Child abuse in England and Wales 2003–2013: Newspaper reporting versus reality

Sample

This research is based on newspaper articles on child abuse and neglect in randomly constructed weeks between 1 April 2003 and 31 March 2013. There were two randomly constructed weeks for each year in the decade; that is, 14 days of newspaper coverage each year for 10 years. The sample contained all newspaper articles from randomly selected Mondays, Tuesdays, Wednesdays and so on, thereby ensuring that for every year, each day of the week is covered equally. These randomly constructed weeks were adopted to control for the bias of cyclical trends in newspaper coverage by preventing the over- or under-sampling of typically light or heavy news days (Mejia et al., 2012).

Articles were selected through a keyword search of the LexisNexis database within the two constructed weeks. The keywords used were child abuse, sexual abuse, physical abuse, emotional abuse and neglect. Articles about current or recent child abuse within England and Wales were the focus. Therefore, articles about historical child abuse (i.e. when the complainant was an adult), stories involving accusations of historical abuses within the Church and articles about child abuse or neglect committed outside of England and Wales were excluded. Historical and church abuses were excluded as their high-profile nature (e.g. Jimmy Savile-related stories) may have obscured regular patterns of child abuse and neglect coverage. These exclusions left some of the randomly selected dates with no articles on child abuse, and so random dates for these days were chosen again and the same criteria adopted as above.

The final sample consisted of 459 newspaper articles, describing 281 different incidents or themes. The 10-year sample included articles from the following UK national newspapers: The Guardian, The Independent, The I, The Observer, The Times, The Telegraph, The Daily Mail, The Sun, The Express, The Mirror, The Daily Star, The People and The Morning Star (including Sunday/weekend editions of these papers). Of these 459 articles, 280 (61%) were from tabloid newspapers and 179 (39%) were from broadsheet papers.

Coding

The 459 articles were imported into the NVivo10 database so that they could be coded into different categories for analysis. Two researchers used an iterative process of coding until satisfactory inter-rater reliability was reached. Cohen’s Kappa was calculated on the ratings of 8.5 per cent of the newspaper articles to determine the extent of inter-rater reliability for each of the three framing dimensions of the coding system: frame of the story (episodic/thematic), frame of the cause (individual/societal) and frame of the solution (individual/environmental). Accounting for chance, inter-rater reliability was high (κ = 0.85) for the coding of all three types of framing.

The definitions of these framing dimensions of the story are outlined further in Table 1.

OPEN IN VIEWER

Table 1.

Definitions of framing.

Frame of story
Episodic	Event or series of events
Thematic	Explores broader context (episodic framing may also be present)
Frame of cause
Individual	Focused only on children or perpetrators in case
Societal	Incorporates roles of non-offending family members/organisations in causing the problems, for example, failure of social services
Frame of solution
Individual	Focused only on children and or perpetrators in case, for example, prison, psychiatric hospital or supervision orders
Environmental	Solution incorporates offender community rehabilitation orders a and/or broader efforts to change the media, government or other institutions (Mejia et al., 2012)
None	Includes justice system intervention if not convicted  b

a

Incorporating ‘offender community rehabilitation orders’ into the ‘Environmental Solution’ category was decided upon as this form of punishment considers the effect of the offender’s actions on the wider community and demands that the offender ‘gives back’ to the community as punishment. It aims to help the offender become a law-abiding, responsible member of the community and therefore incorporates the role of the offender’s role in society as a potential cause of and solution to his or her unlawful behaviour.

b

For example, a newspaper article may be reporting that someone has been arrested on suspicion of committing an offence, but they have not yet been charged with an offence or been found guilty of committing the offence. Therefore, there is no solution suggested as the report is merely mentioning their potential involvement in a case.

