with thehighest intensities of journalism skills, for each year from2014 to 2018. Intuitively, this allows us adaptively toidentify occupations that become more or less similarto ‘Journalism’, based on their underlying skill usage. Italso provides a means to assess likely transitions betweenoccupations, as workers are more likely to transition tooccupations where the underlying skill requirements aresimilar (Bechichii et al. 2018). Higher similarity lowers thebarriers to entry from one occupation to another.Fig. 11b highlights eight top occupations and their jour-nalism skill intensity scores for 2014 and 2018. ‘Reporter’,‘Editor’, and ‘Copywriter’ cover the highest percentage ofjournalism jobs in the dataset, respectively. While the jour-nalism skill intensities of these occupations were relatively
Prepared using sagej.cls awson et al. Journalists Y ea r s o f E du c a t i on R equ i r ed ( a v e r age ) All Jobs
Minimum years of education (a)
Journalists All Jobs
Minimum years of experience Y ea r s (b) Figure 8. (a)
Years of Education demanded by employers from job ads are consistent with the market average ; (b)
Years ofExperience required by employers have consistently remained below the market average, according to job ads. However, this gaphas closed considerably since 2014.
37 51 53 52 6929 57 52 67 82 100 81
Permanent N u m be r o f J ob s Journalism Jobs: permanent vs. temporary
Temporary
Figure 9.
Temporary positions represent the majority ofjournalism job ads in Australia. However, the proportion ofpermanent positions have been increasing according to job adsdata.
Yearly Rank of Journalism Skills
JournalismEditingWriting
Demand for journalism skills (a) (b)
Figure 10.
The absolute posting frequency (a) and relativeyearly rank (b) of three major journalism skills have increasedbetween 2012 and 2018. high in 2018, their growth since 2014 was relatively low.In comparison, ‘Photography’, ‘Communications’, ‘SocialMedia’, and ‘Public Relations’ experienced higher journal-ism skill intensity growth from 2014 to 2018. This providesevidence as to where workers with journalism skills might befinding employment outside of journalism.
Discussion
Volatility of journalism jobs
Drawn from job ads and employment statistics, our findingsreveal the highly volatile nature of the journalism industry.Compared to other industries, journalism experiencesdramatic fluctuations that are unpredictable and irregular.The data also confirms that journalism is an industry incrisis, particularly since the spread of COVID-19 (seebelow). However, the data also reveals surprises, includingthat the number of journalism jobs ads and employmentlevels increased from 2012 until 2016. Since then, though,journalism jobs in Australia have been in decline.The volatility of journalism jobs in Australia is clearlyapparent in Posting Frequency & Employment levels.Posting frequency of job ads have ranged from near zerolevels in 2012 and 2014 to more than 200 posts per quarter in2016. These violent swings are also apparent in the quarterlyemployment statistics of ‘Journalists and Other Writers’.Following the mass redundancies of 2012, employmentlevels plummeted, reaching their lowest levels in 2013.They have since increased. However, the data confirms thatvolatility of employment has been a constant for journalism,and that this has worsened during COVID-19.Fig. 5 reveals this extreme volatility. The error metricsfrom the Machine Learning model used to predict dailyposting frequencies of job ads (as detailed in TrendAnalysis & Predictability) highlight the difficulties ofmaking predictions about journalism employment. This lackof predictability is indicative of volatility. The higher theerror scores for a given occupation, the higher the likelihoodthat the occupation is experiencing significant disruption.This becomes apparent when we compare journalism toother occupations. For example, the volatility of ’Journalists’dwarfs that of ‘Data Scientists’, an occupation experiencingsignificant demand and volatility in Australia (Dawson et al.2019).The volatility of journalism jobs is further revealed bya time series analysis of journalism compared to otheroccupations (Fig. 4), a gender-based analysis (Fig. 6),a geographical analysis (Fig. 7) and an analysis of thetemporary nature of journalism jobs (Fig. 9).
Prepared using sagej.cls Journal Title XX(X)
EditingCopy WritingPublic RelationsSocial MediaSocial Media PlatformsSocial Media Tools Social Media Strategy Social Content . . . . . . Importance of journalism skills (a)
EditorCommunications CoordinatorCommunications / Public Relations ManagerCopywriterPhotographerReporter Public Relations / Communications Specialist Social Media Specialist
20 25 30 35 40 45 50 55 60
Occupations with journalism skills (b)
Figure 11. Skill and occupational similarity analyses : (a) The changing similarity (or relative importance) of specific skillscompared to the skill ‘Journalism’; (b)
Eight occupations that have the highest similarity to the ‘Top Yearly Journalism Skills’.
What is indisputably clear is that the advertising marketfor news and journalism has collapsed, and continues tocollapse. Meanwhile, consumers have consistently shownan unwillingness to pay for digital journalistic content. In2019, Australian news consumers admitted they would muchwould rather subscribe to a video streaming service suchas Netflix (34%), than pay for online news (9%) (DNRAustralia 2019). The Internet has detonated the advertisingmodel that once sustained journalism, and simultaneouslyre-adjusted consumer expectations on the monetary valueof journalism content. The fact that journalism is strugglingis confirmed in several ways by the data, including by theunpredictability of job ads posting frequency and the clearshifts in employment levels, as shown in Fig. 1. To say thatjournalism is being disrupted is an understatement.
