No clear connections: a study of the association between news media use, public trust, and Swedish social services

News and media reports on social work have had policy and legislative consequences for social work practice in Sweden. Generally, it is often argued that media reports have negative implications for public perception and public trust towards social services. However, empirical studies supporting such claims are rare. Going back to the classic theory of cultivation regarding media e ﬀ ects, this study investigates possible associations between news media use, public trust in news, and public trust towards social services in Sweden. Analyses are based on survey data from a representative sample of Swedish adults and young adults ( n = 1 697) and linear regressions have been applied. In contrast to assumptions in the existing literature, no associations were identi ﬁ ed between the level of news use, preferred news outlet and public trust towards the social services, whereas trust in news predicted trust in social services. The ﬁ ndings challenge dominating notions on the implications of news media use and public trust in social services. In short, news reports of the social services do not simply translate into public (dis)trust. This also enable, linear regressions has been applied. In contrast to assumptions in existing literature, no associations.


Introduction
It is often assumed that news and media reports of social work and social services have a strong negative influence on public perception and trust towards the profession (Brunnberg, 2001;Kagan, 2016;McCulloch & Webb, 2020;Sjöström & Öhman, 2018;Zugazaga et al., 2006).However, the empirical evidence of these connections is surprisingly weak (cf.Shehata, 2021).With regard to media coverage of social services, analyses of news reports show that news typically focuses on high-end child protective cases and/or cases where clients have suffered harm from an intervention or lack thereof (Ayre, 2001;Blomberg et al., 2004;Brunnberg, 2001).In Sweden, such media coverage and news reports have had a role in policy and the passing of new legislation and revised guidelines for social service practices.This includes, for example, the implementation of 'Lex Sarah' which regulates whistle-blowers' rights and obligations to report on malpractice (see e.g.Hedin & Månsson, 2012).
The notion that news media influence the public is shared by many different media effect theories such as agenda-setting (McCombs & Shaw, 1972), framing theory (Gamson, 1989), the third-person effect (Phillips Davison, 1983) and the theory of cultivation (Gerbner, 1970).However, they differ regarding perspectives on how, and whom, the media is able to influence (Potter, 2012).While, for example, agenda-setting highlights media's ability to decide what to report on, and thereby influencing the agenda for public opinion and debate, the third-person effect is concerned with the phenomenon that individuals tend to believe that others are more susceptible to media messages than themselves.The cultivation theory, on the other hand, rests on the assumption that news and media influence public opinion by repeatedly presenting particular worldviews and attitudes about specific phenomena (Potter, 2012(Potter, , 2022)).The cultivation effect has been shown to be particularly pronounced in relation to trust and confidence in public institutions such as the police, courts and schools (Gauthier & Grazino, 2018;Moy et al., 1999).It is reasonable to assume that this is valid in relation to other public institutions, such as social services, as well.The potential of media messages to cultivate attitudes and opinions is further dependent on audience characteristics their choice of and trust in news media, as well as the extent of news consumption (Grabe & Drew, 2007) variables that will be empirically examined in the present study.
To date, little is known about the relationship between the public's media consumption and public trust in social services.A Scottish study suggests that public perceptions of social services might be related to the type of media outlet (McCulloch & Webb, 2020).Others have found that news media coverage could negatively influence individuals' attitudes towards disability benefits, especially when they lack direct experience (Briant et al., 2011;Happer & Philo, 2013).However, patterns in news consumption were not accounted for in these studies.Social workers themselves find media and news important in shaping negative public perceptions and distrust in them and that news reports contribute to stigmatising the social work profession (Beddoe et al., 2019;Legood et al., 2016;Staniforth et al., 2022).Similarly, Lundälv (2019) demonstrates that Swedish social workers believe that news coverage of social work is mainly negative (see also Lundström & Andersson, 2004).
Importantly, just as social services differ between countries, so does media coverage of the social services.While UK-based scholars argue that the predominately negative media portrayal of social work in the UK dismantle trust in both social workers and social interventions (Ayre, 2001;Franklin & Parton, 1991;Gaughan & Garrett, 2012).Cordoba (2017), on the other hand, concludes that the most significant feature of newspaper reports on social work in Australia is that they barely exist and if so, representations are described as relatively neutral.Similarly, Iarskaia-Smirnova et al. (2021) found that neutral or positive accounts of social work in two Russian newspapers were more common than negative ones.
Compared to their European counterparts, Swedish news media have been argued to be less critical towards social services (Brunnberg, 2001) and not as inclined to capitalise on scandals and promote polarisation (Nord, 2007).Sjöström and Öhman (2018) further argue that national Swedish news outlets might be less likely to critically investigate social services since the social services are governed by local authorities and not national bodies.However, in press converge on dramatic child protective cases, scrutiny constitutes a considerable part of the narrative (Lundström, 2004).Analyses of one of Sweden's largest broadsheets (Svenska Dagbladet) have shown that social services' interventions for youth offenders are often portrayed as insufficient and misguided (Estrada, 2004).Moreover, Blomberg (2019) concludes that media reports of social workers' poor working conditions contribute to representations of social service in crisis, and Markström et al. (2011) found that newspaper accounts of eldercare are predominantly critical.
Due to scarce empirical evidence, questions remain about whether news reports shape public trust in social services?And if so, what importance does the use of certain news media outlets and levels of use have in this process?It is also possible that the relationship between news consumption and trust in social services depends on the degree to which people trust news in general (cf.Rölle, 2017).In fact, although population trust in public administrations is essential for legitimacy, accessibility, professional status, performance and help-seeking behaviour (Christensen & Laegreid, 2005;Cordoba, 2017;Kagan & Zychlinski, 2016;Van de Walle, 2013), research specifically targeting trust towards the social services is rare.One of the few studies with such a focus was conducted in Sweden and it shows that the lowest levels of trust in social services employees were found among the most disadvantaged groups in terms of income and education, and among those who had experiences of contact with the social services (Nilsson & Landstedt, 2022).The current study expands these findings by including analyses of news consumption in relation to public trust in social services.The aim of the study is to explore if the level of news media use, the use of specific media outlets and trust in news content, are associated with public trust in the social services.News media, albeit difficult to characterise in absolute terms (Zelizer, 2004), is here understood as professional media outlets that focus on delivering information and analyses on current events that are deemed to be of interest to the general public (c.f.Stephens, 1988).