These definitions were adapted from the coding systems developed by Hove et al. (2013) and Mejia et al. (2012), which in turn were derived from both deductive and inductive manual-holistic approaches to frame analysis. In particular, the definition of the frame of the article and frame of the cause as episodic (event-focused reporting) or thematic (context and wider issue focused) was adopted from Hove et al. (2013). Hove and colleagues coded story frames into the categories of episodic or thematic and thus did not account for both frames being present in the same article. While there are statistical advantages to this mutually exclusive frame, we agreed with their analysis that episodic and thematic reporting in one article can be helpful and reporting does not need to be either/or. Therefore, in our coding system, if an article has a thematic trait, even if it also has episodic traits present, it was coded as thematic. We adopted the frame of the solution as individually focused on offenders or victims and/or environmentally focused on broader systems beyond the individual from Mejia et al. (2012) (see Table 1).

In addition, articles were coded for the following:

Types of child abuse reported in newspapers

Throughout the decade, child sexual abuse was more commonly reported than all other forms of child abuse and neglect combined, in both tabloid and broadsheet newspapers (see Figure 1).

OPEN IN VIEWER

Figure 1.

Reporting child abuse in tabloid and broadsheet newspapers.

Overall, from 2003 to 2013, 55 per cent of reporting covered child sexual abuse only, 10.9 per cent concerned neglect only, 6.8 per cent physical abuse only and 0.2 per cent emotional abuse only, with 27.2 per cent of stories being about multiple forms of abuse. In the entire 10-year sample, there was only one article that had emotional abuse as its sole concern. The ‘multiple’ category consists of articles mentioning more than one type of abuse. When the ‘multiple’ accounts that include sexual abuse were included in the total percentage of newspaper reporting, accounts of sexual abuse increased to 75.4 per cent of the articles.

Reporting of child neglect, physical and emotional abuse remained consistently low throughout the decade. While reporting of sexual abuse fluctuated considerably over time, it was far higher than other forms of abuse and neglect in each of the 10 years (see Figure 2).

OPEN IN VIEWER

Figure 2.

Newspaper coverage of child abuse 2003–2013.

Comparison with estimated prevalence of types of abuse

Child protection data give a useful overview of the scope of child maltreatment dealt with by authorities. The number of children with child protection plans gives a better indication of the relative extent of categories of child abuse and neglect than notifications to children’s services because these cases are substantiated by a social worker, that is, there is some confirmation that child abuse and/or neglect has occurred.

Table 2 shows the percentage of child protection plans by type of abuse from 1 April 2003 to 31 March 2013. On average, the majority (43.5%) were for neglect followed by emotional abuse (24.9%). Physical abuse comprised 14.2 per cent and sexual abuse comprised 6.8 per cent of cases, while 10.7 per cent were for cases of multiple forms of abuse.

OPEN IN VIEWER

Table 2.

Percentage of child protection plans by type of abuse in England and Wales 2003–2013.

Type of abuse (Department for Children, Schools and Families, England and Wales figures) – percentages may not tally exactly due to rounding
	Neglect		Emotional abuse		Physical abuse		Sexual abuse		Multiple		Total for all categories
2003–2004	13,530	40.6%	6130	18.4%	6185	18.5%	2985	8.9%	4525	13.6%	33,355
2004–2005	14,175	42.9%	6170	18.7%	5925	17.9%	2875	8.7%	3925	11.9%	33,070
2005–2006	14,755	44%	7110	21.2%	5445	16.2%	2745	8.2%	3510	10.5%	33,565
2006–2007	15,925	44.6%	8270	23.1%	5450	15.3%	2660	7.4%	3425	9.6%	35,730
2007–2008	16,395	45%	9180	25.2%	5315	14.6%	2460	6.8%	3070	8.4%	36,420
2008–2009	16,920	46.1%	9745	26.5%	4785	13%	2175	5.9%	3090	8.4%	36,715
2009–2010	18,480	44.2%	11,535	27.6%	5415	12.9%	2505	6%	3890	9.3%	41,825
2010–2011	19,865	43.6%	12,290	27%	5175	11.4%	2555	5.6%	5695	12.5%	45,580
2011–2012	19,470	42.6%	13,265	29%	5065	11.1%	2370	5.2%	5560	12.2%	45,730
2012–2013	19,125	41.5%	14,705	32%	5070	11%	2190	4.8%	4995	10.8%	46,085

