Public Understanding of Childhood Obesity: Qualitative Analysis of News Articles and Comments on Facebook

ABSTRACT Childhood obesity has become a major focus of public health and subject to increased news coverage. News can shape public understanding of childhood obesity by selective reporting (framing) which can affect policy support and weight stigma. As news is consumed on social media, comments on articles present a novel method to explore public understanding. This study examined how childhood obesity is framed by news articles on Facebook and how individuals commenting understand and react to these articles. This study used a qualitative research design. Facebook pages of 11 national UK news outlets were searched for news articles on childhood obesity published between May 2015 and May 2020. Of those, 30 articles were randomly selected. Framing analysis was used to determine whether childhood obesity was portrayed as a behavioral, societal or medical issue. Responding comments (N = 1,104) were grouped according to the dominant frame of the corresponding article and analyzed using qualitative content analysis. Of the 30 articles, 28 mentioned societal, 26 behavioral and 18 medical aspects of childhood obesity, highlighting that most articles included more than one frame. Three themes were generated from responding comments: Culprits and Remedies, Appraising Childhood Obesity and Making Sense of the News Article. Findings showed that comments related to Appraising Childhood Obesity differed between differently framed articles, while the other themes did not. This study highlights the need for improved communication on childhood obesity to address weight stigma and improve understanding of news articles.


Background
In the United Kingdom (UK), 1 in 10 children aged 4 to 5 and 2 in 10 children aged 10 to 11 were classified as obese in 2018/ 2019 (NHS, 2019). Childhood obesity is associated with an increased risk for depression during childhood and the development of physical illnesses, such as type 2 diabetes or coronary heart disease, in adulthood (Llewellyn et al., 2016;Sutaria et al., 2019). Thus, research and public health efforts are increasingly focused on understanding and targeting childhood obesity (Department of Health and Social Care, 2018;Tran et al., 2019).
Childhood obesity has similarly been subject to increased news article coverage (Nimegeer et al., 2019). As news articles are a major information source for the public (Hodgetts & Chamberlain, 2006;Naidoo, 2016), they might influence public understanding and policy support (Barry et al., 2013;Frederick et al., 2016Frederick et al., , 2019Saguy et al., 2014;Sun et al., 2016). While previous research has shown how childhood obesity is described by news articles (Nimegeer et al., 2019) and discussed in social media posts (Harris et al., 2014;Kovacs et al., 2018), little is known about the portrayal of childhood obesity in social media news stories and the resulting public response. Additionally, most research has focused on portrayals of adult obesity and a gap exists regarding specific findings on childhood obesity (Nimegeer et al., 2019). Thus, this study will take a novel approach to investigate public understanding of childhood obesity by gathering news articles and responding comments on social media -thereby capturing natural sensemaking and integrating two types of data. The following sections will summarize the existing literature on news reporting on childhood obesity, draw on framing theory to explain the influence of news reporting and highlight how previous research has utilized social media to gather public understanding on obesity and childhood obesity.

News reporting and public understanding of childhood obesity
News articles on childhood obesity often use emotive language and evoke parent blaming (De Brún et al., 2013;Nimegeer et al., 2019), as evidenced by a Daily Mail article headlined "Scientists blame working mothers for Britain's childhood obesity epidemic" (Powell, 2019). News articles play a vital role in disseminating research findings and health policies (Hodgetts & Chamberlain, 2006;Naidoo, 2016). However, news articles do not act merely as the megaphones of scientists or policymakers, but are aimed at attracting readership (Hodgetts & Chamberlain, 2006;Schwitzer et al., 2005). Previous research has found that news articles sensationalized research on obesity and reported findings without questioning them critically (Saguy & Almeling, 2008). In the context of childhood obesity, news articles have been found to place greater focus on individual responsibility compared to scientific studies (Bastian, 2011). This suggests that the public may not be presented with an accurate account of research on childhood obesity, which could be problematic if the public builds their understanding of childhood obesity on these news articles.
Previous research into public understanding of news on obesity and childhood obesity has examined comments on news outlets' online pages (Brooker et al., 2017;Gregg et al., 2017). One study explored newspaper headlines and comments in response to the UK's childhood obesity strategy (Gregg et al., 2017). Findings showed that both comments and headlines adopted a primarily negative stance toward the strategy. Commenters debated responsibility for childhood obesity, with a focus on parental responsibility, but also acknowledging the role of individuals, schools and the government. Additionally, commenters voiced anecdotal solutions, such as changing societal norms around food or nutritional education, and expressed worry about the government prioritizing industry interests over children's health.
A further study analyzed comments on an article covering European Union (EU) legislation on obesity as a disability (Brooker et al., 2017). Emphasis was placed on personal responsibility and obesity being a result of negative lifestyle choices. Commenters further debated whether weight was a simple issue of exercise and diet or whether it was more complex, with comments highlighting challenges to weight loss and questioning weight as an indicator for health. These studies (Brooker et al., 2017;Gregg et al., 2017) suggest that public understanding of news articles on childhood obesity is interlinked with public understanding of childhood obesity. However, these findings are observational and do not provide a framework for explaining the influence of news articles.