Setting
The Swedish social services and news media landscape The social services are the local authority's provider of social support.It is a tax-funded public institution and the main provider of social support within the Swedish welfare system.Typically, social services are divided into the branches of individual and family care (child protection, addiction treatment, financial aid, etc.), elder care, and disability care.In contrast to most Swedish welfare, care provided by the social services is means-tested and access is granted through the assessment of need.Social workers employed by the social services have a dual role where both investigative and protective, as well as preventive and curative, actions are included (SFS 2001:453).
The news media landscape in Sweden is diverse with an extensive number of news outlets (Nord & Grusell, 2021).Like in other Nordic countries, public service media (Swedish Radio and Swedish Television) have a relatively strong position, (Lindell et al., 2022), which is reflected in both public consumption, coverage and trust (Nord & Grusell, 2021).For the present study, it has been necessary to limit the number of studied news media.However, the seven included outlets are used by a majority of Swedish news media consumers (cf.Medieakademin, 2022) as the most popular public service outlets, daily newspapers and tabloids were included in the sample, together with local news.Drawing on the key principles in the theory of cultivation, outlets were selected based on their characteristics in terms of differences in reach (local/national), format (tv/radio/newspaper), and perceived trustworthiness (see e.g.Gauthier & Grazino, 2018;Grabe & Drew, 2007).News on social media was also included in the sample since news is consumed in high-choice media environments today, not least in highly digitalised Sweden (Djerf-Pierre & Shehata, 2017).

Swedish Radio (SR)
On an average week about 7.4 million individuals, or 70% of the population, listen to public service radio (Sveriges Radio, 2021).SR is deemed trustworthy by 70% of the Swedish population.(Medieakademin, 2022).

Dagens Nyheter (DN)
The national broadsheet DN is the biggest daily published newspaper with 1,111,000 readers on an average day (Orvesto konsument, 2022), and deemed trustworthy by approximately 47% of the population (Medieakademin, 2022).

Local news outlets
Respondents were also asked to what extent they listened to local radio news, watched local TV news, or read local newspapers since this type of media is popular among the population in Sweden (Kristensen and Nord 2021).

News on social media
Today there is also a growing number of social media that distribute news and, therefore, we asked respondents on a general level, if they consumed news on social media.

Design and sample
The present study is based on data from a repeated cross-sectional survey of a representative sample of the Swedish population (16 years or older) (SOM-instutitet, 2021).The survey covers opinions on society and media and was administered by the SOM Institute (University of Gothenburg, Sweden) and the data were collected in the fall/winter of 2019.All respondents were informed that participation was voluntary and gave their written consent before participation.Ethical approval for all research based on data from the national SOM survey, including this project, was given by The Regional Ethical Vetting Board in Gothenburg (dnr.Göteborg 130-15).
Data were collected by means of a survey distributed by post and through a web-based platform.In total eight reminders, four via post and four via text messages, were sent out between 25 September 2019 and 26 November 2019.The response rate was 50% (n = 1 697) (Falk et al., 2020).The survey covers opinions on society and media.All included items and instruments have been validated (Falk et al., 2020).