The compilation methods of the statistics can change from year to year and from region to- region, and so the figures are not always comparable. For example, a new method of calculating the number of children subject to child protection plans in England was introduced in 2010. l

Wales: 2009–2013 statistics taken from: https://statswales.wales.gov.uk/Catalogue/Health-and-Social-Care/Social-Services/Childrens-Services/Service-Provision/ChildrenOnChildProtectionRegister-by-LocalAuthority-CategoryOfAbuse-AgeGroup – also, NSPCC pdf available at http://www.nspcc.org.uk/Inform/research/statistics/wales_wdf49861.pdf

England: Department for Education (2013) Characteristics of children in need in England, 2012–2013:https://www.gov.uk/government/publications/characteristics-of-children-in-need-in-england-2012-to-2013 – also, NSPCC pdf available at: http://www.nspcc.org.uk/Inform/research/statistics/england_wdf49858.pdf

Older statistics on child protection were gathered by the Department of Children and Family Services (DCFS) in Referrals, Assessments and Children and Young People who are the subject of a Child Protection Plan. These have now been archived but can still be accessed online at Referrals, Assessments and Children and Young People who are the subject of a Child Protection Plan, England – Years ending 31 March 2004–2008 – Table E: Children who became the subject of a Child Protection Plan (CPP) 1, by category of abuse: http://webarchive.nationalarchives.gov.uk/20130313205001/http://education.gov.uk/rsgateway/DB/SFR/s000811/sfr24-2008.pdf

Figure 3 demonstrates that neglect and emotional abuse were the most common sources of child protection plans and that these issues were relatively invisible in newspaper articles.

OPEN IN VIEWER

Figure 3.

Child abuse and neglect depicted through newspaper articles and the number of children subject to child protection plans.

It is important to note here that some child abuse is not notified to authorities and thus is not reflected in child protection data (Harker et al., 2013; May-Chahal and Cawson, 2005; National Crime Intelligence Service (NCIS), 2005). One national self-report survey found that 10 per cent of respondents reported having been sexually abused as a child involving sexual contact, with a further 6 per cent reporting experience of non-contact abuse (May-Chahal and Cawson, 2005) – figures which far exceed cases coming to the attention of government agencies.

In another study, 2275 children and young people from 11 to 24 years were interviewed about their experiences of abuse and neglect (Radford et al., 2011). In 2011, young people between 11 and 17 years reported some form of sexual abuse (11.3%), physical violence by a parent or guardian (2.4%), physical violence by another perpetrator (36%), emotional abuse by a parent or guardian (3%) and/or emotional abuse by another perpetrator (31.9%).

It is clear from the discrepancy between self-report surveys and officially recorded identified cases that some child abuse remains hidden in England and Wales, and indeed in all countries where research has been conducted (e.g. Millard and Flatley, 2010; Sedlak et al., 2010), perhaps particularly the incidence of child sexual abuse. So perhaps while the reporting of sexual abuse does not over-emphasise the problem of sexual abuse (Cheit, 2003), it does appear to under-emphasise other forms of child maltreatment, particularly emotional abuse and neglect.

Victims and offenders

The ages and genders of all victims and offenders were coded where reported. If one article included more than one victim and/or offender, the age and gender of the victim(s) and offender(s) were coded separately. Where evident, the number of offenders and victims in cases reported in newspapers were also noted, alongside the relationship between the offenders and children.

