There is a great debate among parents, teachers, politicians, and marketers as to how the media contributes to the development of childhood obesity [1]. Recently, meta-analytic data [2-4] and literature reviews [1,5] summarizing a large corpus of empirical studies have suggested that children’s confrontation with media content, particularly persuasive content—such as traditional TV commercials [6,7], product placements in movies [8], brand presentations in YouTube videos [9], or online advertisements [10,11]—impacts their food preferences and eating behaviors, particularly regarding the consumption of foods high in fat, salt, or sugar.

Media content creates a very narrow food environment for children. Most of the content children are exposed to presents foods high in fat, salt, or sugar [12-17]. This kind of content is not only shown most often, the characters in children’s movies also consume these items most frequently and evaluate them predominantly very positively [15]. The consumption of highly processed products that are high in fat, salt, or sugar but low in nutritional value (ie, minerals or vitamins) is contributing to the development of overweight and obesity and therefore can be categorized as unhealthy. The overrepresentation of these foods in audiovisual media content is indeed worrying [18].

Drawing on the notion of cue reactivity, this paper describes a nonexperimental study linking content analytical data with panel survey data on the food preferences of children. We first review the literature on media effects on children’s food preferences and then explain the goals, methods, and findings of the study.

Regarding children’s media environment, it has been argued that products integrated within editorial content, interrupting the content, can be more influential on product choices than classical advertisements [19]. This is especially true for children because they are still developing their cognitive abilities and might therefore not be aware of attempts to persuade them that are integrated with the entertainment content [20]. A study on the Reactivity of Embedded Food Cues in Advertising Model (REFCAM) [21] found that when such presentations were integrated within the editorial content, it first led to a kind of cue reactivity. This cue reactivity revealed itself with an increased heart rate [22] or a higher likelihood of thinking about the presented product [23]. In the next step, this cue reactivity influenced children’s eating habits [21]. The authors described this pathway as being influenced by individual susceptibility factors, such as children’s BMI or age, and also by “the level of integration” of a product [21]. Thus, how a product is presented might influence children’s reactions.

Although scholars have theorized that the “food-related media diet” is also represented in children’s actual diet, this assumption has never been formally tested to date. The available evidence comes from experimental and survey research focusing on unhealthy [24-28] and healthy foods [28-30]. Whether healthy food presentations have the power to influence healthy food choices is not entirely clear from current empirical evidence [19,31].

It might not be sufficient to only consider what food is presented in media content targeted at children; it will also be necessary to consider how the presented food is shown [32]. Since drawing attention to the food is crucial for arousing appetite, reinforcing appreciation for that food and thus triggering eating behaviors [21,33] is more likely when the depicted food plays a central role in the media content. Therefore, presentation centrality, which is commonly connected to interaction with a product (ie, whether one of the characters on screen handles a product, consumes it, or verbally mentions it) [34], is an important factor that needs to be considered. A study by Charry [30] showed that audiovisual presentations of fruit led to higher intentions of choosing fruit compared to a presentation that was only visual. Similarly, findings by Naderer and colleagues [35] suggest that a character handling or consuming the food elicits higher food consumption for the presented snack compared to the food being only visually presented.

Furthermore, social factors can play a crucial role in influencing obesity in children [36]. Parents heavily shape their children’s food environment and thus play a significant role in both establishing children’s food preferences and gatekeeping the media food environment of their children [37]. The theory of parental mediation states that parents’ behaviors can influence how children access, receive, process, and react to media content and to what extent children adopt behaviors presented in media [38]. Parental mediation encompasses a variety of distinct social behaviors [39] that can be important in the prevention of negative media effects in children [40]. One behavior is called social coviewing, which is the joint viewing of media content by parents and children without necessarily talking about the content [39,41].

Experimental data measuring food choices and eating behaviors shortly after exposure to persuasive content suggest that media depictions do significantly shape unhealthy eating behaviors among children [3,4]. However, as valuable as experimental studies are, they do not allow conclusions about long-term effects, and they typically test specific food depictions, not the food-related media diet as a whole. Cross-sectional data examining the correlation of media consumption and children’s BMI indicate that children with factors such as extensive TV viewing also are more prone to be overweight or obese [42]. Some first longitudinal examinations also speak to this relationship [43,44]. However, these studies do not take into account the actual content that children are exposed to.