Dependent variable
The dependent variable trust in the social services was operationalised via the question: 'Generally speaking, to what degree do you trust how the following institutions and groups do their job: the social services'?A five-point Likert scale response format was used and response options were (1) to a very low degree, (2) to a low degree, (3) neither high nor low degree, (4) to a high degree, (5) to a very high degree.The response rate for the question was 89.3%, n = 1516.

Independent variables
The included independent variables were chosen based on dimensions of news media use that are considered to be of importance for news to influence individuals: exposure to content (news use and preferred news outlet) (Gauthier & Grazino, 2018;McCulloch & Webb, 2020;Potter, 2012;Wakefield et al., 2010) and trust in content (personal views on news trustworthiness) (Grabe & Drew, 2007;Happer & Philo, 2013;Lee, 2005;Potter, 2012).
Exposure to content.Respondents were asked to rate how often they watch, read, or listen to the news media outlets listed: (1) Never, (2) Less than one day per week, (3) 1-2 days per week, (4) 3-4 days per week, (5) 5-6 days per week, (6) Every day.For local news, a corresponding index ranging from zero to six was computed.
Trust in news content.A composite index was computed of the following statements rated on a scale from (1) Disagree to (5) Agree: News is in general very trustworthy; Swedish media often report falsely; I do not trust that news is impartial; Journalists often twist news according to their own beliefs, and Swedish media do not tell the whole truth in relation to important societal issues.The index had a Cronbach's Alpha of 0.88 and ranged from 5 (disbelief of content) to 20 (belief of content).

Covariates
Based on previous research regarding factors influencing trust in social services and public administrations (Blind, 2006;Nilsson & Landstedt, 2022;Putnam, 2000;Rothstein & Uslaner, 2005) the following control variables were included: Respondents were asked what level of education they had and given eight response options from (1) Less than primary school to (8) Ongoing PhD studies or PhD degrees.The variable was recoded into four categories: low (values one and two), medium-low (values three and four), medium-high (values five and six) and high (values seven and eight).
Previous contact with the social services was operationalised through the question: 'In the last twelve months have you, for your own benefit or for the benefit of a close relative, been in contact with the social service?'.The response options were: (1) Yes, for my own benefit, (2) Yes, for the benefit of a close relative and (3) No.
General feelings of trust.Respondents were asked to rate if individuals, in general, can be trusted on an eleven-point (range: 0-10) Likert scale, 0 = 'In general, it is not possible to trust other individuals', 10 = 'In general, it is possible to trust other individuals'.
Corruption in public administration was measured through the question 'According to you, to which extent does any form of corruption occur in the following sectors?':Health care, Construction, Restaurants and Schools.Response options ranged from zero to eight where higher indicates more perceived corruption.A corruption index (Cronbach's Alpha .84)including the above questions ranging from zero to 28 (high corruption) was constructed and used in the analysis.
To assess satisfaction with democracy respondents were asked how satisfied they were with democracy in Sweden.They were given four response options ranging from (1) Completely dissatisfied to (4) Completely satisfied.

Statistical analysis
Chi-squared statistics were used in descriptive analyses.Associations were estimated via linear regression.Only covariates that were significant in descriptive analysis or unadjusted models or made considerable impact on the regression coefficients, were included in the final model.All analyses were performed in Statistical Package for the Social Sciences (SPSS) version 26.

Patterns of trust
The most frequently indicated response regarding trust in social services was 'neutral', which is also reflected in the mean level of trust (Table 1).Approximately one-third stated that they have a quite low or low sense of trust, while almost 17% reported a quite high or high sense of trust.
No age or gender differences in trust were identified.Higher trust was reported by those who had higher levels of general trust in others (n = 193; 76.9%, p < .001)and those who were satisfied with democracy (n = 200; 79.1%, p < .001).Lower trust was reported by those with higher income levels (n = 64; 14,3%, p = .014),those who displayed higher levels of perceived corruption (n = 319; 68.2%, p < .001)and those who had been in contact with the social services on behalf of a close relative (n = 37; 38.7%, p < .001).(The results from group comparisons are not shown in tables although provided upon request).Statistically significant results, indicating that those who report higher trust in news also report higher trust in social services, were retrieved for all included measurements on general trust in news media (see Table 2).

