‘Social Workers Failed to Heed Warnings’: A Text-Based Study of How a Profession is Portrayed in UK Newspapers

Previous research has indicated that social workers are portrayed negatively in the UK press, particularly in child protection cases. But what is the nature of this negativity? And are social workers also mentioned in more positive contexts? To explore these questions, a collection of three months of newspaper articles was compiled (early May to early August 2019), using the seed term ‘social worker(s)’. Almost 1,000 occurrences were located and categorised as ‘positive’ (6 per cent of instances), ‘negative’ (25 per cent) or ‘neutral’ (69 per cent). Further classification of negative instances indicates these concern social workers’ perceived failure to act rather than perceived over-zeal-ous behaviour (ratio 5:2). Findings also suggest that the press tend to hold social workers to a higher moral standard in their everyday lives than is the case for other members of society. Understanding how social workers are portrayed in the press is important for practitioners in terms of recruitment, job satisfaction and retention. In addition, as newspapers are the major source of information on social work for members of the public, greater awareness could reduce the societal tendency towards finding individuals to blame.


Introduction
Previous research indicates that social work as a profession and by extension, the social workers who carry out this work are frequently portrayed negatively in the UK press, particularly in child protection cases (e.g. Ayre, 2001;Reid & Misener, 2001;Warner, 2014;Hughes & Houston, 2019). Social workers are in the news when a case goes wrong, particularly when a child dies (Parton, 2014), and are often subjected to a 'blame culture' by the press and the public (Munro, 2010(Munro, , 2011(Munro, , 2019, whereby a person or persons are held accountable when something goes wrong. But what is the nature of this negative portrayal? And are social workers also mentioned in more positive contexts? Whilst previous studies of the portrayal of social work or social workers in the press have focused on negativity within news stories as 'whole texts' (Warner, 2014;Hughes and Houston, 2019), the current study is concerned with the 'language' surrounding the professional category of 'social worker' across newspaper texts. Epistemologically, the study takes a social constructionist approach in exploring how press discourse surrounding the search term 'social worker(s)' constructs, reinforces and extends the wider societal view of the profession (cf. studies in Taylor and Marchi, 2018). To explore this, a database of three months of UK newspaper articles featuring 'social worker(s)' was compiled and investigated both quantitatively and qualitatively. The collection of texts (or 'corpus') was examined using corpus linguistic software as this allows the analyst to slice through the data in different ways and thus obtain fresh perspectives whilst retaining the wider context of individual articles.
The way in which social workers are portrayed in the press influences both how the public see these professionals (this may be the primary source of information for many people) and social workers' own self-esteem and pride in their profession, thus affecting both recruitment and retention (Reid and Misener, 2001;Legood et al., 2016). Press portrayal of social work is of particular importance as, according to a Unison survey, 56 per cent of social workers are considering leaving their role (Community Care, 2019; see also Ayakwah and Cooper, 2019). Exploring the extent to which press depictions of social workers are negative, and analysing the form this negativity takes through language analysis, have the potential to help journalists and the general public better understand the challenges facing the profession and reduce the tendency towards blame.
The study aims to provide a snapshot of how social workers are depicted and seeks to answer the following research questions: (i) What are the proportions of positive, negative and neutral mentions of social worker(s) in the UK press? (ii) What are the dominant subcategories within each of these three groupings? The article first explores relevant literature on press coverage of social work, making the case for this study, then discusses the value of the corpus linguistic methodology adopted before providing first an overview of quantitative findings for each category followed by discussion of the more detailed classifications.

