Edited by: Rabindra Ratan, Michigan State University, United States
Reviewed by: Benjamin J. Li, Nanyang Technological University, Singapore; Kaitlin Fitzgerald, University at Buffalo, United States
This article was submitted to Human-Media Interaction, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The incorporation of narratives helps to enhance children’s engagement in active video games (AVGs), thus increasing moderate-to-vigorous physical activity (MVPA). Specific narrative elements, such as the visual representation of the characters’ body shape, have been rarely manipulated to explore their role in modulating children’s narrative engagement (NE) and exercise motivation.
To investigate the effects of character body shape manipulation (overweight/obese, average, or athletic slim) on children’s narrative immersion (NI), NE, wishful identification (WI), as well as their mediating effect on AVG and PA motivation.
Children ages 8–12 years old (
Controlling for social desirability, the analysis indicated that participants with a BMI of greater than the 75th percentile had a significantly higher NI, NE, WI, and PA motivation when video characters were set to the overweight/obese condition, than they did for video characters set to the average or athletic slim conditions. On the other hand, children of equal or less than the 75th percentile exposed to the average character body condition had a greater NE, WI, and PA motivation than overweight/obese or athletic slim conditions. A mediation analysis with structural equation modeling indicated that NE mediated the effect between character body shape and AVG and PA motivation.
Narrative cartoon characters that mirror the target participant’s body shape can increase NE, which in turn mediates AVG and PA motivation. Content producers should identify optimal strategies in character body shape design to encourage children of different weight status to participate in PA with engaging stories to maximize health narratives’ persuasive potentials.
∗The term average in this sense is not in reference to the national average body weight, but rather an average of the body weights represented in conditions A and C.
Entertainment media narratives have great potential in shaping children’s perceptions and behavior (
In this context, the “ideal body shape” of media characters portrayed in media is conveyed to children across different developmental stages and may be a motivational factor in altering their own body shapes (
As a crucial element in narrative persuasion, characters, especially their body shapes, have an important role to motivate or deter children to exercise. For children with overweight or obesity, the goal of a persuasive narrative would be for their weight loss, resulting in a body shape of healthy body weight. How should the characters’ body shape look in the narratives for them? Research so far has not provided a consistent answer to this question. The tailoring theories suggest that creating characters that look similar to the actual self, i.e., overweight/obese for overweight/obese children, may improve motivation to increase PA for better health outcome (
Thus, the question remains: What type of body shape moves children? This study aims to address this question through an experimental study of character body shape manipulation in an animated narrative series among 8–12-year-olds.
This section will review a series of interdisciplinary literature. We will first situate this study in the larger public health context of the importance of reducing child obesity in terms of both adverse physical and psychological effects, a goal that this study shares. Then we will discuss the disappointing results of engaging children in physical activity (PA) and the potential benefits of AVG in increasing PA. In the next section, we will discuss how children do not play with AVGs for a long enough duration to have an impact and the potential positive impact of narrative that captures the child’s attention with interesting plots and extraordinary characters in PA promotion via AVGs as well as reviewing several potential mediators of the narrative effect such as narrative immersion (NI), narrative engagement (NE), and wishful identification (WI). We also explore gender differences in the psychological effects of being overweight and obese. Children may have emotional responses to the body images of the cartoons that lead to dissatisfaction and potentially a decrease in motivation to participate in PA. Last but not least, we will review the visual representation of mediated character’s body shapes and the effect of children’s exposure to these body shapes.