News media and trust in social services
Trust in news media was significantly and positively associated with trust in social services.However, none of the included news media outlets were associated with trust.The adjusted model explains approximately 17% of the variation in the dependent variable (Table 3).When excluding the media trust index the explanatory power of the model drops to approximately 13% (Adj R 2 = .131).Consequently, about 4% of the model's ability to explain the level of trust is accounted for by the news media trust.Diagnostics of the model revealed no problems with multicollinearity (see Table 3.), autocorrelation (Durbin Watson = 2.039) or level of extreme outliers, 0.13%, ± 3 SD (see e.g.Field, 2017).
Interaction analysis between all included news outlets, media trust and trust in the social services revealed no significant results except for the biggest tabloid Aftonbladet (data not shown in tables, the results available upon request).Higher use of Aftonbladet impacted positively on the relationship between news media trust and trust in social services.Sensitivity analyses indicated that media trust was associated with trust in other human services organisations (HSOs).(The results provided upon request).These results also show that the use of different news outlets and the level of use can be of importance for trust in HSOs in general.

Discussion
This study shows that most individuals neither trust nor distrust Swedish social services.However, there is a greater proportion that explicitly states a low level of trust in Swedish social services compared to a high (31.6%respectively 16.7%).In comparison, more than half of the population report trust in healthcare (Swedish Municipalities and Regions, 2023) and the police (The National Council for Crime Prevention, 2022).The lower level of trust in social services is in line with previous research suggesting that means-tested services are less trusted than universal ones (Rothstein, 2003;Jessen, 2010).
The results of this study indicate, in contrast to assumptions in the existing literature and the theory of cultivation (Gaughan & Garrett, 2012;Grabe & Drew, 2007;Lundström & Andersson, 2004;Kagan, 2016) that neither level of use of news media, nor preferred type of news outlet, matter for level of trust in the social services.Our findings, therefore, suggest that news media reports might not be as influential on public perception and trust in social services as previously stated, at least with regard to public trust in social services in Sweden.This might be related to the characteristics of the Swedish media landscape.Compared to their European counterparts, Swedish tabloids are e.g. less inclined to scandalise and create narratives based on polarisation and drama, which in turn tends to influence the consumers more than more balanced narratives (Nord, 2007;Markström et al., 2011;Kepplinger et al., 2012).Our findings partly confirmed this; higher use of the tabloid Aftonbladet impacted positively on the relationship between news media trust and trust in social services.Previous research (Sjöström & Öhman, 2018) further suggests that Swedish national news media are less likely to critically scrutinise social services since they are locally governed.This might contribute to less criticism towards social services within the Swedish news media, which could explain our findings, at least partly.
Our findings could also be understood in relation to research indicating that social services might not be as poorly depicted in Swedish news as in, for example, the UK (Sjöström & Öhman, 2018).At the same time, other researchers illustrate how representations of social services in Swedish media are primarily negative (Blomberg, 2019;Lundström, 2004;Markström et al., 2011) and that social workers in Sweden, as well as in other countries, generally understand news media coverage of them as negative (Legood et al., 2016;Lundälv, 2019;Staniforth et al., 2016).Furthermore, given the social policy effects of Swedish national news media coverage (e.g.Lex Sarah), it is questionable if news media, in general, disregards critical investigations of local incidents.It is possible that for national news media to investigate (local) social services, the incident must be particularly severe and upsetting (Ayre, 2001).Such reports might influence public trust negatively even if the presence of long-term effects is unclear.Previous analyses (Nilsson & Landstedt, 2022) also reveal that trust towards social workers is stable over time, despite multiple reports of high-end child abuse cases in the Swedish news during the time of the study (cf.Djerf-Pierre & Shehata, 2017;Nord & Grusell, 2021;Lindell et al., 2022).
The key finding of this study is that trust in social services is primarily associated with trust in news media in general, also when adjusting for other aspects known to be of importance for trust.In line with notions of the theory of cultivation that individuals are more likely to accept and believe messages they perceive to be real from actors they trust (Gauthier & Grazino, 2018;Busselle & Greenberg, 2000), our findings challenge the idea that news and media reports have direct damaging effects on the social services.Rather, it points to the oppositeindividuals who trust news content are also more likely to trust social services.This result corresponds with Rölle (2017), showing that higher trust in news is associated with greater trust in public administration in general (cf.Lindell et al., 2022).
The present study indicates that news media reports and social workers' assumptions cannot automatically be assumed to reflect public trust in social services.This warrants a more nuanced understanding of the relationship between news and public trust towards social services that also recognises that the public might be more resilient towards 'negative' news coverage than what has previously been assumed (see also McCulloch & Webb, 2020).