Relevant literature
Negative press coverage in children's social work can be traced back at least as far as the 1973 death of Maria Colwell (e.g. Warner, 2014;Parton, 2014), leading to a climate of fear, blame and mistrust throughout the 1970s, 1980s and 1990s (Ayre, 2001) and was particularly prevalent in the highly publicised case of seventeen-month-old Peter Connelly ('Baby P') in 2007. The social workers and manager involved in Peter Connelly's case suffered '[i]ntensely hostile media coverage and political reaction' and ultimately lost their jobs' (Warner, 2014(Warner, , p. 1638. Arguing that politicians and the press have a 'shared mutual interest in the co-authorship of "bad" stories about social work', Warner (2013Warner ( , p.1637 points to the newsworthiness of the Peter Connelly case serving the purposes of the press whilst enabling politicians to publicly take a moral stance and thus increase their own profile. Following the Peter Connelly case and the ensuing intense media discussion, social work became increasingly risk-averse, with more referrals, child protection plans, care proceedings and ultimately an increase in the number of children taken into care, leading Parton (2014Parton ( , p. 2042 to claim that social work 'has been reduced to a very narrow concern with child protection'. Parton's (2014) overview of research studies examining how social workers are presented in the UK asserts that there are far more negative than the positive stories and that these negative stories are largely within child protection. Media magnification of perceived 'moral panics' in society, first discussed in Cohen's (1972) seminal work, contribute towards the negative perception of social workers, particularly in child protection, and the 'blame culture' around social work. For example, Franklin's (1998) analysis of almost 2,000 UK newspaper articles of social work and social services revealed these to be 'overwhelmingly negative and critical' with 67 per cent of stories concerning the domain of children's social work (Franklin, 1998, p. 5;Reid and Misener, 2001, p. 195). Similarly, Ayre (2001) noted that British newspapers focus on scandals within child protection. A recent literature report for the professional standards authority concluded that the UK media give a 'highly distorted picture' with positive stories 'comparatively rare' (Penhale and Young, 2015, p. 172). Whilst rarely conducted, studies of comparative coverage between the UK and other countries suggest that the UK press adopt a more severe attitude towards individual social workers (e.g. Lonne and Parton's (2014) comparison of press coverage in the UK and Australia; Reid and Misener's (2001) comparison of the UK and USA). Notably, Reid and Misener's (2001) empirical study of almost 400 newspaper articles revealed that 58 per cent of stories are positive in the USA compared with just 13 per cent in the UK. The authors hypothesise that this difference is due to the UK press' 'preoccupation with a number of notorious scandals involving child maltreatment ' (p.198).
Whilst negativity is a widely recognised feature contributing to the newsworthiness of press stories in general (Caple et al., 2020), the intense negativity of the UK press towards social workers, in particular, led Parton (2014) to suggest that social workers have become the 'cultural scapegoats' for UK society (p. 2049). This seems to particularly occur in cases of child abuse resulting in apparently preventable deaths with participating social workers presented in the press as being 'both inefficient and. . . repressive, overly interventionist and insensitive ' (2014, p. 2048). For many members of the public, newspaper stories are a significant source of information on the role of social workers, meaning that the UK press plays an important role in constructing, reproducing and exacerbating the culture of blame towards social workers within society. The negativity and blame surrounding mentions of social workers in the UK press have in turn contributed to social worker unease (Ayers, 2016;Legood et al., 2016). For example, practising social worker Ayers (2016, n.p.) points to 'a climate of fear. . . [in which] The guilt and anxiety can be all consuming'. Further evidence for the importance attached to media coverage is provided through Legood et al.'s (2016) interview study (n ¼ 16) which revealed a 'widespread view that the media are fundamental to the shaping of the public's understanding and perception of social services' with one social worker interviewee wryly commenting that 'you only ever hear the bad news' (2016, p. 1880) and many in Legood et al.'s study speaking of a 'stigma' attached to the job title of 'social worker ' (p. 1878). Similarly, Parton (2014Parton ( , p. 2053 points to the increasing public anxiety around child protection and the consequent 'anger and hostility projected onto social workers' with the media playing an important role. In addition to the effect on social worker recruitment and retention, negative press representation (whether actual or perceived) is likely to further stigmatise individuals and families who require social worker support.
Much has changed since Reid and Misener's seminal 2001 study of the press. The years 2008-2010 saw an 'exceptional cluster of media stories about child abuse with the focus on social workers and their presumed failings' (Jones, 2012, p. 85). More recently, social workers were heavily criticised following child sex abuse scandals in Rochdale, Oxford and Rotherham (e.g. Bingham, 2015). The proliferation of 24 h news channels, decline in print press and increase in online media have led to an ever-growing appetite for new stories but also open up opportunities for social workers to feed into the demand for new angles (Jones, 2012). The current study aims to provide an up-to-date snapshot of how social workers are portrayed in UK newspapers, and to explore and categorise these depictions.