Volatility exacerbated by COVID-19.
In a fragmentingnews ecosystem, consumer demand for news and journalismis difficult to quantify. The
Digital News Report: Australia2019 has found that many consumers are disengaging,with the proportion of Australians avoiding news increasingfrom 57% in 2017 to 62% in 2019 (Fisher et al. 2019).Demand for ’quality’ and ’public interest’ journalism iseven harder to quantify, given ongoing debates as to whatexactly constitutes ’quality’ and ’public interest’ (Wildinget al. 2018). Nonetheless, demand for journalism has surgeddramatically since the outbreak of COVID-19. The ironyof the coronavirus pandemic is that even as it has beenkilling off journalism jobs, it has also created a heighteneddemand for, and appreciation of, journalism among thegeneral public. As news analyst Doctor (2020) wrote of theUS situation in late March, ’The amount of time Americansspend with journalists work and their willingness to pay forit have both spiked, higher than at any point since Election2016, maybe before ... [but] how many journalists will stillhave jobs once the initial virus panic subsides?’. In the UKin March,
The Guardian received 2.17 billion page views, anincrease of more than 750 million above its previous record,set in October 2019 (Bedingfield 2020). Since the outbreak of COVID-19, the volatility of thejournalism jobs market has worsened dramatically. Wenoted above that in early April News Corp suspended thepublication of 60 newspapers nationally. Then, on April14, Australian Community Newspapers, which publishes170 community titles, said it was suspending publicationof some of its non-daily newspapers; as a result, fourprinting presses were closed and an unspecified number ofstaff were stood down (Meade 2020a). The following day,the federal government announced a $50million package tosupport public interest journalism across TV, newspapers andradio in regional and remote Australia (Hayes and Rubbo2020). And on April 20, the government announced thatdigital platforms including Google and Facebook would beforced to pay for content as the internet advertising businesswould be overhauled to help local publishers survive theeconomic fallout of the coronavirus crisis (Crowe 2020). Thescheme, which would involve a mandatory code imposed ondigital giants, would potentially set a global precedent. Thecombined and ongoing impact on journalism jobs of thesesudden, cumulative developments are hard to predict, but willno doubt be profound.
Gender Wage Gap
At first glance, the data seems to suggest that gender equity isfinally arriving in Australia for journalism - an industry thathas traditionally been male-dominated - as more women thanmen are employed. As the data shows, in 2014 there were 0.7female journalists employed for every male Journalist, butby 2018 the proportion of female-to-male employment morethan doubled, with almost 1.8 female journalists employedfor every male Journalist. It then declined in 2019 to 1.35, aproportion still almost double that of 2014.However, further detail reveals that equity remainselusive. Specifically, wage inequality has worsened. Since2014, annual salaries for female journalists increased byAU$3,000, compared with an increase for male journalists ofover AU$30,000 over the same period. From 2014 to 2018,
Prepared using sagej.cls awson et al. average wage growth for Male journalists was more thanten times greater than for female journalists. Meanwhile, theaverage male Journalist has been getting older, while theaverage female Journalist has been getting younger. In 2014,the average age for a Journalist, whether male or female, wasroughly the same: late 30s. By 2018, the average age for amale journalist was 42, whereas for a female journalist it was34.The potential impacts of this worsening disparity areconcerning. It is likely that senior positions responsible formajor editorial decisions are increasingly being dominatedby men, whereas junior roles are being filled by women whoare younger and worse-paid. This may be having a flow-oneffect as to which news stories are being covered, and howthose stories are being covered. In other words, the gendergap and age gap may be having an impact on the content ofthe news. Further research is needed into related issues of theindustrys composition, including, for instance, the ethnicityof journalists. A vast body of literature exists regarding theimportance of diversity in news (Rodrigues and Paradies2018; Budarick and Han 2017). Further work is needed intodiversity (and its various sub-categories), and what effectdiversity has, for instance, on the proportion of people whoare actively avoiding the news. Location
As discussed above, the sustained pressures on regional andlocal journalism have led to a worrying growth of newsdeserts in countries including Australia and the US. Thistrend has been accelerating alarmingly since the outbreak ofCOVID-19, leaving many areas without any regional or localnews coverage. Hence we might assume that journalism jobsin regional and local areas have been drying up, and that anever-increasing proportion of journalism jobs are in urbancentres.The data, however, is not so clear. As Fig. 7a shows,in 2012 fewer than a quarter of Australias journalism jobads were for jobs outside Sydney, Melbourne, Brisbane,Canberra and the ACT or Perth. In every subsequent year, theproportion of job ads for journalism positions outside theseurban centres has been considerably higher. The peak camein 2017, when nearly half of all job ads were for positionsoutside the major cities. Does this suggest that in 2017 therewere as many jobs for journalists in the regions as in thecentres? Surely not. The explanation, we would suggest, liesin various factors. These include that regional journalismjobs are hard to fill, perhaps because they offer relativelylow salaries, and are hence re-advertised. It is also possiblethat there is a high turnover for some regional positions. Inshort, the job ads data may simply be an indication that thejournalism industry is even more volatile in the regions thanin major urban centres.Research consistently and emphatically reveals thatregional and local journalism are suffering, with anincreasingly bleak prognosis of cuts and closures. While thedata shows a surprisingly high proportion of journalism jobads for positions outside the main metropolitan centres, thiscannot be taken to suggest that journalism is holding steadyin these areas.