In all, 97 per cent of all event-focused articles (i.e. episodic framed articles) reported at least one piece of information relating to the offender (316/327 articles). This does not necessarily mean that all information relating to the offender was reported; it simply means that in 316 of the episodic articles, at least one incident of child abuse was reported and the report gave at least one piece of information relating to the perpetrator (e.g. age, gender and/or relationship to the child). Overall, child sexual abuse committed by men outside the family was the most common form of abuse reported in the newspaper articles.

Information about victims

Information about perpetrators

Framing child abuse

Articles were analysed to determine whether the newspaper articles used an episodic or thematic frame overall (see Table 1 for definitions). Of the 459 articles, 71.2 per cent used a purely episodic frame (n = 327), only discussing specific details about a crime without any contextual comment, whereas 28.8 per cent of articles also used a thematic frame. Figure 4 shows that episodic framing is more prevalent in tabloid newspapers (82.5% articles) than in broadsheet newspapers (53.6% articles).

OPEN IN VIEWER

Figure 4.

Episodic and thematic frames in broadsheet and tabloid newspaper articles 2003–2013.

Articles were also analysed for the way they depict the causes of child abuse (see Table 1 for definitions). The causes of child abuse were framed in terms of individuals in 76.3 per cent of the total sample of newspaper articles. All these articles were framed as the perpetrators’ responsibility. There were no articles that blamed the victim for the abuse.

Societal causes of child abuse were evident in nearly a quarter of the articles analysed (23.7%). Forty-four of these articles (40.4%) blamed children’s services in reports on the outcomes of Serious Case Reviews. Religious or cultural practices were identified as a cause of child abuse in 18 of the articles (16.5%), for example, ritual child abuse in witchcraft. The third most common societal cause reported related to problems with formulation and/or implementation of the law (n = 15; 13.5%). Most of these articles were concerned with the courts giving sentences that were too lenient.

Tabloid newspapers were more likely than broadsheet newspapers to frame the causes of child abuse in terms of individuals (see Figure 5). In all, 84.6 per cent of the articles in tabloid newspapers framed the causes only in terms of individuals, whereas 63.1 per cent of the broadsheet articles were framed in this way.

OPEN IN VIEWER

Figure 5.

Broadsheet and tabloid framing of the causes of child abuse 2003–2013.

Finally, articles were analysed for the ways that they framed the solutions to child abuse (see Table 1 for definitions). In total, 37 per cent of the articles framed the solutions in terms of individuals; 27.2 per cent incorporated environmental solutions and 35.9 per cent offered no solutions in their articles.

Once again, broadsheet newspapers are more inclined to include environmental solutions than tabloid newspapers. Only 16.1 per cent of tabloid newspaper articles, compared to 44.7 per cent broadsheet articles, offered solutions that incorporated broader societal efforts to deal with child abuse. However, still nearly a quarter of broadsheet articles (23%) incorporated solutions only framed in terms of individuals. There was no notable difference between tabloid and broadsheet papers in the number of articles that offered no solutions (Tabloid – 37.9%; Broadsheet – 33%).

Limitations

There are a number of limitations to this study. First, this study excluded reporting of landmark cases of abuse in the Catholic Church and widely reported abuses by the celebrity, Jimmy Savile (due to the historic nature of these abuses and not wanting their inclusion to distort general patterns and trends in reporting of child abuse). Such high-profile cases may have led to increased reporting of other incidents of child sexual abuse and therefore impacted the frequency and content of that coverage. However, our findings were similar to those of comparable studies (e.g. Hove et al., 2013; Mejia et al., 2012).

Second, similar to other studies, this study relied on LexisNexis. It is possible that this database did not capture all of the newspaper articles that covered all types of child abuse. There are also limitations to only including newspapers and excluding other forms of media reporting such as television, radio and Internet sites. This is significant, as a larger portion of the population is exposed to other mass media while newspaper readership is in decline. However, the agenda-setting role of the newspapers for television renders newspapers’ place in framing the issue critical. It also means that the frequency and frames of child abuse on television news may not differ markedly from those of newspapers (Hove et al., 2013).