Methodology
The dataset for this study is a collection of newspaper texts, explored using specialised software. The process of compiling the texts and analysing them is explained next.

Building and exploring the SWP2019 corpus
This study employs the methodology of corpus linguistics; that is, it involves the compilation of an electronic collection of texts (a corpus), and explores this using computer software. This methodology enables more textual data to be analysed than by manual analysis alone. Corpus linguistics allows quantitative techniques to be employed whilst also providing a 'way in' to more detailed qualitative analysis and avoiding the danger of 'cherry picking' interesting instances (see McEnery and Hardie, 2012, for an accessible introduction to corpus linguistics; Baker and Levon, 2015). The wide-reaching focus here is on the portrayal of social workers in general, rather than on particular cases or the profession overall. The combination of systematic computer searches with close reading of each instance of the search term and extensive surrounding text enables detailed classifications to be made and has been widely used in recent years within applied linguistics (e.g. Leedham et al., 2020;studies in Taylor and Marchi, 2018). Ethical approval was not required as all articles are in the public domain.
The online database of newspaper and business information Nexis UK was used to search for 'social worker(s)' in all print and online UK national newspapers in the three months period 7 May 2019 to 6 August 2019 (filtered to include only UK national press, search conducted on 7 August 2019). The profession 'social work' was not included as a search term as the study is designed to uncover press portrayal of social workers rather than the profession. This search gave 987 'hits' within 736 articles covering 5,170,947 words. The articles were saved as text files and the resulting corpus (named 'SWP2019') was loaded into the corpus linguistics software package WordSmith Tools version 7 (Scott, 2019). Duplicate texts were removed in WordSmith (defined by the Tool as 500 identical characters centred on the search word), and rechecked in Excel. The remaining 974 hits were concordanced, that is, given with the search term in the middle of each line so that the immediate language surrounding each hit (the co-text) can be viewed ( Figure 1). Lines should be read vertically, rather than horizontally, to scan the immediate co-text occurring to the left and right of the search term. Concordancing thus enables the analyst to search for any word used in the corpus, to view the search term in different ways, and also to expand each concordance line to see the fuller context: these different 'ways in' bring a fresh perspective to the dataset.
Each instance of 'social worker(s)' was manually categorised as negative, positive or neutral through an iterative process of analysing concordance lines combined with researcher close reading of text extracts. Subsequent levels of categorising were conducted in Excel and used substantial co-text; for example, 'positive' instances were subcategorised using iteratively derived terms, such as 'effective individual or dream job'. Each concordance line was also classified according to domain (children's, adult generic or adult mental health) where this is retrievable from the text, and the particular case or named social worker, to check for dominance of a single case.
In addition to the exploration of SWP2019 using WordSmith Tools and Excel, further holistic analysis was provided through semantic tagging using the corpus software WMatrix (Rayson, 2009). Semantic tagging denotes the process of automatic computer annotation of lexical items (words or phrases) in the corpus with a meaning-based tag or label. The purpose of this is to group words in the same meaning group together, for example, 'brother, sister, grandparent' are grouped together as KIN, thus enabling a focus on broader word groupings rather than individual lexical items. The semantic groupings from SWP2019 were then compared with those from the general British National Corpus written sampler (a collection of 3 million words of written texts taken from the domains of fiction, newspapers and academic writing) in terms of which semantic groupings occur more frequently in the former, whilst taking into account the different corpus sizes. This comparison means the analyst can focus attention on prevalent meaning areas of the corpus under investigation even though a particular lexical item may not appear frequently.