Evolving journalism skills
Skills are the building blocks of jobs and standardisedoccupations. In this regard, occupations can be characterisedas ‘sets of skills’. Intuitively, skills that are similar can beinterpreted as complementary when they are paired togetheror relatively easy to acquire (in either direction) when oneskill is already possessed.This intuition provides insight into how journalism skillsare evolving and where journalists might be finding alternatecareer paths. As Fig. 1 shows, both the demand for andsupply of journalists have been declining in Australia since2016. Therefore, a growing number of former journalists,who presumably possess an assortment of journalism skills,have needed to transition between occupations to find newwork. There are, however, significant transition costs movingbetween jobs (Bechichii et al. 2018; Bessen 2015). Thesecosts can come in the form of education, training, physicallymoving for new employment and other barriers. To reducethe friction of these transition costs, workers tend to leveragetheir extant skills, in concert with acquiring new skills, tomake career transitions.As seen in Fig. 11a, the skill ‘Journalism’ has becomemore similar to ‘Social Media’ and more ‘generalist’communications skills. After applying the
Skill Intensity formula from Eq. (2), we identified the top occupationswith highest intensities of journalism skills from 2014-2018. The Fig. 11b chart reinforces that top journalismskills are becoming more important to other occupations,such as ‘Photographers’, ‘Social Media Strategists’, ‘PublicRelations Professionals’, and ‘Communications Specialists’.From the data, we suggest, three conclusions can bedrawn. First, to be hired, journalists are required to havea wider array of skills, such as photography and socialmedia aptitude. Second, jobs in journalism are increasinglyjobs in social media, generalist communications, and publicrelations rather than in reporting and editing. And third, wesee hints as to where onetime journalists are finding alternatecareer paths. As employment conditions progressivelyworsen, journalists are seemingly pursuing new careers inthe occupational areas seen in Fig. 11b, such as photographyor public relations.At a time of great uncertainty, with employment prospectsdeteriorating, it is no wonder that journalists look beyondtraditional journalism for their futures. For society, however,the implications are significant. In this time of economicuncertainty and polarising politics, the people who possessthe journalism skills required to keep the public informedand hold leaders to account are, in many cases, employingtheir talents elsewhere. This places enormous strain on thehealth and quality of journalism in Australia.
Conclusion
The data reveals a contradiction: demand for journalismskills has increased at the same time that demand andemployment for journalists has declined. Indeed, this isone of several contradictions in a volatile industry. Foran increasing number of news media organisations, asustainable business model remains elusive.Our findings give a clearer outline of the problem.Unfortunately, the solutions remain less clear. Quality
Prepared using sagej.cls Journal Title XX(X) journalism is expensive. Good reporting is often slow andlaborious, fixed to the unfolding story. What is required ofquality journalism is, therefore, at odds with the prevailingemployment conditions.This paper highlights the stresses experienced byjournalism in Australia by analysing jobs data. We observethe volatility and downward trajectory of the occupation bothin job ads and employment statistics. These unfavourableemployment conditions are being worsened by the unfoldingCOVID-19 crisis. Our longitudinal analysis also yieldsimportant findings regarding gender inequity. While womenare representing a greater share of employed journalists, theyare earning less, and the wage gap is growing.Further, this paper has also identified top journalismskills. Adopting a data-driven method, we described whichskills are most similar to ‘Journalism’. We then usedthese yearly skill sets to adaptively similar occupations.This enabled us to quantitatively show that the skilldemands of journalists are becoming similar to those of‘Social Media Strategists’, ‘Public Relations Professionals’,‘Communications Specialists’, and others. This suggestswhere people with journalism skills are likely findingalternate career paths, but also raises a related concern. Onthe face of it, the journalism jobs data we have analysed doesnot look so bad after all. On reflection, however, it suggeststhat the thinning ranks of ‘journalism’ are populated by fewerjournalists, and more public relations specialists.Future research could compare these results to otherlabour markets to assess the validity of these findings. Forexample, the skill similarity methodology could be applied inother labour markets to compare the resulting top journalismskills in different locations. Additionally, labour demandanalyses could be conducted on occupations most similar tojournalists to better understand the incentives to transition toother vocations.The results from this research both reinforce the well-documented difficulties of journalism in Australia andprovide granular details that isolate and reveal thesechallenges. The implications are global. The hope is thatthese analytical methods and insights can contribute to thehealth and well-being of the Fourth Estate, and hence to thehealth and well-being of society.