Framing: Theory and public understanding
One theory on how news articles influence public understanding is framing theory (Chong & Druckman, 2007;Entman, 1993;Scheufele, 1999). Framing theory posits that an issue can be viewed from different perspectives and therefore be constructed to have different meanings (Chong & Druckman, 2007). Obesity, for example, can be constructed as a chronic disease, a source of discrimination or the freedom to consume whatever one wishes (Kwan, 2009). These different meanings are referred to as frames. Frames are constructed in the news by framing -the process of selectively reporting certain aspects of an issue (Entman, 1993). This increases the salience of these aspects while other aspects are omitted and different interpretations are inhibited. For example, reporting on the physiological health consequences of obesity inhibits alternative approaches to health, such as focusing on health behaviors (Kwan, 2009). Approaches focusing on health behaviors promote physical activity as well as fruit and vegetable intake, placing the focus on increased physical wellbeing rather than on weight loss (Clifford et al., 2015;Dugmore et al., 2020). UK news articles have been found to commonly define childhood obesity as problematic due to its prevalence, its rising rates and associated health risks (Nimegeer et al., 2019). Additionally, news tends to endorse individual causes and solutions more often than societal or biological ones (Nimegeer et al., 2019). While understanding the type and content of framing present in news articles about childhood obesity is useful, it is equally important to understand how these frames influence public understanding.
According to framing theory, not every individual will be influenced by news framing in the same way, as individuals interpret them in the context of their pre-conceptions (Entman, 1993;Scheufele, 1999). However, adopting a certain frame can influence an individual's cognitions, attitudes and behavior (Scheufele, 1999). Experimental studies have shown that framing obesity as a public health crisis rather than using an approach focusing on health behaviors, leads to an increased perception of the health risks of obesity and increased negative attitudes toward individuals with obesity (Frederick et al., 2019;Saguy et al., 2014). Additionally, news articles endorsing societal, compared to individual, frames of obesity seem to influence attributions of responsibility respectively (Barry et al., 2013;Sun et al., 2016). Whether news framing influences support for obesity policies is less clear and studies have reported mixed findings (Barry et al., 2013;Frederick et al., 2016Frederick et al., , 2019Saguy et al., 2014). These experimental studies indicate that framing affects public understanding of childhood obesity. However, they do not provide insight into how individuals make sense of and react to news articles in a natural setting.

Gathering public understanding on social media
A recent survey showed that 40% of UK respondents (N = 2,023) reported consuming news on social media in (Newman et al., 2019. Thus, public understanding of and reactions to news articles can be explored using comments on news articles on social media. These comments are expressed in a natural setting (Sugiura et al., 2017) and can be used to explore public understanding of and reactions to news articles. Su et al. (2020) explored whether news framing of obesity influences the support (notions of inclusivity, positive affect, and encouragement) and stigmatization (discrimination or derogatory jokes) expressed in comments on social media. The results showed that while defining obesity as a health risk, burden to health services and economic cost did not lead to significantly more stigmatizing than supporting comments, defining obesity as a cosmetic issue did. Attributing obesity to the environment, biology, genetics and parenting led to significantly more supporting than stigmatizing comments, while no significant difference was observed for attributions to personal responsibility. While these findings investigate framing in a natural setting, they do not account for the nuances of response that individuals could display. Thus, more qualitative enquiries are needed.
Previous qualitative studies have examined comments on social media unrelated to news articles (Chou et al., 2014;Harris et al., 2014;Kent et al., 2016;Kovacs et al., 2018). These studies showed that public understanding of obesity and childhood obesity revolves around responsibility, causes, solutions, expression of stigma, as well as support and countering this stigma (Brooker et al., 2017;Chou et al., 2014;Gregg et al., 2017;Harris et al., 2014;Kent et al., 2016). Understanding of childhood obesity entails additional nuances such as the influence of parents and schools (Gregg et al., 2017;Harris et al., 2014;Kovacs et al., 2018). Governmental efforts to address childhood obesity were perceived as useful by some (Harris et al., 2014;Kovacs et al., 2018), but also met with critique and mistrust (Gregg et al., 2017;Kovacs et al., 2018). This critique concerned the collection and use of weight data gathered from children, as well as the adequacy of government involvement in parenting (Kovacs et al., 2018). While this provides a useful starting point on public perception of childhood obesity, little is known about how news articles on childhood obesity are received on social media and whether this is influenced by news article framing.
This study aimed to advance current knowledge on public understanding of childhood obesity. The research questions posed were (1) How is childhood obesity framed by UK news articles on Facebook? and (2) How do individuals commenting express understanding of childhood obesity in comments on these articles?

Research design
This study used an experiential qualitative design exploring textual data. This design was chosen as it allowed us to examine how understanding might be socially constructed in a natural setting. Facebook was chosen for data collection as it is the most common social media platform for consuming news (Newman et al., 2019). Facebook is a social media platform created in 2004 (Facebook, n.d.a) where users can create posts, comment on other user's posts and exchange messages (Franz et al., 2019).