The exclusion of historical cases because of their sometimes high-profile nature may not have achieved its intention of increasing representativeness, given that the ‘politics of outrage’ (Parton, 2014) and the strident media coverage (Lonne and Parton, 2014) may have become, at least temporarily, the norm in the United Kingdom.

Our definition of ‘individual’ as encompassing only the child and/or perpetrator ignores the fact that some of the reports falling under our definition of ‘societal’ and ‘environmental’ include discussion of individual professionals.

By their nature, our data are more descriptive than explanatory. It does not permit us to understand why journalists under-report neglect, intra-familial abuse and contextual factors. Qualitative interviews may be useful in addressing these questions in future.

Finally, this analysis cannot demonstrate the effect of episodic/thematic news reporting on audiences; it can only demonstrate the frequency of the framing. However, it is reasonable to assume that people gain at least some of their understanding of the prevalence, causes and prevention of abuse from newspapers (Hove et al., 2013; Lonne and Gillespie, 2014). Furthermore, a recent analysis of child protection and media reporting in Quebec suggests that newspaper articles not only influence public attitudes but also reporting behaviour to child protection services (Saint-Jacques et al., 2012).

Why are physical and emotional abuse and neglect being ignored?

Our findings are strikingly similar to analyses of newspaper coverage in the United States (Hove et al., 2013; Kunkel et al., 2002) and Australia (Lonne and Gillespie, 2014) in that child sexual abuse is by far the most commonly reported form of child maltreatment. A word frequency search of the 459 newspaper articles in this sample found that the word ‘sex’ was the third most frequently used term, after the words ‘child(ren)’ and ‘police’. A cynic could argue that these words sell newspapers within a highly competitive market in England.

Unlike in Australian newspapers (Lonne and Gillespie, 2014), most articles did report the number of children involved in cases, the gender and age of the victims and the perpetrators. However, 71 per cent of the articles had no thematic analysis. The episodic crime reports of individual child sexual abuse cases mirror the majority of child victimisation that reach the criminal courts (Lonne and Parton, 2014); they do not represent the majority of child abuse and neglect depicted through self-report surveys and child protection data (Gilbert et al., 2009a; Harker et al., 2013; Radford et al., 2011). In part, this is due to the confidentiality inherent in child protection cases. It is the criminal court cases that are easily accessible to journalists, although this does not explain why there is so often little context to the articles.

There was only one article that featured emotional abuse, approximately one in 10 articles that featured neglect and approximately one in 17 articles that featured physical abuse. Articles on physical abuse appeared more prominent in Australia (Lonne and Gillespie, 2014) and in Quebec (Saint-Jacques et al., 2012) than in this sample of newspapers articles in England and Wales.

The reasons for the fluctuations in newspaper reporting of child sexual abuse are unclear. It might be worth noting some of the high-profile cases at the time, which could have influenced the reporting of child sexual abuse in this sample. For example, in 2012–2013, there was major coverage of Jimmy Savile (although these articles were excluded from the sample). The spike in 2007–2008 may have emanated from the high-profile inquiries into the ‘Baby P’ case that dominated the news in 2008 and caused an increase in reporting of cases in the same local authority. In 2009–2010, there was significant reporting about the Catholic Church’s reaction to allegations of child abuse (although these articles were also excluded from the sample).

Newspapers help to ‘shape the conceptual borders of child maltreatment’ (Saint-Jacques et al., 2012). Lonne and Gillespie (2014) and Hove et al. (2013) found similar lack of reporting of neglect and emotional abuse in Australia and the United States, respectively, the most common forms of maltreatment to underpin child protection plans. Hove et al. (2013) argue that the lack of reporting of neglect is partly because it is difficult to define. At its most extreme, the impact of neglect and physical abuse is clear. In England and Wales, physical punishment of children is considered ‘unreasonable’ if it involves an implement (e.g. whip, cane or belt) and/or if it leaves a mark on the child. However, it is legal to hit children with ‘reasonable punishment’ (Section 58, Children Act, 2004), yet this term is not defined in law. Problems with definition of physical abuse can arguably present the same definitional problems as neglect in all but the most extreme of cases.