Acknowledgements
We would like to thank Burning Glass Technologies for generouslyproviding the data for this research. We would also like to thankGoogle for generously providing cloud computing resources for thisresearch.
References (2020) Governments around the world crack down on journalistscovering COVID-19. https://cpj.org/2020/03/the-torch-governments-crack-down-on-journalists.php . Accessed: 2020-4-19.Abernathy PM (2018) The expanding news desert. Technical report,University of North Carolina at Chapel Hill.ACCC (2019) Digital platforms inquiry - final report. Technicalreport, Australian Competition and Consumer Commission. Alabdulkareem A, Frank MR, Sun L, AlShebli B, Hidalgo C andRahwan I (2018) Unpacking the polarization of workplaceskills.
Sci Adv . Accessed: 2019-8-1.Australian Bureau of Statistics (2019a) 6291.0.55.003 - LabourForce, Australia, Detailed, Quarterly, May 2019.Australian Bureau of Statistics (2019b) 6306.0 - EmployeeEarnings and Hours, Australia, May 2018.Bechichii N, Grundkei R, Jameti S and Squicciarini M (2018)Moving between jobs: an analysis of occupation distances andskill needs. Technical Report 52, OECD.Bedingfield W (2020) Coronavirus news fatigue is real and it couldbecome a big problem. . Accessed:2020-4-22.Benton J (2020) On a rough day for american newspapers, investorsaren’t buying gannett’s story and tribune’s not done chopping .Bessen J (2015)
Learning by Doing: The Real Connection betweenInnovation, Wages, and Wealth . Yale University Press.Blake A (2019) Dynamics of data science skills. Technical report,The Royal Society.Boseley M and Knaus C (2020) Australia’s coronavirus socialdistancing rules explained: state by state guidelines.
TheGuardian .Budarick JN and Han GS (2017)
Minorities and Media:: Producers,Industries, Audiences . Palgrave Macmillan.Carnevale A, Jayasundera T and Repnikov D (2014) Understandingonline job ads data. Technical report, Georgetown University.Carville O, Court E and Brown KV (2020) Hospitals tell doctorsthey’ll be fired if they speak out about lack of gear.
BloombergNews .Crerar S (2020) Coronavirus and the crisis in regionalnews. . Accessed: 2020-4-27.Crowe D (2020) ’a level playing field’: digital giants willhave to pay for news. . Accessed: 2020-4-20.Dawson N and Rizoiu MA (2020) Coronavirus infectingaustralian jobs: vacancy rates down since early february.
TheConversation
URL https://theconversation.com/coronavirus-infecting-australian-jobs-vacancy-rates-down-since-early-february-134234 .Dawson N, Rizoiu MA, Johnston B and Williams MA (2019)Adaptively selecting occupations to detect skill shortages fromonline job ads. In: . pp. 1637–1643.Dawson N, Rizoiu MA, Johnston B and Williams MA (2020)Predicting Labor Shortages from Labor Demand and LaborSupply Data: A Machine Learning Approach. URL http:
Prepared using sagej.cls awson et al. //arxiv.org/abs/2004.01311 .Department of Employment, Skills, Small and Family Business(2019) Sixty per cent of job vacancies in australia areadvertised online. . Accessed:2019-7-7.Doctor K (2020) Newsonomics: What was once unthinkableis quickly becoming reality in the destruction of localnews. . Accessed: 2020-4-19.Fenton N (2011) Deregulation or democracy? new media, news,neoliberalism and the public interest. Continuum . Accessed:2020-4-19.Flynn K (2020) Coronavirus is speeding up the collapse of localnewsrooms.
CNN .Gardiner A, Aasheim C, Rutner P and Williams S (2018)Skill requirements in big data: A content analysis of jobadvertisements.
Journal of Computer Information Systems
Business Insider .Guo L and Volz Y (2019) (re) defining journalistic expertisein the digital transformation: A content analysis of jobannouncements.
Journalism Practice
ABC News .Hidalgo CA, Klinger B, Barab´asi AL and Hausmann R (2007) Theproduct space conditions the development of nations.
Science
Journalism : 1464884918800077.Khadem N and Pupazzoni R (2020) AAP newswire service closesafter 85 years with 500 job losses, including 180 journalists.
ABC News .Kuo L (2020) ‘they’re chasing me’: the journalist who wouldn’t stayquiet on covid-19.
The Guardian .Makridakis S (1993) Accuracy measures: theoretical and practicalconcerns.
International Journal of Forecasting
The Guardian .Meade A (2020b) News corp australia warns of coronavirus crisisjob cuts as smaller regional papers close.
The Guardian .Nieman-Lab (2020) About nieman lab. . Accessed: 2020-3-10.OECD (2019) OECD skills strategy 2019 - skills to shape a betterfuture. Technical report, OECD.Ricketson M, Dodd A, Zion L and Winarnita M (2020) “likebeing shot in the face” or “i’m glad i’m out”: Journalists’experiences of job loss in the australian media industry 2012–2014.