In the United States, approximately 16.6% of children are overweight, and an additional 18.5% of children are considered obese (
Physical activity (PA) helps to prevent childhood obesity and has significant psychological benefits on children with obesity, including improvement in emotional well-being, self-perception, and self-confidence (
An alternative PA intervention to promote exercise comes in the form of exercising using AVGs, also known as exergames, which are video games that require movements that mimic “real-life” exercise (
Despite their seemingly engaging qualities, AVGs are less likely to be played for a long periods of time (
The use of narrative could be a solution to this problem. Narratives, defined as two more events occurring in succession (
Not all narratives elicit the same motivational effects in children. Previous studies found that narratives with interesting plots, including cliffhangers, and extraordinary characters are capable of promoting PA engagement in children aged 8–12 years (
Most of the previous studies on obesity combating narratives have targeted the 8–12-year age group. Without intervention, children with obesity in this age group are highly likely to become obese young adults (
Researchers have found that narratives Immersion, NE, and WI are potential mediators that improve or deteriorate the narrative effect. NI was adapted from the term “narrative transportation” (
Narrative engagement, a relevant concept to narrative transportation/immersion, is based on the notion that NE is a process of constructing mental models of narrative events (
Wishful identification, defined as the desire to become or mimic actions of characters in the narrative (
In other words, seeing characters similar in body shape for children with overweight/obesity may enhance the perceived self-relevance, enabling them to become more immersed in a narrative, resulting in a higher WI with the characters. On the other hand, seeing characters engaging in PA in a body shape that the children wish to resemble, may also enable them to be more engaged with the narrative and having a higher WI. Either way, the increased engagement with narratives and WI has the potential to motivate them to participate in higher PA to combat obesity. More research is needed to determine whether similar or aspirational body shapers are more likely to encourage children to engage in PA.
The visual components of the video narratives added to AVGs, such as the body shape of characters, may affect AVG motivation and amount of time spent engaging in PA. Narratives, especially those with vivid imagery, may affect players’ attention and adherence to health messages, such as encouraging PA (
The body shape of characters may also decrease motivation to engage in PA as well. The body shape is an inextricable aspect of contemporary American media, partially due to the prominence of character-based stories, and can contribute to negative psychological effects (
The body shape, media representation, and subsequent associations of the characters in media differs by gender as well. In cartoons, female characters were four times more likely to be depicted as underweight; overweight female characters were more likely to be presented as unintelligent or unhappy (
The portrayals of the ideal body shape in both genders could lead to internalization, which causes body dissatisfaction and even lead to a decrease in motivation to participate in PA. A previous study demonstrated that people are especially susceptible to internalizing media messages about the idealized body shape, as determined by an internalization score. Those who were presented with a high internalization score were more likely to report higher body dissatisfaction (
For younger girls specifically, the internalization of the appearance ideals conveyed through characters, as represented in conversations with friends and criticism by peers, was significantly related to body dissatisfaction (
The negative effects of internalization are further highlighted through the narrative persuasion theory, which states that narratives can have an impact on beliefs, behaviors, and attitudes (
Therefore, building upon previously research, we plan to create a narrative in which the visual components, specifically the body shape of the characters, resemble the viewer with a goal to increase PA. More specifically, to explore the influence of the variation of narrative character’s body shape on 8–12-year-old children’s exercise motivation (AVG and PA motivation), we manipulated the character’s body shape and examined their NI (
Due to the fact that WI is based on the notion that the character’s body shape matches that of the children, we divided the participants into two groups: children with a BMI greater than the 75th percentile vs. a BMI equal or less than the 75th percentile. Therefore, we could compare the effects of the different character body conditions on children of different weigh classes. The 75th percentile BMI marker is also correlated with onset of certain conditions, such as asthma and heightened allostatic load (
We hypothesized that presenting the children with characters with similar character body shapes would ultimately enhance NI, NE, and WI, all of which in turn will increase PA and AVG play motivation.
To investigate the effect of character body shape on the aforementioned outcomes, we presented children aged 8–12 with a single animated narrative video in which we had varied the body shapes of the main characters to three conditions (overweight/obese, average, and athletic slim). These three conditions were chosen to represent a graded scale of the body shape spectrum presented in media narratives, from the idealized body shape to characters that more closely resemble the overweight/obese population our research team works with. As part of a pilot study, children were presented with sample images of potential characters representing each condition and asked for their opinion. From their answers, we noted that the expression “thin” should be avoided and therefore chose “athletic slim” to avoid a negative connotation for children of the body shape. Furthermore, a team of research assistants were also interviewed about the perception of body shapes of three different conditions and contributed to the current naming convention.
A professional media production company, FableVision, was hired to create a 15-min animation about a dystopian Sci-Fi narrative,
During the animation production, the production team created multiple layers of different animation panels to allow various modifications of different elements. This allowed for varying each element of a character (i.e., body shape independent of any other alterations such as hair color, background settings). With this approach, three different character types based solely on body shape (all other elements were kept equal) were created by a single animator to ensure a consistent visual style: overweight/obese (condition A), average (condition B), and athletic slim (condition C; see
Image stills taken from the narrative trailer for each of the three body shape conditions. Produced by FableVision Studios for Northeastern University. Reproduced with permission from Northeastern University.