With an extensive and externally valid design, this study aimed to link content analytical data on children’s actual audiovisual media content—that is, food depictions in movies and TV series they were exposed to—with their food preferences, measured with survey data. This study used panel data from children and their parents over 6 months. The content analytical data represent the appearance and centrality of unhealthy foods, but also—as another urgent area for research—those of healthy foods. The survey data included parental coviewing measures that may diminish (or enhance) the effects of unhealthy (or healthy) food depictions. As dependent variables, unhealthy, but also healthy, food consumption of children was integrated within the model.

The study was approved by the ethics committee of the University of Vienna (00343) and the principals of the respective schools. Prior to each wave, the children were asked to take home an information sheet, a consent form, and a survey for their parents. All children who returned a consent form that was signed by one of their parents at the first measurement point (T1) were then asked for their oral consent; if they agreed, we interviewed them individually and measured their weight and their height. Data were anonymized.

A moderated regression analysis was performed controlling for healthy or unhealthy food consumption as autoregressive paths. All predictors were entered simultaneously and terms were mean-centered prior to computing interaction terms.

Evidence is abundant for the effects of unhealthy food presentation in children’s media on children’s consumption behaviors [4,19,48]. The aim of this study was to revisit this evidence, not only with respect to the presentation of unhealthy foods, but also healthy ones. For the first time in the existing research, content analytical data from movies and TV series that children were exposed to over a time span of 6 months were combined with panel data from children and their parents. With this data linkage design, this study could test how effects evolved over time.

As the findings reveal, even though unhealthy food exposure was much higher overall, exposure to unhealthy and healthy foods alone was not related to children’s healthy and unhealthy consumption behaviors. The results show that the centrality with which the products were presented greatly mattered for the relationships. When foods were presented centrally, that is, when they were zoomed in on or when the characters interacted with them, consumed them, or verbally mentioned them, exposure to unhealthy food presentation was positively related to unhealthy food consumption. This effect can be explained by the fact that centrality eases the perception of unhealthy foods, thus increasing cue reactivity, leading to a “wanting” of that product [33]. That is, centrality plays a role in reminding children about their food preferences.

Interestingly, this relationship was different for the presentation of healthy foods. For healthy foods, centrally placed healthy food products were negatively related to unhealthy food consumption. One could argue that children are made aware of the importance of healthy foods for their diet and thus their preference for unhealthy foods decreases. Yet to create such an awareness, a central placement of the food is necessary. Of course, more empirical evidence is needed to corroborate that claim.

However, when healthy foods were placed noncentrally, they were positively related to children’s consumption of unhealthy food over time. This finding is in line with a prior study [19] suggesting that healthy food placement can promote unhealthy eating behaviors. Subtle presentation of healthy foods can activate children’s inherent preference for unhealthy foods, for instance, by serving as a cue for appetite [32]. When appetite is cued, then children automatically prefer unhealthy over healthy options. However, when the centrality of healthy foods rises, this automatic process may be impeded; children may be reminded about the importance and necessity of healthy food and consume less unhealthy food over time. Again, the precise underlying mechanisms remain to be studied. Overall, even though unhealthy foods were presented more often, the amount of food exposure was not a key element influencing unhealthy food consumption. However, persuasive strategies (ie, centrality) in connection with unhealthy as well as healthy food presentation are more important in that regard.

Against expectations, foods presented in the media that children were exposed to did not show any relationship to healthy food consumption. Healthy food consumption could hardly be explained empirically. One explanation could be that children have an inherent preference for unhealthy foods [49,50]. As a consequence, unhealthy food consumption may be triggered more easily as compared to healthy food consumption. This is also in line with former research testing the effects of different persuasive strategies concerning healthy food consumption [23,51,52]. It seems that more than just a central placement is needed to positively impact children’s healthy eating habits [32].