Journalism Studies
Journal of Intercultural CommunicationResearch
Long-Range Forecasting: From CrystalBall to Computer . 2 edition edition. Wiley-Interscience.Sherwood M and O’Donnell P (2018) Once a journalist, always ajournalist?
Journalism Studies
Urban Stud.
The Guardian .Taylor SJ and Letham B (2018) Forecasting at scale.
The AmericanStatistician
Weltwirtsch. Arch.
The Atlantic .Wilding D, Fray P, Molitorisz S and McKewon E (2018) The impactof digital platforms on news and journalistic content. Technicalreport, University of Technology Sydney.Young S and Carson A (2018) What is a journalist? the view fromemployers as revealed by their job vacancy advertisements.
Journalism Studies https://euobserver.com/coronavirus/147956 . Accessed:2020-4-19.Zion L, Sherwood M, O’Donnell P, Dodd A, Ricketson M andMarjoribanks T (2016) ‘It has a bleak future’: The effects ofjob loss on regional and rural journalism in Australia .Zion L, Sherwood M, O’Donnell P, Marjoribanks T, RicketsonM, Dodd A and Winarnita M (2018) New beats report:mass redundancies and career change in australian journalism.Technical report.
Prepared using sagej.cls Journal Title XX(X)
Contents (Appendix)
Technical Appendix 14
Data Sources . . . . . . . . . . . . . . . . . . . . 14Skill Similarity . . . . . . . . . . . . . . . . . . . 15Trend Analysis & Predictability . . . . . . . . . . 16Quantify Labour Demand Volatility . . . . . . . . 16Top Journalism Skills by Year . . . . . . . . . . . 16This document is accompanying the submission
Layoffs,Inequity and COVID-19: A Longitudinal Study of theJournalism Jobs Crisis in Australia from 2012 to 2020 . Theinformation in this document complements the submission,and it is presented here for completeness reasons. It isnot required for understanding the main paper, nor forreproducing the results.
Technical Appendix
Here, we describe the data sources we used to analysejournalism jobs. We also outline the skill similaritymethodology that enables us to construct temporal (yearly)sets of top journalism skills. Lastly, we describe how thesetemporal sets of top journalism skills then allow us toadaptively identify occupations that are ‘most similar’ tojournalism, at the granular skill level.
Data Sources
Journalism job ads.
This research draws on more than6.7 million Australian online job ads from 2012-01-01until 2019-02-28, courtesy of data provided by BurningGlass Technologies ‡ (BGT). BGT also granted access to theaggregated job ads data from 2019-03-01 to 2020-03-31,allowing us to address the early impacts of the unfoldingcoronavirus pandemic (COVID-19) on journalism jobs inAustralia. BGT collected the job ads data via web scrapingand systematically processed it into structured formats. Thedataset consists of detailed information on individual job ads,such as location, salary, employer, educational requirements,experience demands, and more. The skill requirements havealso been extracted (totalling > , unique skills) andeach job ad is classified into its relevant occupational andindustry classes. There are two occupational ontologies inthe job ads dataset. The first is ANZSCO, which is theofficial occupational classification standard in Australia andNew Zealand. The other is the BGT occupational ontology,which has been developed due to shortcomings of officialoccupational standards (as described in Related Work &Background).To ensure selection accuracy, we instituted the followingsearch query conditions over the dataset:1. All job ads with ANZSCO Occupation labels of‘Newspaper or Periodical Editor’, ‘Print Journalist’,‘Radio Journalist’, ‘Television Journalist’, and ‘Jour-nalists and Other Writers nec’ (where ‘nec’ stands for‘not elsewhere classified’).2. OR All job ads with the BGT Occupation label of‘Journalist / Reporter’ and ‘Editor’ (the two primaryBGT occupational classes for journalists); 3. OR All job ads with the ‘Journalist’, ‘News’, or‘Editor’ in any part of the job title.After manually reviewing the returned job ad features foraccuracy, the selection process resulted in a sample of 3,231Australian journalism job ads from 2012-01-01 until 2019-02-28. We used the same search query and approach for the2019-03-01 to 2020-03-31 period to supplement this sample.This returned 467 journalism job ads, amounting to a totalof 3,698 journalism job ads from 2012-01-01 to 2020-03-31.The job ads during the period are observed aggregated daily,with limited skill level details. However, much of the analysisthat follows requires access to the features within individualjob ads, so only Fig. 2 leverages the 2020 data. Further details on job ads data.
It is estimated thatapproximately 60% of Australian job ads are postedonline (Department of Employment, Skills, Small andFamily Business 2019). At aggregate levels, online jobadvertisements (ads) provide valuable indicators of relativelabour demands. This includes demand features, such assalaries, educational requirements, years of experience, and,most importantly, skill-level information. Here, a distinctionmust be made between skills, knowledge, abilities, andoccupations. ‘Skills’ are the proficiencies developed throughtraining and/or experience (OECD 2019); ‘knowledge’ isthe theoretical and/or practical understanding of an area;‘ability’ is the competency to achieve a task (Gardiner et al.2018); and ‘occupations’ are standardised jobs that are theamalgamation of skills, knowledge, and abilities used byan individual to perform a set of tasks that are required bytheir vocation. Throughout this paper, the term ‘skill’ willincorporate ‘knowledge’ and ‘ability’. Skills, in this sense,are the constituent elements that workers use to performtasks, which ultimately define jobs and occupations.