News article and comment selection
Eleven UK national news outlets (see Table 1) were selected to obtain a broad range of news coverage including various political orientations, reporting styles and target audiences.
News outlets were categorized based on political orientation (centrist, left-wing, right-wing) and type (broadsheet, tabloid) ("Daily Star (United Kingdom)," (2021); "List of Newspapers in the United Kingdom," 2021). This resulted in six news outlet categories, with the news outlets selected covering five. The classification of news outlets into news outlet categories is presented in Table 1. Facebook pages of news outlets were manually searched for posts using twenty search terms, resulting from all possible combinations of an agespecification (children OR child OR childhood OR kids OR kid) and an obesity-related word (obesity OR overweight OR obese OR fat).
Two separate maximum sample sizes for news articles and comments were determined using previous studies as guidance. A previous study on news articles and comments on weight loss surgery used 19 news articles (Glenn et al., 2012). Therefore, a sample size of 30 articles was deemed to include sufficient data for analysis. The maximum sample size for comments followed a previous study that included 100 comments per article (Laestadius & Caldwell, 2015). To obtain this number, the first 100 eligible comments per news article were extracted. Comments on Facebook news posts were sorted as "All comments -Show all comments. including potential spam. The most relevant comments will appear first" (Facebook, n.d.b). Responses to eligible comments were also extracted but did not count toward the sample size of 100 comments per article.
News articles were screened by the lead researcher (BB) using the following inclusion criteria: Articles must be: (1) textbased; (2) in English; (3) published since May 1 st 2015 and May 1 st 2020; (4) focused on childhood obesity; and (5) responded to by at least one eligible comment. After an initial screening, N = 162 articles were obtained. The results of this first screening were discussed, and some articles (n = 64) were removed due to not fulfilling criterion (4) as judged by both researchers (BB and ESM). Next, articles were grouped based on news outlet category (see Table 1). In each category six articles were randomly selected using the "sample" command in R Studio (version 1.2.1335). While the random selection process was based on news outlet category, all news outlets were represented in the final sample. This random selection process led to three articles needing to be resampled, as two articles could not be accessed on Facebook at the time of comment extraction and one article did not fulfil criterion (5).
Subsequently, comments on those 30 articles were screened by BB using the following inclusion criteria: Comments must be: (1) in English; (2) posted within one month of the article; (3) text-based (comments just containing gifs, emojis or memes were excluded); and (4) related to childhood obesity. After an initial screening N = 1,184 comments were obtained. The screening results were discussed, and some comments (n = 80) removed due to not fulfilling criterion (4) as judged by both researchers. The final sample contained 30 news articles ranging from 239 to 1070 words in length and 1104 responding comments ranging from 3 to 99 comments per article. The data screening process with the number of articles and comments per news outlet category can be seen in Figure 1. News articles were numbered 1 to 30 and references are provided at the end of the article. The news outlet category broadsheet-centrist was not covered by any news outlet selected.

Data collection
News outlets' Facebook pages were searched between 05.05.2020 and 14.05.2020, with articles extracted on 15.05.2020. All data was extracted manually to comply with Facebook's Terms and Conditions (Facebook, n.d.c) and to avoid extraction of personally identifiable information (e.g., commenter names). Data collection was undertaken by BB.

Data analysis
NVivo v12 (QSR International, 2018) was used to organize data analysis. All coding was conducted manually. This study used passive analysis as articles and comments were created prior to this study and without the influence of the researchers (Franz et al., 2019). However, the research design and the analysis are influenced by the lead researcher's background.

Researcher description
The lead researcher (BB) conducting this study was a 24year-old white, female health psychology Masters student from Germany. Since health psychology has historically focused on public health and changing health behaviors (Murray, 2014), the researcher might be more likely to identify behavioral causes and solutions to childhood obesity. To be aware of these influences, the researcher kept a reflective journal documenting thoughts and ideas throughout the study.

News articles
News articles were analyzed deductively using framing analysis (Foley et al., 2019). The aim of framing analysis is to determine which information is made salient, thereby promoting a certain understanding (Entman, 1993;Foley et al., 2020). At first, all news articles were read to promote familiarization (Foley et al., 2019). Then, nodes in NVivo were set up to represent relevant frames (behavioral, societal and medical) adopted from Stefanik-Sidener (2013) and relevant frame subcomponents (problem definition, moral evaluation, causal attribution and treatment recommendation) adopted from Entman (1993). Operational definitions of frames and frame subcomponents are presented in the Appendix (Table A1). Article fragments were then deductively coded into one or more subcomponents under the most appropriate frame (Foley et al., 2019). After two rounds of coding by BB, a second researcher (ESM) reviewed article fragments coded under one frame (behavioral), and both researchers discussed the coding.