Limitations
One limitation of this study is the high rate of indifferent responses concerning trust in social services.It is not possible to know why 50% of respondents indicate neither a high nor low level of trust.However, it is likely that the indifference is due to the nature of the social services as such.Only a limited number of people personally encounter social services, and individuals might not invest in forming an opinion concerning a service they do not perceive as relevant to them.The general focus on social services implies that branch-specific differences are not visible.However, asking branch-specific questions would most likely have generated even lower response rates.In addition, when social work is discussed in news and other public forums the term 'social services' is the most commonly used.Using the most common concept outside academia is always preferred when designing survey questions (Gideon, 2012).This applies also to the survey from which data were drawn for the present study.All items, including the item measuring trust in social services, were tested and validated before inclusion in the survey (Falk et al., 2020).Another limitation is that the cross-sectional design only gives limited insights into the causal pathways between exposure, audience characteristics and attitudes highlighted as influential within the theory of cultivation (Gauthier & Grazino, 2018).However, this study constitutes a starting point for understanding better the relations between public trust towards social services and aspects of news media use within that framework, even if more empirical studies with a longitudinal design are needed.
Finally, due to the importance of context regarding the diversity of news media environments (Ahva et al., 2017), trust (Putnam, 2000) and how welfare and social services are organised across countries (Esping-Andersen, 1990), generalisations of the findings outside the Nordic, or similar European, countries should be done with caution.

Implications and concluding remarks
Our findings indicate no association between news media use and public trust in social services.The presumed effects of higher news media consumption outlined in the cultivation theory are thus not visible in relation to the social services.However, trust levels varied in relation to, for example, level of income and perceived corruption, which are in line with the proposition that audience characteristics are important.The significance of trust in news is also supported as it predicted trust in social services (cf.Grabe & Drew, 2007).This suggests that it is time to review previous dominating notions, within academia and the social work profession, about the relationships between news, public trust, and social services.News reports do not simply translate into public (dis)trust.Given social workers' previous hesitance to take part in public debates about social work and social issues in Sweden and elsewhere (Lundälv, 2019), our findings open up for social workers to raise their public voices since the consequence of doing so appear not as 'risky' as expected.Social workers' perspectives on their profession, service users and social work are, on the other hand, highly relevant and important, both for public understanding and for social workers' abilities to push for the consideration of their standpoints.
The findings indicate that there might be other aspects than media use and trust in news that are of importance for trust in social services.Given that low trust can influence legitimacy, accessibility, and individuals' help-seeking behaviour it would be valuable to undertake in-depth investigations on how age, gender, level of income and education or experiences of the social services affect trust.Importantly, analyses of representations of the social services in the news coverage would also provide key insights regarding the correlation between news media, professional understanding, and public trust in the social services.Pseudo-experimental research or focus groups using news on social services as stimuli could also contribute new knowledge on how individuals interpret and react to, for example, news where scrutiny constitutes the narrative.Research on the content of news, news consumption and trust across specific branches of social services would also be valuable to deepen our understanding of connections.In addition, research is needed to advance knowledge concerning social workers' experiences of the role of media regarding representations of social services and public trust, including how it might impact their daily work or create feelings of stigma (cf.Beddoe et al., 2019).
Worth noting is that in the Swedish context, it seems like the level of trust in social services is lower than the trust in social workers (Nilsson & Landstedt, 2022).One explanation could be a dissonance between the perceived objectives of the social services as an organisation, versus social workers as providers of support.As pointed out by McCulloch and Webb (2020), high levels of trust towards social workers imply that they have a licence to operate.However, our findings indicate that this does not mean that social services automatically are perceived in the same way.Given that news media might not be the nemesis of Swedish social work as previously believed, perhaps news media could rather be viewed as a resource?Much focus on social services in relation to news has previously been on media management to promote organisational self-interest, usually through a representative with a specific responsibility for media relations (Enbom et al., 2014;Sjöström & Öhman, 2018).A constraint for frontline social workers' public participation is, of course, that they always have to protect their clients and not reveal any sensitive information about them.Consequently, as Lundälv (2019) shows, very few social workers are actively taking part in public debates about social work and social services.We argue that public trust towards social services could be improved if social workers to a greater extent engaged in broader public debates and discussions on social issues and social work.This ought to be possible without revealing sensitive information and news media constitute a platform for such engagement.It could also promote broader reflections on professional practice (Sicora, 2019), ameliorate public knowledge and raise awareness of organisational issues constraining client-centred social work (cf.Hedin & Månsson, 2012), and ultimately enhance trust and improve conditions for both social workers and their clients.

Table 1 .
Trust in the social services 2019.

Table 2 .
Frequency in reported news media use in absolute and relative numbers.

Table 3 .
Final regression model on trust in the social services including news media use and trust and control variables.