Advantages of job ads data.
Understanding how thecomposition of skill sets evolve within an occupationis essential to understanding trends in that occupation.However, occupational data rarely captures skill-level data.Most often, official occupational standards are static, rarelyupdated classifications, which fail to capture the changingskill demands of occupations, or to detect the creation of newtypes of jobs.
Journalist employment statistics.
Employment data(labour supply) were collected from the ‘Quarterly DetailedLabour Force’ statistics by the ABS (Australian Bureauof Statistics 2019a). These employment data are organisedinto standardised occupations called the Australia and NewZealand Standard Classification of Occupations (ANZSCO).ANZSCO provides a basis for the standardised collection,analysis and dissemination of occupational data for Australiaand New Zealand. The structure of ANZSCO has fivehierarchical levels - major group, sub-major group, minorgroup, unit group and occupation. The categories at the mostdetailed level of the classification are termed ’occupations’.A shortcoming, however, is that the lowest level ofoccupational employment data available by the ABS isat the 4-digit Unit level, which is one hierarchical levelabove specific occupations. As our research is focused ‡ BGT is a leading vendor of online job ads data.
Prepared using sagej.cls awson et al. on the employment Unit class of ‘Journalists and OtherWriters’, all ABS employment statistics cited in thisresearch include the following occupations: ‘Copywriter’,‘Newspaper or Periodical Editor’, ‘Print Journalist’, ‘RadioJournalist’, ‘Technical Writer’, ‘Television Journalist’, and‘Journalists and Other Writers nec’. While the inclusionof the ‘Copywriter’ and ‘Technical Writer’ occupations inthese statistics could distort results pertaining to ‘Journalists’to an extent, we consider this impact to be limited inscope. As we describe in Jobs Data Analysis and Results ,the employment statistics highlight important trends injournalism occupations, which are confirmed by findingsfrom the job ads data.Another shortcoming of employment statistics is their‘lagging’ nature. The inertia of labour markets means thatit takes time for changes to materialise in employmentstatistics. Additionally, the official reporting of employmentstatistics takes time. Employment statistics are oftenpublished several months or years after the reported period.As a result, these ‘lagging’ characteristics are not availablefor the most recent periods in our work (such as for thesecond half of 2019 and later.)
Skill Similarity
In this section, we detail the methodology previouslyemployed in (Alabdulkareem et al. 2018; Dawson et al.2019) to dynamically measure skill similarity. Here, wepresent the building blocks for this method, applying it forjournalism related skills and occupations.
Intuition.
Two skills are similar when the two are relatedand complementary, i.e. the two skills in a skills-pair supporteach other. For example, ‘Journalism’ and ‘Editing’ have ahigh pairwise similarity score because together they enablehigher productivity for the worker, and because the difficultyto acquire either skill when one is already possessed by aworker is relatively low.Our goal, therefore, is to calculate the similarity of eachunique skill relative to every other unique skill in the dataset.Such a measure allows us to identify which skills have thehighest pairwise similarities to a specific skill or set of skills.We also want to identify how skill similarity evolves overtime. To achieve this, we have instituted a temporal split of acalendar year. This enables us to assess yearly changes to theunderlying skill demands of journalism jobs.
The Revealed Comparative Advantage of a skill.
We implement a data-driven methodology to measurethe pairwise similarity between pairs of skills that co-occur in job ads. One difficulty we encounter is thatsome skills are ubiquitous, occurring across many jobads and occupations. We address this issue by using the
Revealed Comparative Advantage (RCA), which maximisesthe amount of skill-level information obtained from eachjob ad, while minimising the biases introduced by over-expressed skills in job ads. Formally, RCA measures therelevance of a skill s for a particular job ad j as: RCA ( j, s ) = x ( j, s ) / (cid:80) s (cid:48) ∈S x ( j, s (cid:48) ) (cid:80) j (cid:48) ∈ J x ( j (cid:48) , s ) / (cid:80) j (cid:48) ∈J ,s (cid:48) ∈S x ( j (cid:48) , s (cid:48) ) (3) where x ( j, s ) = 1 when the skill s is required for job j ,and x ( j, s ) = 0 otherwise; S is the set of all distinct skills,and J is the set of all job ads in our dataset. RCA ( j, s ) ∈ (cid:34) , (cid:80) j (cid:48) ∈ J,s (cid:48) ∈ S x ( j (cid:48) , s (cid:48) ) (cid:35) , ∀ j, s , and the higher RCA ( j, s ) thehigher is the comparative advantage that s is consideredto have for j . Visibly, RCA ( j, s ) decreases when the skill s is more ubiquitous (i.e. when (cid:80) j (cid:48) ∈ J x ( j (cid:48) , s ) increases), orwhen many other skills are required for the job j (i.e.when (cid:80) s (cid:48) ∈ S x ( j, s (cid:48) ) increases). RCA provides a method tomeasure the importance of a skill in a job ad, relative tothe total share of demand for that skill in all job ads. Ithas been applied across a range of disciplines, such as tradeeconomics (Hidalgo et al. 2007) (Vollrath 1991), identifyingkey industries in nations (Shutters et al. 2016), and detectingthe labour polarisation of workplace skills (Alabdulkareemet al. 2018).