Comments
Comments were analyzed using qualitative content analysis (Erlingsson & Brysiewicz, 2017). The aim of qualitative content analysis is to use textual data to infer meaning (Erlingsson & Brysiewicz, 2017;Hsieh & Shannon, 2005). Qualitative content analysis typically presents data in words and themes, especially when the analysis focuses on latent meaning (Bengtsson, 2016). Since the present study focused on the latent content rather than tone or speakers as in Glenn et al. (2012), it was deemed most appropriate to present the results with words and themes only rather than presenting prevalence data. Comments were analyzed using a combination of induction and deduction. Firstly, all comments were read to promote familiarization. Then, a coding scheme was inductively developed. Both researchers independently coded comments on articles from one news outlet category. The category Tabloid-left-wing was chosen as it contained the largest number of comments, thereby allowing for comprehensive coding. After coding, both researchers collaboratively developed a coding scheme. Comments were grouped depending on the dominant frame (behavioral, medical, societal) of the article they responded to. The dominant frame was determined by assessing which frame covered the largest proportion of the article and reflecting whether this frame was the most salient. Next, all comments were deductively coded by BB using the coding scheme, adding codes if necessary and discussing the coding with ESM. The researchers used ad hoc unitization to code comments. The minimum unit of coding was a single word, the maximum an entire comment. The researchers could assign comments to none, one or multiple codes if appropriate. After two rounds of coding, categories and themes were generated separately for comments on differently framed articles. In this process, researchers could assign each code to one subtheme only.

Ethical considerations
Ethical approval was gained from the Research Ethics Approval Committee for Health at the University of Bath (Reference number: EP 19/20 025). The need to obtain informed consent was waived, as comments on publicly available news articles were deemed as public space where individuals would expect to be observed (British Psychological Society, 2017;Burkell et al., 2014). To protect comment authors, no personal identifiable information was extracted. Additionally, all illustrative comment quotes presented in the findings section are paraphrased versions of the original comments (British Psychological Society, 2017). These were reviewed by ESM and doublechecked using a Google search engine (Rodham & Gavin, 2006).

Methodological integrity
This study considered reflexivity, sensitivity to context, rigor, transparency and relevance (Malterud, 2001;Mays & Pope, 2000;Yardley, 2017). To promote reflexivity and sensitivity to context, the researcher kept a reflective journal, documenting thoughts arising throughout data collection and analysis. Additionally, the researcher detailed expectations before the analysis and utilized them to evaluate its rigor. The researcher was further supervised by ESM with frequent discussions about data collection and analysis. The criterion of transparency was implemented by clear reporting, while relevance was given by methodological novelty and implications for health communication.

RQ1: How is childhood obesity framed by UK News Articles on Facebook?
Almost all (n = 27) news articles contained elements of more than one frame, with 15 including elements of all three frames (medical, behavioral, societal). The societal frame was most prevalent (n = 28), followed by behavioral (n = 26), and lastly medical (n = 18). The number of news articles coded at each frame and frame subcomponents is presented in Table 2. An overview of the content of article fragments coded at different frame subcomponents is presented in Table 3.

Medical frame
Within the medical frame, childhood obesity was defined as a problem due to associated health risks such as "diabetes" (Articles 1, 15, 26 & 28), which were morally evaluated as "cruel and avoidable" (Article 7). Articles reported medical causes of childhood obesity, such as breastfeeding or epigenetics uncovered by scientific research (Articles 2, 26 & 30). Medical treatment suggestions were seldom (n = 4), although weighing children at health appointments and bariatric surgery were suggested (Articles 23 & 27).

Societal frame
More than half (n = 17) of the news articles used prevalence statistics to define childhood obesity as a societal problem. Causal attribution and treatment suggestions centered around the government setting the framework for healthy choices (Articles 3,7,17 & 20). Moral evaluations represented the biggest share of the societal frame (n = 20), with childhood obesity being evaluated as a "scandal of child ill-health" (Article 15) or a "ticking time-bomb for the NHS" (Article 1). Additionally, contributions to addressing childhood obesity were evaluated with frequent demands for politicians to be "bold" (Articles 3 & 15), and strategies criticized as "weak" (Articles 17 & 28).

Behavioral frame
Within the behavioral frame, childhood obesity was defined as a problem due to an individual being "heavier than average" or "biggest" (Articles 23 & 29). Most article fragments in this frame reported causal attributions (n = 22). These were mainly focused on physical activity and diet, as well as parenting "We find that children whose mothers work are more likely to have increased sedentary behavior and poorer dietary habits" (Article 6). Therefore, parents, especially mothers were frequently morally evaluated. Additionally, treatment suggestions mainly revolved around diet and change coming from parents "Nutrition professionals must consider the influence of family and siblings to provide appropriate and tailored nutrition education" (Article 13).

RQ2: How do individuals commenting express understanding of childhood obesity in comments on these articles?
Across differently framed articles, comments revolved around three themes with three subthemes each (see Figure 2). The first theme, Culprits and Remedies, explores perceived causes and solutions to childhood obesity. These were represented very similarly across comments on differently framed articles and topical subthemes were created. The second theme, Appraising Childhood Obesity, illustrates whether commenters problematized childhood obesity. These appraisals varied between comments on differently framed articles and are presented in three subthemes. The last theme, Making Sense of the News Article, explores how commenters understood and perceived news articles. There were similarities across comments on differently framed articles, which will be explored in topical subthemes.