Finally, coviewing by parents did not moderate the effects on healthy and unhealthy food consumption; however, it was positively related to unhealthy food consumption overall. This finding is arguably hard to explain, as theory would suggest the opposite. Perhaps coviewing exerts an indirect effect: when coviewing, parents, too, are exposed to unhealthy food presentations (which are clearly dominant in children’s media) [15], and they may also be affected by them. These effects on parents may then, in a second step, facilitate the unhealthy eating habits of their children. In such a scenario, potential effects on children could run via 2 paths: unhealthy food presentation could shape children’s eating behaviors directly (ie, by affecting them), and indirectly (ie, by affecting their parents, who then affect their children). This indirect mechanism is certainly speculative and cannot be properly tested in a panel survey; it would thus necessitate strict experimental designs.

This study relied on self-reported data when assessing the specific movies and TV series that children were exposed to. As always, such self-reported data are prone to perceptual biases, as specific movies and TV series may be more likely to be mentioned for reasons that cannot be measured. However, considering that in this study both children and parents were asked about movies and TV series, and content viewed over 6 months would generally be possible to remember, we are confident that we obtained an account of the content the children were exposed to. Also, when it comes to the dependent variables, self-reported consumption behaviors, which are not the same as actual food choices, were assessed [53]. Related to that, this study can only make correlative conclusions regarding the relationship between mediated exposure to foods and children’s consumption behaviors.

The sample relied on the most frequently was movies and TV series. However, this study was unable to include smartphone content or determine the type of screen on which the content was consumed. For the age group of this study (between 5 and 11 years) and their typical content preferences, this seemed reasonable. In addition, the models controlled for overall media consumption to rule out effects of the frequency of exposure to other content not assessed in this study, but future research should take a 360 degree account of children’s media diet. This may be challenging for practical reasons, but integrating several types of content in one study is important. Relatedly, we did not predict reciprocal relations (ie, food presentation by consumption), because food presentation refers not only to exposure, but also to content, and the content of the movies cannot be predicted by consumption. However, future research should follow up on this.

This study focused on the centrality of presentation. As important as centrality is [33], it is only one presentation factor that may shape food preferences [32]. Also, this study did not explicitly distinguish between several types of centrality, such as interacting with a product, consuming it, or mentioning it. When it comes to coviewing, several dimensions should also be distinguished in future research, such as intentional versus passive coviewing. Moreover, the age range used in this study deserves further scrutiny. Children younger than 5 years are frequently exposed to media content containing food. These children are, arguably, particularly susceptible to the presentation of unhealthy and healthy food and thus deserve more attention. Finally, no data on the specific schools the children were attending were collected since these would not have been in line with the ethical standards regarding anonymization. Therefore, this study was not able to control for schools in the models.

When discussing the roots of children’s unhealthy food consumption and childhood obesity, educators, journalists, and policy makers have been very quick to point their fingers at the media, particularly media that targets a young audience. At first sight, the evidence for the media’s role in obesity and unhealthy consumption is overwhelming. Not only are unhealthy foods presented more frequently, more prominently, and more positively as compared to healthy foods [15,54], experimental [19,24,28,48] and survey research [43,44] also suggests that the presentation of unhealthy foods can significantly shape unhealthy consumption habits among children.

This study demonstrates 3 things. First, the findings of this study suggest a more nuanced picture. Linking panel survey data to content analytical data on the content that children were exposed to showed that unhealthy or healthy presentation alone were not related to healthy or unhealthy consumption behaviors. Such relationships may be observable in forced-exposure experimental studies with strong stimuli and measures briefly after stimulus exposure, but not in a longitudinal study sampling actual content [55].

Second, and related to that, this study shows that the way in which foods are presented matters. Centrally placed unhealthy foods do in fact show a positive relationship to unhealthy eating behaviors; centrally placed healthy foods, however, appear to have the opposite effect. Also, healthy foods placed noncentrally seem to prime unhealthy eating habits, not healthy ones.

Third, by and large, children’s movies and TV series seem to be more likely to be positively related to unhealthy than healthy eating behaviors. Healthy consumption behaviors were completely unrelated to exposure to media content. Also, when centrality was low, even healthy foods in the media seemed to foster unhealthy consumption.

Overall, these findings inform and qualify the debate about the media’s impact on healthy and unhealthy consumption behaviors among children. Most importantly, the call to simply place more healthy foods in children’s media may, according to our findings, not be sufficient to combat unhealthy eating and childhood obesity.