Measure skill similarity.
The next step is measuring thecomplementarity of skill-pairs that co-occur in job ads. First,we compute the ‘effective use of skills’ e ( j, s ) defined as e ( j, s ) = 1 when RCA ( j, s ) > and e ( j, s ) = 0 otherwise.Finally, we compute the skill complementarity (denoted θ ) asthe minimum of the conditional probabilities of a skills-pairbeing effectively used within the same job ad. Skills s and s (cid:48) are considered as highly complementary if they tend tocommonly co-occur within individual job ads, for whateverreason. Formally: θ ( s, s (cid:48) ) = (cid:80) j (cid:48) ∈ J e ( j, s ) .e ( j, s (cid:48) ) max (cid:32) (cid:80) j (cid:48) ∈ J e ( j, s ) , (cid:80) j (cid:48) ∈ J e ( j, s (cid:48) ) (cid:33) (4)Note that θ ( s, s (cid:48) ) ∈ [0 , , a larger value indicates that s and s (cid:48) are more similar, and it reaches the maximum value when s and s (cid:48) always co-occur (i.e. they never appear separately). Top journalism skills.
Following the procedure outlinedin (Dawson et al. 2019) for building sets of highlycomplementary skills, we use the θ function togetherwith ‘Journalism’ as the ‘seed’ skill to create top yearlylists of journalism skills. More precisely, we compute θ ( Journalism, s ) – i.e. the similarity between the skill‘Journalism’ and each unique skill that occurs during agiven year. Skills on each yearly list are ordered by theirdescending pairwise skill similarity scores. When inspectingthe yearly skill lists, we make two observations. First, theskills in 2012 and 2013 appear of notably lower qualitythan from 2014 onward. We posit that this has to do withimperfect skills extraction methods during the early yearsof the BGT dataset. As a result, we decided to measurethe top yearly journalism skill sets from 2014 to 2018 (thelast available full year of data for which we had access). § Second, we decided to retain only the top 50 skills on eachyearly list. Through qualitative analysis, we determined thatthis threshold of 50 is both sufficiently exclusive for defining § We did not notice a deterioration of quality regarding other features, suchas salaries, education, experience etc. Therefore, these 2012 and 2013 willbe used for parts of the analysis.
Prepared using sagej.cls Journal Title XX(X) journalism skills and reasonably inclusive for detectingthe evolution of new, emerging skills in journalism. Thepurpose of these top journalism skills lists is to capturejournalism labour trends; it is not intended to represent acomplete taxonomy of journalism skills. The yearly listsof top journalism skills, and their similarity scores, can beobserved in the Supplemental Material Sec.
Top JournalismSkills by Year . Compute journalism skill intensity.
For the occupationalsimilarity analysis in Sec.
Journalism Skills , we decided touse the BGT occupational ontology as opposed to ANZSCO.This is because the BGT occupational classes appear morereflective of current job titles. For example, a job titleadvertised for a ‘Social Media Manager’ is classified by BGTas a ‘Social Media Strategist / Specialist’. Whereas the samejob title would be classified by ANZSCO as an ‘AdvertisingSpecialist’ or ‘Marketing Specialist’.
Trend Analysis & Predictability
We use the Prophet time-series forecasting tool developedby Facebook Research (Taylor and Letham 2018). Prophetis an auto-regressive tool that fits non-linear time-seriestrends with the effects from daily, weekly, and yearlyseasonality, and also holidays. The main model componentsare represented in the following equation: y ( t ) = g ( t ) + s ( t ) + h ( t ) + (cid:15) t (5)where g ( t ) refers to the trend function that models non-periodic changes over time; s ( t ) represents periodic changes,such as seasonality; h ( t ) denotes holiday effects; and (cid:15) t is theerror term and represents all other idiosyncratic changes. Quantify Labour Demand Volatility
We evaluate the forecasting performance using a temporalholdout setup. That is, we split the available time-seriesinto a training part (the first part of the sequence) anda testing part (the latter part of the sequence). We trainthe Prophet model on the training part, and we generatejob ad posting forecasts by “running time forward” inEq. (5) for time t in the testing period. Finally, we measurethe accuracy of the forecast against the observed postingvolumes using the Symmetric Mean Absolute PercentageError (SMAPE) (Scott Armstrong 1985; Makridakis 1993).SMAPE is formally defined as: SM AP E ( A t , F t ) = 200 T T (cid:88) t =1 | F t − A t | ( | A t | + | F t | ) (6)where A t denotes the actual value of jobs posted on day t ,and F t is the predicted value of job ads on day t . SMAPEranges from 0 to 200, with 0 indicating a perfect predictionand 200 the largest possible error. When actual and predictedvalues are both 0, we define SMAPE to be 0. We selectedSMAPE as an alternative to the more widely used MAPEbecause it is (1) scale-independent and (2) robust to actualor predicted zero values. To evaluate the uncertainty ofthe forecast, we adopt a ‘sliding window’ approach. Thisconsists of using a constant number of training days (here , days) to train the model, and we test the forecastingperformance on the next days. We then shift both the training and the testing periods right by one day, and theprocess is repeated. Consequently, we train and test themodel 365 times, and we obtain 365 SMAPE performancevalues. Top Journalism Skills by Year
Top journalism skills calculated by skill similarity method-ology in Sec.