Culprits and remedies
This theme was generated across comments from all frames and discusses perceived causes and solutions for childhood obesity. These varied in their scope of influence. They ranged from immediate behavioral influences, to influences outside of behavior but still close to the child, and influences acting on a societal level. This is described in the three subthemes You Are What You Do, Looking Close to Home, Seeing the Bigger Picture.
You are what you do. This subtheme illustrates commenters' views of obesity as originating from an individual's diet and physical activity behavior On the whole if you eat too much and you don't do exercise you get fat. Regarding diet, commenters perceived specific types of foods as contributing to childhood obesity such as junk, processed, fast food and sugary drinks and food. In contrast, home-cooked meals were often seen as healthy and protective. Some commenters, however, believed However, the bulk of responsibility was attributed to parents. As a result of seeing the causes of childhood obesity in behavior, individual interventions such as fat camps or nutritional education were voiced. Comments on behaviorally framed articles additionally mentioned restricting children's access to food Padlock the cupboards, padlock the fridge, don't give them money so can't buy junk from the shop. These comments often used highly derogative language such as Slap a muzzle on the greedy sod. At times quite graphical, these comments depicted a sense of readiness to at least virtually assault children with obesity.
Looking close to home. This subtheme explores influences outside of behavior, but close to the child, such as the influence of medical conditions, family and schools. Firstly, commenters voiced medical reasons for childhood obesity, such as genetics or medical conditions, as well as emotions and mental health conditions. These influences were often met with more acceptance There are conditions causing this. Don't judge what you don't know.
Secondly, children were not seen as autonomous individuals, but highly influenced by their family and school. Parents especially were viewed as guardians of children's waistlines Preferably the parents teach them to eat healthy to start with and encourage them to exercise, with children almost inevitably being the weight of their parents fat parents = fat kids. Many comments debated parental effort and influence on children's health behaviors. On one hand, providing children with a healthy lifestyle was described as easy. On the other hand, commenters discussed challenges, such as working full-time or living on a budget. Parents of children with obesity were evaluated depending on the stance adopted. Commenters endorsing great parental choice expressed that having children with obesity was a sign of parental failing and parents should feel ashamed. More extreme views even compared childhood obesity with neglect or abuse Fat kids undergo some kind of child abuse -by their families and demanded that social services were involved. However, these extreme views were only expressed in response to societally and behaviorally, but not medically, framed articles. Nevertheless, commenters acknowledging parental challenges expressed sympathy for parents trying to do the best they can with what they have.
Lastly, schools were perceived as responsible for providing healthy nutrition and physical activity. Commenters claimed that, currently, school meals were not sufficiently healthy Make schools give children nutritious food and not the junk they do now! Additionally, commenters stated that schools could improve child health by increasing physical education, education about nutrition and cooking classes.
Seeing the bigger picture. In this subtheme, commenters voiced societal influences on childhood obesity, such as policy, financial pressures, and the food industry. Limitations on individual choice were acknowledged It is not fully down to individual situations there are many societal factors at play such as tv, advertising, societal pressures, government policies and so on. Some commenters voiced that current societal trends were limiting healthiness and thus contributing to childhood obesity. Planned selling off of children's safe playing fields was said to reduce opportunities and safety for outside play. Furthermore, societal influences on parents were explored With the many roles that society and financial pressures put on individuals these days, is it surprising that people go for 'fast' food? Commonly, family income, lack of time and support for parents were discussed as barriers to providing healthy lifestyles for children.
Commenters discussed societal stakeholders such as the medical profession, the food industry and the government. Healthcare professionals were criticized for their handling of child weight. Specifically, weight assessments using body mass index (BMI) were criticized It does not account for levels of activity, body type, muscles and medical conditions. A further stakeholder, the food industry, was said to crucially influence an individual's food environment I blame the availability of junk on the shelves being at child eye level and also cheap it's horrific. This food environment was characterized as unhealthy due to high costs of healthy foods, ubiquitous advertising and availability of unhealthy foods. Consequently, the government was criticized for not setting the right framework for health behaviors by accommodating the vested interests of the food industry, and not providing enough support or education. One commenter voiced With this government, the burden and inconvenience are consistently forced onto people and not the food manufacturers.
As a result of seeing the causes as societal, commenters recommended governmental strategies such as banning or taxing unhealthy foods, subsidizing healthy foods, as well as controlling food advertising. One comment expressed Maybe if we could cap the growing number of fast food outlets, ban all advertising of unhealthy drinks and food.

Appraising childhood obesity
This theme describes how commenters conceptualized childhood obesity. Differences between comments on medically, societally and behaviorally framed articles were evident. Therefore, separate subthemes for each of the conceptualizations were created.