Classification issues
Classification of mentions of the search term into positive, negative and neutral is a subjective decision-making process. Initially, a randomly selected sample of 200 concordance lines were classified by the author, then re-classified two months later to determine the level of intra-rater reliability. In round 1, 35 instances of the 200 were classed as negative; in round 2, this increased to 49. Many of the reclassifications were due to close reading of extended co-text: 1) Jurors were told how two days later, a social worker had seen Scarlett looking 'chatty, alert and happy'. But that night, with Scarlett dying in the back seat of her car, Porton was filmed on CCTV stopping to fill up her car with petrol as she drove the child to hospital. [Dailymail.co.uk,02/08/19] In Example 1, reading just the sentence containing the search term without knowledge of the context led to a neutral classification. When the longer section is included, the social worker description of Scarlett appears in stark contrast with later events, signalling the writer's implicit view that insufficient questioning was carried out during the home visit. Thus, the more co-text read around the search term, the greater the opportunity to perceive implicit negativity in the portrayal of the social worker. Whilst the sentence containing the search term could be taken as the unit of analysis, this would not always give sufficient context to make a fully informed decision as to the evaluative stance taken, as shown by Example 1. (Note that whilst only a single sentence is given in most examples, due to reasons of space, in analysing the data, I read at least one sentence before and one after the sentence containing the search term.) In total, 18 of the 200 instances in the sample were classified differently the second time, giving an intra-rater reliability count of 91 per cent. In general, the greater co-text examined in round 2 enabled more precise classifications to be made of each occurrence of the lexical item. Subsequent analysis followed round 2 in examining at least one sentence before and after the sentence featuring 'social worker(s)'. Classification was iterative, involving reading and rereading extracts and sifting using an Excel spreadsheet to ensure the process was as consistent as possible.

Findings and discussion
This section begins with an overview of findings before exploring each broad categorisation (neutral, negative and positive).

Overview of quantitative findings
The broadest level of classification of each of the 974 instances of 'social worker(s)' in the whole corpus into 'negative', 'positive' and 'neutral' mentions gave 25 per cent of instances as negative (242 raw count), 6 per cent positive (57) and 69 per cent neutral (675). It should be noted that no single case dominated the newspapers during the three-month period of the study. All instances were retained, including neutral occurrences where the mention of a person being a social worker was only incidental (Examples 2 and 3).
2) Forbes, 67, was ordained as an unpaid minister in 2004 when she was still working as a social worker [The Guardian, 22 May 2019].
3) The social worker panicked as her hands got trapped and dragged into the shutter door mechanism on 14 August, 2016, the hearing in Huntingdon was told [independent.co.uk, 11 June 19].
Note that the study counts each and every occurrence of the search terms 'social worker(s)' within a text; for positive and neutral, this is frequently the sole mention of 'social worker(s)' in a story. In contrast, newspaper articles which portray social workers negatively tend to feature multiple mentions (some of which may individually be neutral). Whilst counting by instances rather than by stories boosts the number of negative instances, it can also be argued that a clustering of negative instances in a single story is likely to leave a strong impression on the reader.
Overall, findings confirm the press preoccupation with children's social work, particularly child protection (Parton, 2014), and point to a tendency towards a blame culture (Munro, 2010(Munro, , 2011(Munro, , 2019, where social workers are vilified when a child death occurs. The following exploration of subcategories explores this tendency further.

Negative classifications
Over two-thirds (69 per cent, 166 instances) of the 242 negative occurrences of 'social worker(s)' concern children with just 10 per cent (twenty-four instances) relating to adult generic, <1 per cent to mental health and 20 per cent where the social work domain is unnamed. This reflects the overall dominance of UK social work news stories concerning children in the literature. Figure 2 gives the iteratively derived subclassifications for the 242 negative instances.
By far the largest negative classification category is 'failure to notice/ act', denoting occurrences ranging from social workers being dismissive of concerns raised by others (Example 4), or failing to take sufficient action (Examples 5 and 6) to (rarer) instances where professionals are taken in by the perpetrator (Examples 1 and 7). The language surrounding 'social workers' in the examples highlights the opportunities to notice and act and the apparent failure to do so. Thus, in Example 4 despite the known 'histories' of the two men, social workers 'dismissed concerns and did not intervene': the repetition in this final clause emphasises the lack of action. The use of service user quotation in (Example 6) gives a personal perspective and implies a lack of attention on the part of social workers who did not 'pick up on' the 'vacant' expression. Whilst the Daily Mail uses the relatively restrained comment on social workers (and teachers) 'facing serious questions' after perceived failing (Example 5), The Sun presents social workers as gullible individuals who were 'conned' (Example 7). From Figure 2, it seems that social workers are criticised more frequently for 'in'action (41 per cent of all negative instances) than for over-zealous activity (16 per cent of negative instances). This observation is supported by the keyness of the semantic grouping FAILURE, FAILING, LOSING in the corpus when tagged using WMatrix (Rayson, 2009) and compared with a larger, general written corpus (the British National Corpus written sampler). The dominance of this semantic grouping supports Munro's (2010) argument of the tendency within a person-centred approach towards a culture of blame. If something bad happens, there must be someone to blame such that '[a]ccidents are no longer accidents at all. They are failures of risk management' (Dekker, 2005, p. x. in Munro, 2010), meaning child death is more palatable to the public as an individual can be held responsible and others can take a moral stance (cf. Warner, 2014).