We partnered with a charter school and eight different youth centers in the Greater Boston Area. Site coordinators helped to distribute information sheets to the parents of 8–12-year-old children attending afterschool activities. Research assistants (RAs) also placed posters/flyers with a brief overview of the study on stationary bulletin boards in the school and at the centers.
Regardless of the recruitment method, interested parents completed a screening questionnaire either online or in person, including parent contact information and basic demographic and physical information regarding their child. If they completed the questionnaire in person, it was then returned to their afterschool coordinators, who later passed them to the research team. All parents and children also completed consent and assent forms to participate in this project.
All data collection took place between January 8, 2019 and June 17, 2019. Inclusion criteria included the following: children who were between 8 and 12 years, able to speak and understand English, and able to complete the protocol. Exclusion criteria included the following: children who did not speak English, who had an intellectual disability (preventing them from understanding the narrative video), or who had a physical disability (preventing them from participating in AVGs).
After RAs obtained completed consent and assent forms from the site coordinators, they visited the sites to conduct further assessments of the children. The children’s height and weight were measured by the RAs two times: height (to the nearest 0.1 cm) using a ShorrBoard (Weight and Measure, LLC, Olney, MD, United States) and weight (to within 0.1 kg) using a SECA scale (SECA Inc., Chino, CA, United States). A third measurement was taken if there was more than a 0.2 cm (height) or a 0.2 kg (weight) difference between the first two measurements. Afterward, the demographic information previously provided was confirmed. Body mass index was calculated using the mean of height and weight measurements (BMI; kg/m2) and using the CDC growth charts (
To ensure children provided us with their unfiltered opinions, we used as the primary data collection site the media classrooms of their afterschool facility, where they usually gather to watch TV or play video games. Before watching the video, children were directed away from their afterschool activities and invited into the room to sit in chairs pre-arranged in front of a high-definition TV set. They were then told that the study was about their opinions of an animation movie created to encourage them to exercise more. They were told that the RAs would ask them for their opinions in a questionnaire that they would complete and that there was no right or wrong answer to the questions. The RAs told them that they should feel free to express themselves confidentially in the questionnaires. They were also reminded that they could stop their participation at any time. Children were then placed in groups of 4–15, depending on the room size, time of the visit, the afterschool program’s schedule, and the child participant’s availability.
After watching the animation clip, children were handed hard copies of questionnaires consisting of the following: a demographic questionnaire (age, sex, race, and ethnic background) and a psychological survey questionnaire measuring NI (
The NI scale (
Given children’s limited understanding of the Likert scales and the typical field practice in psychology (
Similarly, the 12-item NE scale was also adapted from the original study with adults (
The 3-item wishful identification (WI) scale was adapted without any modification as it was originally developed for children (
The assessment of AVG play motivation (AVGM) is based on a 5-item scale developed for this project. The scale assessed the degree of children’s willingness to play an AVG if the story is included as its plot. Sample questions include “I intended to play this AVG” and “I plan to exercise through this AVG,” with 1 = disagree, 3 = neutral, and 5 = agree (Cronbach’s α = 0.792).
The PA motivation (PAM) employed a 16-item scale adapted from the validated exercise enjoyment scale (
The 9-item SD scale (
As children completed the paper copies of questionnaires, RAs observed their attention to the questionnaire completion and answered questions from them. They also took notes of children who simply answered all questions without reading them (e.g., selecting “1” or “5” for all questions on multiple pages or completing the questions in extremely short period of time (e.g., finishing over 50 questions in less than 3 min), which suggested the lack of validity for these answers. After children have completed the questionnaires, children were given a $25 gift card for their participation, thanked, and brought back to their afterschool activities. The RAs then collected the questionnaires and entered them into the database. All data entries were triple checked to ensure accuracy. The questionnaires from those less attentive participants were identified and later removed for analysis.