As long as the child is healthy.
Comments on medically framed articles discussed what constitutes a healthy child. Regarding infant feeding, many commenters asserted that a fed baby is a healthy baby, while only a minority of commenters viewed that infant obesity might lead to future ill health. Moreover, comments expressed that weight should not be used as a sole indicator for child health, as not everybody who weighs the same actually has the same body shape and has the same muscle/fat ratio. The definition of health adopted by commenters was linked to the perceived need for intervention, as many commenters voiced that child growth should not be interfered with as long as the child was healthy, If my children are healthy and happy who cares just let parents be parents and children be children. Interestingly, the notion of health was often intertwined with the notion of happiness, indicating that commenters considered both physical and mental health. This shows that childhood obesity is only one of various factors commenters considered relevant to child health. Commenters who focused on more immediate indicators of child health, such as mental health or hunger, did not appraise childhood obesity as a major concern.
Dismantling the ticking timebomb. Comments on societally framed articles debated whether childhood obesity was a societal problem. On one hand, comments asserted that childhood obesity was a ticking timebomb. This was justified by comments about the dangerous health implications and the high rates of childhood obesity. One commenter noted In the 80s when I was at school u rarely saw any obese children, just the odd one but now it seems normal showing that individuals assessed the scope of the childhood obesity by the visibility of larger bodies, and compared current perceptions to perceptions of the past. In contrast, other comments stated that rising childhood obesity rates were due to the methods of measuring and defining obesity, Even though the main reasons behind any rise have been statistical tomfoolery including using BMI on kids who haven't finished growing. This was supported by other comments stating that children would outgrow their obesity as it was believed to be puppy fat. Further comments voiced that childhood obesity rates were falling rendering it no real problem. Additionally, commenters criticized the societal pressure on children to be a certain size, The problem in today's society is that they want all children to have the same body shape, same weight and height as each other for the same age group.

Othering or welcoming -Discussing size acceptance.
Discussions in response to behaviorally framed articles reflected a dichotomy between problematizing weight and othering and viewing weight as a varying characteristic and including children with obesity. On one side of the dichotomy, commenters viewed that children with obesity do not have an acceptable body weight Children should not be fat. The notion of normative weight was often supported by discussing health impacts of larger bodies and being able to visually assess whether an individual's weight was normative. One comment expressed Go down any high street in the U.K and you see fat children, you see their parents, they're also fat. There was a strong sense of individual and parental responsibility to attain this normative weight, with commenters often using the term excuses to refer to challenges regarding health behaviors. This was accompanied by the view that individuals should not be supported by collective resources such as taxpayers' money Why should taxpayers pay for obese people it's their own fault. Children with obesity were also othered by derogatory jokes and labels, such as greedy or lazy. On the other hand, a few comments discussed the inappropriateness of having one normative child weight standard. Commenters expressed that children's bodies naturally differ from each other and thus should not be pressured toward certain body standards, but that individuals of all sizes should be accepted. This was illustrated by one comment Others will always be larger than their peers, so perhaps it's time to stop being terrified about that, (. . .) and be more understanding of size diversity.

Making sense of the news article
This theme was generated across comments on differently framed articles and explores how commenters perceived and made sense of the content of news articles.
Research and policies from the ivory tower. Scientific research was perceived as originating from an ivory tower, as some commenters treated scientific findings as having the power to create knowledge that will be considered true. One commenter noted I think if it's proven by science there's nothing to argue about?. At the same time, commenters voiced that science was too far removed from the real world. Scientific articles were dismissed as nothing new and researchers were perceived as not understanding the reality of parenting Yeah but the inexperienced just out of uni going 'this baby is fat' because she looks at a chart, doesn't know that, cos she doesn't have children.
Likewise, commenters argued that policymakers did not understand their lived reality. Talking about a proposed policy, one commenter voiced They obviously never travel on public transport, nor do they need to feed children on a budget. Therefore, proposed policies were often dismissed as ridiculous or unenforceable surely. Despite also being perceived as originating from an ivory tower, policies were not treated as facts or truth. Instead, commenters viewed ordinary people to have greater authority, with many demanding the application of more common sense when addressing childhood obesity.

Experience as expertise -Assessing news plausibility.
When commenters assessed scientific research or policies covered, they relied on their own knowledge and experience as forms of expertise. Commenters used their understanding to evaluate the underlying logic of a proposed policy. It was dismissed as implausible if commenters did not understand the causal pathways through which a policy would decrease childhood obesity. Similarly, a policy was viewed as implausible if it addressed causes that commenters did not believe to be contributing to childhood obesity. One commenter voiced Banning them from eating 2 or 3 hours a week will address childhood obesity? (. . .) children get the majority of their food (. . .) from their parents. When evaluating scientific findings, many comments adopted an all or nothing approach where they compared scientific findings to their personal experiences, and then classified them as true or false. One commenter noted God knows where they get their data from, my daughter was bottle fed and she's slim now. Nevertheless, some commenters tried to introduce a more nuanced understanding through stressing that a certain factor represented an increased risk for childhood obesity but was not a definite determinant. Others, however, used personal experiences to support their scientific reasoning Correlation does not mean causation. My daughter is an only child and none of us are obese.