The second most prevalent subcategory is termed 'separate to job' where the fact that someone is a social worker appears to be secondary to the negative news story concerning their behaviour. Whilst the perceived misdemeanour is not carried out as part of the social worker's professional role, these examples are retained in the categorisation as the instances add to the negativity around social workers. Examples range from crimes committed by social workers which may have an impact on their job (Example 10) to moral judgements on the part of the article writer (Example 11). Whilst 'fleecing' a partner could never be classed as moral behaviour, it seems from (Example 11) that social workers' personal lives are also held up for scrutiny with less conventional relationships highlighted.
The third subcategory in Figure 2 classifies the social worker as 'overzealous', referring to the portrayal of social workers as acting disproportionately. Note that this category accounts for just 16 per cent of negative instances, compared with over twice as many for the category of failing to notice/act (43 per cent), giving a ratio of 5:2 for inaction as opposed to perceived over-action. Often the over-zealous behaviour takes the form of pressurising someone at a vulnerable point in their lives (Example 12). Sometimes the social worker is described as scrutinising or questioning in an overly strong manner (Examples 13 and 14). 13) I was quizzed by medics and social workers, then made to wait for a long time [The Guardian, 24 July 2019].
14) The social workers asked me the most prying questions [The Guardian, 03 August 2019].
On occasion, the journalist suggests that the social worker is pursuing their own agenda (15). In Examples 12-15, the lexis chosen in the surrounding co-text highlights the negativity: 'under pressure, accusing, quizzed, prying'. Examples 12 and 15, in particular, suggest that individual social workers may be too vehement in the pursual of their goals. The most extreme example of the press describing over-zealous behaviour is Example 16, where an article on the British police force refers to officers as social workers and-metaphorically-soldiers intent on pursuing a leftwing politically correct agenda (this appears somewhat ironic given the widespread discussion of 'canteen culture' in the police force). Whilst describing police officers rather than social workers, this final example has been retained as it draws on a discourse of social workers as powerful 'pc gone mad' vigilantes. Whilst such extreme language is rare, it is likely to leave a lasting impression on the reader and links both police officers and social workers with extremist, state-sanctioned behaviour.
The fourth subcategory in Figure 2 contains almost 10 per cent of negative classifications grouped together as 'other' as the individual subgroups are each low. These include three (raw) occurrences categorised as 'malpractice', four instances where a social worker is alleged to have claimed a victim was to blame and one concerning the excessive cost of agency social workers.
The 'bureaucracy' category illustrates how social workers are at times portrayed as 'jobsworths', acting as uncaring automatons (Examples 17 and 18). 17) There were no foster places available, so he was driven to a B&B on the outskirts of Chesterfield by a social worker and told to wait for someone to tell him what to do next [The Guardian, 26 July 2019].
18) But Mr Moore, who attended last month's D-Day commemoration in Normandy, died knowing he had won his final battle because his wife of 56 years was 'reassessed' by Kent social workers after the Mail highlighted the case [dailymail.co.uk, 28 July 2019], These instances were perceived as different due to the lack of action subcategory as the focus is on excessive red tape at the expense of human feeling, though clearly there is overlap.
The category of 'media' here refers to the entertainment media of television (TV), film and books. Whilst a minor subcategory with just ten instances, this grouping is interesting as it shows how TV tends to feature social workers as well-intentioned but flawed characters, serving to perpetuate inaccurate stereotypes. Child protection cases where social workers remove a child from their family often feature in TV dramas (Example 19), portraying over-zealous behaviour rather than the more common perception in newspapers of a lack of action. The social worker in example 19 is depicted as the enemy and rather than supporting the parentless siblings, they 'loom' and wait to perhaps remove children, acting as the faceless 'state'.