To compare the perceptions of conditions A, B, and C by children of different weight status, we dichotomized our participant population using the 75th percentile of BMI, which also happened to have resulted in a median split of the total sample as a cut-off point. The Shapiro–Wilk and Levene’s tests were applied for normality and homogeneity, respectively. All variables presented a normal and/or homogeneous distribution and were analyzed using parametric tests. Characteristics of the sample were compared between the group ≤ 75 and >75 BMI percentile with an independent
Since the Cronbach’s Alpha of NI was relatively low (0.544), we then conducted a factor analysis for NI and after excluding two items (“When I was watching the story, other activity going on in the room around me was on my mind.” and “After the story ended, I found it easy put it out of my mind.”), the new Cronbach’s Alpha was 0.610. Nevertheless, due to the high correlation between NI and NE (
General linear models with two factors (two-way ANOVA) were ran for each outcome. BMI percentile (≤75th/>75th) was added as fixed factor, and Condition as a random factor (conditions A, B, and C, were coded as 1, 2, and 3, respectively). When interactions were found (
Structural equation modeling (SEM) was conducted using Mplus 7 (
The statistical significance level was set at 5% (
Data was originally collected from 96 children. Of them, nine children were identified as improperly answering questions (giving the same rating to each of the questionnaire questions throughout or too fast in answering all questions) and were therefore removed from data analysis, resulting in 87 participants.
The participants’ average age was 9.8 (
Participant demographics and social desirability results (
≤75th BMI percentile ( |
>75th BMI percentile ( |
||||||
Condition A (overweight/obese) ( |
Condition B (average) ( |
Condition C (athletic slim) ( |
Condition A (overweight/obese) ( |
Condition B (average) ( |
Condition C (athletic slim) ( |
||
Age, mean ( |
9.8 (1.3) | 9.8 (1.2) | 0.919 | ||||
Weight (kg) | 33.8 (6.7) | 48.7 (10.3) | <0.001 | ||||
BMI, kg/m2, mean ( |
16.6 (1.5) | 23.0 (3.7) | <0.001 | ||||
BMI percentile, mean ( |
43.5 (22.2) | 91.6 (6.3) | <0.001 | ||||
Sex (boy/girl, n) | 8/3 | 12/9 | 5/7 | 9/4 | 11/7 | 3/9 | 0.831 |
Race ( |
0.170 | ||||||
Asian | 2 | 0 | 0 | 1 | 2 | 0 | |
African American | 1 | 5 | 3 | 5 | 4 | 7 | |
American Indian/Alaska Native | 0 | 1 | 1 | 0 | 0 | 0 | |
Caucasian | 4 | 3 | 0 | 2 | 0 | 0 | |
Hispanic or Latino | 2 | 7 | 4 | 2 | 10 | 3 | |
Mixed | 1 | 2 | 3 | 1 | 1 | 0 | |
Other | 1 | 3 | 1 | 2 | 1 | 2 | |
Social desirability ( |
3.3 (0.8) | 3.4 (0.9) | 0.458 |
A two-way between-subject multivariate analysis of variance indicated a significant Condition effect,
Means of the main outcome variables per group (
Condition A (overweight/obese) | Condition B (average) | Condition C (athletic slim) | |
≤75th BMI percentile | |||
Narrative immersion (NI) | 3.39 (0.60) | 3.42 (0.59) | 3.12 (0.58) |
Narrative engagement (NE) | 3.19 (0.61) | 3.74 (0.50)#, |
3.10 (0.54) |
Wishful identification (WI) | 3.09 (1.03) | 3.86 (1.20)a | 3.11 (1.04) |
AVG play motivation (AVGM) | 3.56 (1.16) | 4.35 (0.82)a | 3.35 (1.36) |
PA motivation (PAM) | 3.84 (0.69) | 4.27 (0.68)# | 3.57 (0.79)b |
Narrative immersion (NI) | 3.74 (0.47)c | 3.15 (0.59) | 3.13 (0.80) |
Narrative engagement (NE) | 3.73 (0.48)*,c | 3.34 (0.68) | 3.12 (0.77) |
Wishful identification (WI) | 4.41 (0.78)*,c | 3.50 (1.02) | 3.33 (1.33) |
AVG play motivation (AVGM) | 4.15 (0.93)d | 3.96 (0.96) | 3.35 (1.33) |
PA motivation | 3.93 (0.76)c | 3.39 (0.90) | 3.15 (0.66) |
In terms of the structural equation modeling analysis,
Structural Equation Model for the relation between character body shape, children’s BMI percentile, AVG motivation and PA motivation, with narrative engagement and wishful identification as the mediating variables. X1: the overweight/obese condition in reference to athletic slim condition. X2: the average condition in reference to athletic slim condition. All paths’ estimates are standardized. For the sake of simplicity, the direct and indirect paths from the IVs to the DVs as well as the correlations between the mediators and DVs are not shown. ∗
Structural Equation Model for the relation between character body shape and AVG motivation and PA motivation, with narrative engagement as the mediating variable. All paths’ estimates are standardized. X1: the overweight/obese condition in reference to athletic slim condition. X2: the average condition in reference to athletic slim condition. For the sake of simplicity, the direct and indirect paths from the IVs to the DVs as well as the correlations between the DVs are not shown. ∗
To check the robustness of the overall model, two-way interactions of SD with the condition variables were tested. These interaction coefficients were not significant and results overall remained similar.