Research and policies -Benefitting the health of the nation?.
Firstly, commenters debated the impact that research and policies might have on child health. When strategies involved communicating weight information, the majority of commenters viewed this as harmful to children, with concerns about eating disorders and negative self-evaluations frequently raised. One commenter shared In reality it could lead to perfectly healthy kids having body issues. Other commenters felt that communicating weight information was nevertheless important as this would aid informed decisions regarding children's health.
Secondly, commenters discussed the impact of news articles on parents. Often, articles were perceived as blaming parents, especially mothers Always blaming mothers for something!. This was said to negatively impact maternal mental wellbeing, especially when articles covered a sensitive topic such as infant feeding, This leads these mothers to feel terrible. At the same time, other commenters criticized these reactions as overreacting and that articles were not blaming formula feeding mothers.
Lastly, commenters debated the impact they perceived policies to have on society. Policies were viewed as targeting everyone and thereby not being differentiated enough. Commenters criticized that policy measures were not considering individual children, thereby neglecting interindividual differences, What I don't get is they use statistics and class every child the same. Furthermore, commenters voiced that policies were not benefitting everyone's heath but threating personal freedom, especially of those seemingly unaffected by childhood obesity. One commenter expressed Stop punishing everybody when parents aren't in control of their children.

Discussion
This study explored two research questions, namely 1) how is childhood obesity framed by UK news articles on Facebook? and 2) how do individuals commenting express understanding of childhood obesity in comments on these articles? The findings in response to the first research question showed that news articles drew on different frames and frame components. In line with Nimegeer et al. (2019) articles more frequently mentioned behavioral causes such as diet and parental behavior compared to medical and societal causes. This could be due to the British public predominantly holding individuals responsible for their health (L'Hôte et al., 2018). Articles more frequently mentioned societal solutions, such as restricting promotions on unhealthy foods, compared to medical or behavioral solutions. This partly contrasts with Nimegeer et al. (2019) showing increased reporting of societal, yet a dominance of behavioral, solutions. These findings could indicate that while policymakers increasingly focus on childhood obesity by promoting societal strategies, they still endorse behavior as the dominant cause through, for example, restricting the promotion of certain foods (The Health Foundation, 2020;Ulijaszek & McLennan, 2016).
Findings addressing the second research question showed that commenters made sense of news articles using their personal experiences. Commenters did not only mention anecdotal solutions as found by Gregg et al. (2017), but used personal experiences to assess the plausibility of the article. This indicates that many commenters relied on experiential rather than critical thinking when assessing the article. This could be due to Facebook being a platform for sharing experiences (Franz et al., 2019) or due to previous comments voicing personal experiences (Rodriguez et al., 2016). Rodriguez et al. (2016) found that when individuals read anecdotal evidence in news articles, they were more likely to respond with their own experience and less likely to endorse scientific and critical reasoning. Thus, anecdotal evidence in comments could hinder in-depth understanding of research or policies covered in news articles. Further research is needed to understand the influence of the anecdotal evidence given, for example, whether it can influence causal attributions or policy support (Niederdeppe et al., 2014). This could demonstrate how social media contributes to public understanding of childhood obesity.
This study found that commenters discussed causes and solutions of childhood obesity such as physical activity, diet, parenting and societal influences, in line with previous research (Cain et al., 2017;Gregg et al., 2017). Interestingly, causes and solutions voiced were over and above the dominant frame of the article the comments responded to. This suggests that commenters do not merely adopt the dominant frame of an article (Holton et al., 2014). This could be due individuals having their own pre-conceptions about childhood obesity (Entman, 1993;Scheufele, 1999). Individuals could then be motivated to share these, especially if the ideas in question have not yet been mentioned by previous comments and individuals hold strong opinions about them (Duncan et al., 2020).
Comments additionally appraised childhood obesity. In line with previous research (Cain et al., 2017), many comments endorsed weight stigma and expressed concern about the health risks of childhood obesity. Yet, consistent with Cain et al. (2017), a minority of comments countered weight stigma and discussed other approaches to health such as approaches focusing on health behaviors. This study found that comments on medically framed articles invoked the least stigma, whilst comments on behaviorally framed articles invoked the most stigma. This contrasts with Su et al. (2020). However, that study conducted quantitative analyses of Chinese social media, while this study was qualitative using Western social media. Further research is necessary to examine whether reactions to medically, behaviorally and societally framed articles found in this study can be replicated in experimental studies. Successful replication would indicate that differently framed articles elicit different responses, in line with framing theory (Entman, 1993;Scheufele, 1999). Unsuccessful replication could indicate that different individuals are exposed or attracted to differently framed news articles on social media and thus comments differ (DeVito, 2017;Karnowski et al., 2017).