This lack of clarity in terms of legal definitions may be matched by a lack of clarity in terms of moral judgement. While we assume that most people judge sexual contact with children to be wrong, and therefore newsworthy, some people (including journalists and editors) may be ambivalent, for instance, about whether hitting children is morally justifiable in some circumstances and therefore not worthy of condemnation or reporting. This, if true, could constitute a reciprocally reinforcing pair of processes.

Beyond ‘evil’: The importance of reporting possible causes and potential solutions

In other areas, such as mental health, contextual factors – including causal beliefs – have been shown to be powerfully predictive of attitudes and responses (Read et al., 2006). The newspaper articles in this study usually lack contextual information. These results are similar to findings in Australia (Lonne and Gillespie, 2014), South Africa (Corbella and Collings, 2007), Malaysia (Niner et al., 2013) and some of the research in the United States (Kunkel et al., 2002; Mejia et al., 2012). It is interesting to note that Hove et al. (2013) found that recent American newspaper coverage of child abuse commonly uses a thematic analysis. This does not appear to be the case in England and Wales. Perhaps this reflects a difference in the competitive environment of tabloid newspapers in the United Kingdom. The frames tabloid newspapers present are likely to be more influential because they are read by far more people than broadsheet papers. In 2012, the combined circulation of The Sun and The Daily Mail was 4.6 times greater than that of the two best-selling broadsheets (The Daily Telegraph and The Times) combined (Audit Bureau of Communications, 2012). Tabloid newspapers rarely report contextual factors in either the causes or solutions to the widespread health and social problem of child abuse. Instead, they often simply report salacious details of crimes.

One would expect broadsheet newspapers to frame articles with a thematic analysis much more often than tabloid papers. So it was disappointing to see that 63.1 per cent of the sample of articles in broadsheet papers only framed the cause of abuse in terms of a ‘bad’ individual, and only 44.7 per cent offered any solutions that incorporated broader societal efforts to address child abuse. Even when causes were framed in societal terms, more than half of these articles focused on systemic failures identified through Serious Case Reviews (by definition extreme cases) or they focused on perceptions of lenient sentences. The newspaper coverage in England and Wales does not usually serve to connect people to the broader societal problem or effective policies and programmes to prevent it. It would be good to see comparable thematic analysis to American newspaper articles.

There are clearly many barriers for journalists to be able to frame articles differently, not least lack of time to write thoughtful well-researched articles. However, journalists rarely report drops in the financial markets, humanitarian crises overseas or even a poor performance by the England football team, without considerable analysis of causes and identification of potential solutions.

Hove et al. (2013) comment on professional norms and audience expectations preventing journalists from doing a better job. It is interesting that they suggest one criteria for the way that an article is framed reporting is ‘not offending people’. However, perhaps one of the roles of journalists is to challenge public perceptions of the causes and solutions to child abuse. It is perhaps easier to see child abuse as entirely explicable in terms of one ‘evil’ perpetrator, as the tabloid newspaper The Sun frames it, rather than as a public health problem that can be addressed systematically. The focus on bad individuals serves to hide the poverty and social exclusion that underpins much abuse of children (Lonne and Parton, 2014). Use of sensationalist terms like ‘evil’ also suggests that perpetrators of sexual abuse cannot change their behaviour. This is contrary to the largest meta-analysis of sexual offender treatment programmes, which found that cognitive behavioural programmes are effective in reducing recidivism rates (Lösel and Schmucker, 2005).