Skill Similarity . Prepared using sagej.cls awson et al. Table 1.
Top journalism skills calculated by skill similarity methodology in Skill Similarity
Rank 2014 2015 2016 2017 20181 Journalism Journalism Journalism Journalism Journalism2 Editing Editing Editing Editing Editing3 Media Relations Media Relations Copy Writing Content Management Content Management4 Corporate Communications Copy Writing Media Relations Social Media Media Relations5 Copy Writing Content Management Content Management Copy Writing Copy Writing6 Content Management Copywriting Social Media Media Relations Social Media Platforms7 Public Relations Social Media Social Media Platforms Corporate Communications Social Media8 Social Media Public Relations Copywriting Social Media Platforms Content Development9 Content Management Systems (CMS) Social Media Platforms Corporate Communications Content Development Corporate Communications10 Multimedia Corporate Communications Public Relations Social Content Public Relations11 Copywriting Content Development Content Development Public Relations Social Media Tools12 Content Development Content Management Systems (CMS) Social Media Tools Copywriting Copywriting13 Strategic Communications Strategic Communications Digital Marketing Facebook Content Management Systems (CMS)14 Facebook Social Media Tools Online Marketing Strategic Communications Social Content15 Social Media Platforms Multimedia Multimedia Social Media Tools Strategic Communications16 Marketing Communications Facebook Strategic Communications Marketing Communications Social Media Strategy17 Media Coverage Marketing Communications Market Research Content Management Systems (CMS) Content Marketing18 Publicity Social Content Marketing Communications Multimedia Facebook19 Proofreading Digital Communications Content Management Systems (CMS) Proofreading Digital Communications20 Social Media Tools Publicity Writing Content Marketing Media Coverage21 Digital Communications Media Production Content Marketing Digital Journalism Publicity22 Crisis Management Social Media Strategy Photography Digital Communications Proofreading23 Adobe Photoshop Communications Programmes Instagram Publicity Multimedia24 Communications Programmes Media Coverage Publicity Media Coverage Instagram25 Digital Journalism Internal Communications Digital Communications Digital Marketing Video Production26 Community Relations Content Marketing Media Coverage Writing Marketing Communications27 Photography Proofreading Social Content Video Production Adobe Photoshop28 Social Media Strategy Writing Social Media Strategy Graphic Design Content Curation29 Graphic Design Adobe Photoshop Media Production Media Production Video Editing30 Youtube Brand Awareness Generation Proofreading Communications Programmes Adobe Indesign31 Media Strategy Adobe Indesign Facebook Instagram Adobe Creative Suite32 Brand Management Marketing Materials Event Planning Social Media Strategy Adobe Acrobat33 Web Content Management Digital Marketing Adobe Photoshop Video Editing Brand Awareness Generation34 Adobe Indesign Video Editing Meeting Deadlines Adobe Photoshop Adobe Illustrator35 Social Content Adobe Creative Suite Self-Starter Self-Starter Google Analytics36 Marketing Materials Adobe Acrobat Marketing Breaking News Coverage Press Releases37 Event Planning Graphic Design Creativity Creativity LinkedIn38 Digital Marketing Video Production Adobe Indesign Adobe Illustrator Digital Marketing39 Writing Instagram Adobe Creative Suite Event Planning Digital Journalism40 Instagram LinkedIn Adobe Illustrator Adobe Indesign Media Production41 Online Research Media Strategy Community Relations Adobe Creative Suite Communications Programmes42 Adobe Acrobat Photography Adobe Acrobat Adobe Acrobat Crisis Management43 LinkedIn PR Agency Press Releases Promotional Materials Media Strategy44 Google Analytics Meeting Deadlines Internal Communications Photography Writing45 Video Editing Digital Journalism Campaign Management Content Curation Photography46 Website Production Event Planning Creative Writing Marketing Blog Posts47 Proofing Google Analytics Video Production Meeting Deadlines Internal Communications48 Video Production Media Campaigning Blog Posts Google Analytics Event Planning49 Media Planning Press Releases Crisis Management Media Strategy Creative Problem Solving50 Campaign Management Crisis Management Youtube Business-to-Business Creativity