Strengths, limitations and future research directions
This study used a novel approach to examine public understanding of childhood obesity, as previous research has commonly focused on either news framing or online comments. Additionally, research on understanding of childhood obesity has been scarce compared to research on obesity (Nimegeer et al., 2019). A further strength of this study is the use of preexisting articles and comments as this allowed framing and understanding to be gathered as it naturally and spontaneously occurred (Kovacs et al., 2018;Sugiura et al., 2017). This means that individuals were more likely to express themselves openly (Jamison et al., 2018;Rodham & Gavin, 2006). By analyzing comments on social media, this study might additionally include viewpoints of individuals hard to reach, such as individuals with lower incomes (Clarke & Braun, 2013;Kalogeropoulos et al., 2017).
The findings of this study are limited by using Facebook as the research tool. While comments can provide rich data, Facebook was not set up for research (Franz et al., 2019). Therefore, the algorithms that determine the order of news articles or comments are often outside of the researcher's knowledge or influence (DeVito, 2017), but influence the data collected. Furthermore, demographic information of the sample was not available limiting the degree to which the findings can be located within their socio-cultural context (Malterud, 2001;Yardley, 2017). The findings are limited to individuals that chose to comment, thus, understanding of individuals choosing not to comment is not represented.
Additionally, it is unclear whether commenters read the news article or just responded to the Facebook news post (including the headline and a picture). Previous research suggests that for individuals to feel informed and thus rely on Facebook news, exposure to news posts rather than reading the news articles is relevant (Müller et al., 2016;Schäfer, 2020). By frame analyzing the whole article, this study might have underestimated the role of headline framing.
Further research is necessary to develop a deeper understanding of how public understanding of childhood obesity is created by news on social media. Firstly, using social media, future studies should examine how commenters interact with each other to create understanding of these news articles (Franz et al., 2019). Secondly, it would be useful to study how the framing of childhood obesity influences public understanding and reaction to social media news stories. The present study indicates a differential influence of individual, societal and behavioral frames, however, experimental methodologies are needed to fully address this question. Moreover, future research should triangulate the findings from this study with findings obtained by other qualitative methods, such as interviews or qualitative surveys (Mays & Pope, 2000). This would allow for demographic information to be collected, and identify whether findings were specific to social media or childhood obesity.

Implications for childhood obesity communication
This study led to two key findings. The first key finding was that comments on differently framed news articles differed in their endorsement of weight stigma. Although causality cannot be inferred, this study suggests that framing childhood obesity as originating from behavior could be contributing toward weight stigma. It is important to avoid blaming individuals for childhood obesity and avoid framing it as a simple lifestyle choice (BPS Obesity Task and Finish Group, 2019;L'Hôte et al., 2018). As weight stigma can be detrimental to child health (BPS Obesity Task and Finish Group, 2019;Pont et al., 2017), this calls for urgent action. Therefore, communication should explain how psychological, biological, societal and environmental factors influence childhood obesity. Moreover, step-by-step explanations should be used to demonstrate interactions of these factors.
The second key finding was that many commenters made sense of news articles using anecdotal evidence. As this has the potential to hinder understanding of the content covered (Rodriguez et al., 2016), joint action by journalists, research and policymakers is needed to improve reporting on childhood obesity.
There is a need for journalists to improve the clarity of their reporting, especially regarding the inferences that can be made from scientific studies. Previous research suggests that news articles use language that indicates a stronger causal relationship than is warranted by the original scientific study (Bratton et al., 2019;Haber et al., 2018). Therefore, journalists should use appropriate causal language, especially in headlines, that is understandable to readers, for example, using wording recommendations suggested by Adams et al. (2017). Given the difficulties of changing the way news is reported (Holland et al., 2013;Schwitzer et al., 2005) and that journalists often base articles on press-releases (Schwitzer et al., 2005), researchers and policymakers should create a climate for improved reporting and understanding of research and policies regarding childhood obesity.
Previous research has found that press-releases issued by universities used language indicating stronger causality than was warranted by the original study (Bratton et al., 2019;Haber et al., 2018), and that overstating in press-releases was associated with overstating in news articles (Bratton et al., 2019). Therefore, to improve the communication of research findings to the public, researchers should issue accurate press-releases, add lay summaries to scientific articles and/or use social media to disseminate findings (Jucan & Jucan, 2014;Kuehne & Olden, 2015). This could improve access to and understanding of research and help transform the perception of scientists in their ivory towers.
Policymakers should be more explicit about how policies address childhood obesity, as some commenters struggled to understand causal pathways. Furthermore, they should explain how interindividual differences between children, families and communities are considered in policies, a concern found in this study. Thus, policymakers should utilize psychological insights to communicate policies in a way that maximizes support and minimizes reactance (BPS Obesity Task and Finish Group, 2019).

Conclusions
This study explored public understanding of childhood obesity by examining news articles and responding comments on Facebook. Findings showed that commenters used anecdotal evidence to make sense of news articles and comments on differently framed articles differed in their appraisal of childhood obesity. This indicates that communication about childhood obesity should be improved, with joint action from journalists, researchers and policymakers.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.