Perhaps the notion of the ‘evil’ perpetrator also serves to distance journalists and readers alike from the pain of child abuse. This idea of the ‘evil’ perpetrator was a common feature throughout episodic reporting in the sample. Vocabulary that was regularly used included ‘vile’, ‘depraved’, ‘monster’, ‘beast’ and so on. Use of such vocabulary could serve to not only distance the reader from the events (the notion of ‘us and them’) but also to comfort the reader by suggesting the idea that these events do not happen in their world. It suggests that the perpetrators are not part of humanity; they are ‘animal like’ and ‘beasts’. This might be comforting to readers if it makes them think of the events as not part of ‘normal’ society but as separate from it. This has the unfortunate consequence of perpetuating the idea that nothing can be done to change the issue; that it is not a social problem but instead an individual problem that has nothing to do with their lives.

It is perhaps the information that was not present in the articles, more than the reported information that may offer the most insight. The voices of children and non-offending parents are usually absent in the coverage (Lonne and Parton, 2014). Contextual information and social and structural facts known to underpin physical abuse, emotional abuse and neglect are often missing. Prevention, education and policy options are largely absent.

There were 21 articles in which children were perpetrators of sexual abuse (6.2% of the articles in which the age of the perpetrator was identified), none of which offered any thematic analysis. In self-report surveys, sexual abuse by children appears to be a more significant problem. Radford et al. (2011) comment on the basis of a large self-report survey:

It was notable that of the 0.5 per cent of under 11s, 4.8 per cent of 11–17s and 11.3 per cent of young adults aged 18–24 who had experienced contact sexual abuse as defined by the criminal law, 65.9 per cent of contact sexual abuse of children and young people (0–17) was perpetrated by other children and young people under the age of 18 rather than by adults in or outside the home.

Pathways to more accurate and productive reporting

Cognitive behavioural treatment programmes for adult sexual offenders reduce recidivism (Arkowitz and Lilenfield, 2008; Lösel and Schmucker, 2005). Success from treatment programmes for young sexual offenders appears even better as judged by recidivism rates in some studies (Reitzel and Carbonnell, 2006). There are also promising programmes that serve to reduce physical abuse, emotional abuse and neglect, for example, home visiting programmes (Nurse-family Partnership – MacMillan et al., 2009; Olds et al., 1988); parent education programmes (e.g. Triple-P – Prinz et al., 2009) and child sexual abuse prevention programmes (Mikton and Butchart, 2009). It is possible to help children to overcome the trauma of child abuse (Lonne and Gillespie, 2014), and with political and public will, the blight of poverty that underpins much child abuse and neglect can be reduced (Horowitz, 2014). More attention to these solutions in newspaper articles could make the problem seem less insurmountable to readers and journalists alike. Feature articles about examples of successful prevention and offender treatment programmes might be helpful (McCosker et al., 2014).

Although space in newspapers and journalists’ time are inevitably limited, it might prove possible to improve accessibility of relevant data and ease of access to expert opinion, including the voices of children and adults abused as children. Perhaps a better relationship between journalists and agencies with expertise can be created to improve the quality of articles so that the public can get a better sense of context, analysis and solutions to the public health problem. Hove et al. (2013) suggest that ‘child welfare advocates’ should focus less on the ways that journalists frame the issue and more on universal communications that inform people what they can do. These two approaches, however, are not mutually exclusive.

While this analysis focused on newspaper articles, it is well known that newspapers are increasingly using the web to promote their journalism. The future of journalism includes videos, links and multi-media presentations of issues, thereby expanding the potential space to cover issues. This may offer opportunities to contextualise child abuse and neglect in a broader framed narrative that includes promising solutions. In this way, newspapers could contribute more to public understanding of the phenomena, to amelioration of its consequences and, perhaps most importantly, to its prevention.

An example of the potential contribution of newspapers, published since the study period, was headlined ‘Two million children are referred to social services’. The article focused equally on neglect and abuse. It identified a range of possible causes for increased referral rates, including increasing poverty and recent welfare reforms. Interviews with experts referred to possible solutions, including, besides addressing poverty, improved recruitment, training and retention of social workers (Dugan, 2014).