In a further entertainment example, this time a reality show, a social worker is presented as non-mainstream and struggling with their personal issues (Example 20); perhaps the 'traumatic past' is commented on to prompt the reader into wondering how this individual can help others. 20) In the mix this season are a single mom DREAMer, a Catholic virgin, a pansexual Trump supporter, a small-town farmer, a leader of the black community, a social worker with traumatic past and a pansexual feminist artist [Mail Online,19 June 2019].
In the positive categorisation, there was just one instance deemed 'media' with the neutral grouping having a further nine, meaning that half of all social worker mentions connected with entertainment were negative portrayals. Further exploration of the depiction of social workers in the media is beyond the scope of the current study but warrants investigation since for many people, newspaper reporting and entertainment media coverage are the only time that social workers are seen in action (e.g. see Edmondson and King, 2016), and if this portrayal is often negative this may leave a lasting impression.
The final category of 'abuse rings', whilst small, is an extreme one so is important in the overall portrayal of social workers in the press. There are five instances of social worker extreme misconduct, spread over two cases: one concerns a group of care professionals brainwashing children in Italy, the other (one instance) details a social worker in the UK jailed for child sexual abuse. (Whilst SWP2019 is compiled entirely from UK national press articles, a small number of these refer to non-UK contexts.)

Positive classifications
Positive mentions of social workers are the smallest of the three overarching categories forming just 6 per cent (n ¼ 57) of the total. As with negative instances, where the social work domain is given, this is predominantly that of children's social work (thirty-four instances or 60 per cent).
The main subcategory within positive mentions is 'praise for a group', such as social workers within a Local Authority, often including a quotation relating to the work carried out (n ¼ 16): Example 24 depicts an instance where a social work student is denied their dream job, due to accusation of personal prejudice which would be particularly unacceptable in social work.
Taking the stance that social work requires particular skills and personal strengths is the subcategory of being an 'effective individual' (eight instances). This group comprises depictions of current or former social workers who are utilising their social work skills in other areas of life (as a football manager in Example 25 and politician in Example 26). Where positive mention concerns an individual social worker, rather than a group of social workers or a Local Authority, the person is frequently named (Examples 22,24 and 26). This chimes with Legood et al.'s (2016Legood et al.'s ( , p. 1886 finding that members of the public will 'hold a distinct and more positive view of an individual social worker' even if the overall negative impression of the group endures. The naming of 'good' or 'expert' individual social workers contrasts with the absence of names in the negative mentions of individual social workers (exceptions being managers who are named and shamed in high profile cases) and also perhaps serves to reinforce the underlying notion that these praiseworthy individuals are 'the exception that proves the rule'.

Neutral classifications
The corpus contains a total of 675 instances of 'social worker(s)' categorised as neutral, forming the largest broad grouping at 69 per cent. The largest subcategory within neutral was straightforwardly named as 'doing their job', denoting social workers going about their role with no judgement conveyed (265 instances, 39 per cent of the neutral category): 28) When the social worker arrived that morning in September 2015, I was in the grip of addiction [The Sun, 14 July 2019]. 29) Smith was given a community payback order which will see him supervised by social workers for 12 months [mirror.co.uk, 24 July 2010].
A similar subcategory is that where 'social worker' is mentioned as a 'job title' (139 instances, 21 per cent) (Example 30)-sometimes as a victim of crime (Example 31)-or where the job title is one within a list of professionals. Smaller subcategories include discussions of social workers in the contexts of 'government cuts' (Example 32), 'training' for social workers (Example 33) and as characters in 'TV' dramas. A total of thirty-five occurrences (5 per cent) were subcategorised as 'empathetic' towards social workers (Example 34). 34) the sheer terror that the women and children who live with violent men experience on a daily basis is the same terror that social workers confronting these same men have to deal with, often with no protection [The Guardian, 10 June 2019].
The next section brings together findings overall.

Discussion and conclusion
Using a snapshot of three months of newspaper articles featuring social workers, this study confirms that the domain of children's care is the most common social work area featured in the press, with adult generic and adult mental health rarely depicted. In response to the first research question posed (what are the proportions of positive, negative and neutral mentions of 'social worker(s)' in the UK press), findings indicate that, whilst the majority of occurrences are neutral (69 per cent), there are far more negative than positive portrayals of the profession (25 per cent negative and 6 per cent positive). The second research question asked: what are the dominant subcategories within each of these three groupings? Categorising of concordance lines through iterative rereading and sorting, employing substantial co-text, indicates that the negative mentions are dominated by the subcategory 'failure to notice/act' (43 per cent within 'negative') in contrast to 'over-zealous' behaviour (16 per cent). The observation that social workers are criticised far more for 'in'action than perceived 'over'-action is supported by the keyness of the semantic grouping FAILURE, FAILING, LOSING in the SWP2019 corpus. Implicit in many newspaper reports is the view that social workers should have known and should have acted. Thus, cases of neglect and abuse (largely of children) are viewed not as a resourcing or societal issue but as social workers failing to act. Positive instances of the search term relate largely to individuals' qualities as effective practitioners, experts in their field or to achieving their 'dream job', whilst the neutral category is dominated by matter-of-fact mentions of social worker as a job title or to social workers simply doing their job with no overt judgement given.
Whilst negative instances form just a quarter of all instances in the corpus, these are likely to be highly influential on newspaper readers as most members of the public have little or no personal contact with social workers and thus may base their views on what they hear or read in the media. Similarly, from the available literature, it appears that the perception amongst social workers themselves is that most newspaper reporting is highly negative with social workers feeling they were 'always being singled out as the "failing" profession' (Legood et al., 2016(Legood et al., , p. 1878. Press portrayal of social workers as negligent professionals who frequently fail to either notice or act in cases of suspected child abuse reinforces and perpetuates the dominant discourse of negativity and culture of blame surrounding the profession and is highly likely to harm both recruitment and retention (see Introduction section). Few other professions appear to receive such consistently negative press coverage.
Social work is notably portrayed in the press as a 'dream job' which is immensely challenging, requiring dedicated, efficient individuals; this positive categorisation, however, has the negative consequence of any social worker failing to live up to this vocational ideal being quickly perceived to be inadequate and blamed for systemic failures. In a significant minority of negative occurrences in the study, the newspaper story concerns an individual who happens to be a social worker and who has behaved badly in their personal life; this suggests that the press tend to portray social workers as individuals who should be 'good people' in all aspects of their lives and who are held to a higher moral standard than rest of us. Increasing awareness of the wide-ranging role of social workers and the multi-agency nature of many social work cases could help to alleviate the tendency of press and public towards blame for inaction or over-zealous action in difficult cases.
Limitations of the current study include the compilation of a single corpus containing all sections of national UK newspapers rather than divided by newspaper and/or by section, though the provenance of each example is indicated throughout. A more fine-grained analysis could compare popular and broadsheet representations of social workers and take account of different readerships, whilst social workers as educated professionals are more likely to read the Guardian, The Sun has a far higher circulation (e.g. Mayhew, 2019). A further limitation is the short time-period explored, and future work could include replication over a longer time period, or broadened in scope beyond the UK, perhaps using constructed week samples (cf. Reid and Misener, 2001). Further work could focus on media platforms, such as TV, radio and social media in order to build a broader view of the impacts of various media. As this is a study of press reporting only, a missing factor is the views of social workers themselves and a future study could include interviews with social workers on their views of media portrayal. Whilst not including a specific question on the media, the Writing in professional Social work Practice (WiSP) project (Lillis et al., 2019) interviews include several social worker mentions of the press, many in the context of a fear of being named and shamed in high profile cases. A final suggestion for future research is to conduct a similar study on how other professions are portrayed (e.g. police, health care) to provide a comparison point. Questions to explore in future studies of media portrayal include: How far do UK news stories drive the culture of blame surrounding particular professions? To what extent does the negativity surrounding social work in the UK press harm recruitment and retention rates?
Findings from this study illustrate the dominance of child protection cases in the UK national press, an intense focus on social workers' perceived failure to act and on perceived over-zealous behaviour, indicating that a culture of blame remains prevalent. For social work practitioners, managers and service users, the study highlights the broad range of both negative and (more limited) positive areas of social work in the UK press. It is also hoped that the study will encourage unions and other professional bodies (i.e. those responsible for representing social workers), as well as social workers themselves to consider engaging with the media and presenting more positive stories of the profession (e.g. Maier, 2009) in order to promote a shared understanding amongst journalists and the public of social workers' role in society (cf. Jones, 2012).