To increase children’s engagement in PA-inducing AVGs, we investigated the effects of manipulating the character body shape in an animated narrative on children’s NI, NE, WI, AVG, and PA motivation.
We have found that the visual presentation of different character body sizes in a narrative impacted children’s NE across different weight groups. Children with a higher than 75th BMI percentile had a significantly greater NI, NE, WI, and PA motivation in the overweight/obese (condition A) character condition than those who were in the average or athletic slim conditions (conditions B and C). On the other hand, children equal to or below the 75th BMI percentile had higher NE, WI, and AVG play motivation when the characters body shape was average (condition B) than in condition A (overweight/obese) or C (athletic slim). Furthermore, the overweight/obese character body shape (condition A) resulted in higher NE and WI among children with higher than 75th BMI percentile than those with 75th or less BMI percentile while condition while the average character body shape (condition B) resulted in higher NE and PA motivation in the 75th or less BMI percentile than those whose BMI percentile are 75th or higher. Last but not least, the NE had a stable and significant mediating effect between the character body manipulation and the AVG play motivation and physical activity motivation.
In the mediation model, the NE repeatedly demonstrates the capability of narratives to translate the power of entertainment to health behavior change. We have consistently demonstrated the importance of the engaging factor, that is, a health intervention, be it in the form of narrative media or not, must be engaging to have the potential to induce health behavior change. On the other hand, this also suggested the psychological crux of successful narrative health intervention for children. A narrative message must be fully understood, grabbing children’s attention, capturing their emotion, and drawing them into the narrative world.
Wishful identification yielded no significant mediating path results. While this could be potentially attributed to the relatively small sample of the study, the result also suggests that children’s engagement with a story may be more relevant to their motivation to exercise than their desire to become like the characters presented in the story. In addition, the original WI scale adopted in this study did not include items asking about the physical characteristics of the characters, incorporation of these items may alter the results.
While WI was not found to have significant effects on children’s PA or AVG play motivation in this study, this concept can still be relevant to a child’s engagement with the story and thus the relationship between WI and NE should be further explored. For example, children showed significantly higher WI with characters that looked like their own body shapes in both weight groups. Since these characters may already appear more similar to the children, the increased desire to be like the shape of those characters may not translate directly to exercise behaviors among the audience, despite the fact these characters themselves are engaged in physical activities in the animated story. On the other hand, when WI was higher, so was NE, which according to the mediation results did mediate AVG and PA motivation. A higher WI may contribute to higher NE as well as the health behavior intentions in an indirect fashion.
From a practical perspective, given the higher NE, NI, and WI presented when characters have body shapes that match the intended audience, when executing potential PA interventions among children population, media creators should be conscious of creating body shapes of various sizes if they are catering to diverse children body shapes. Creating the appropriate visual aspect of the character body shape could not only increase children’s engagement with the story, but also motivate them to participate in the physical activities suggested and potentially enhance their adherence to the intervention. This way, children of all body shapes can have access to an engaging and relatable narrative.
Due to the importance of NE in PA intervention, and potential clinical use, further research is needed on how to best engage children of various backgrounds to ensure the narrative truly serves the entire child population. For example, while we have found that the children in our population were more likely to be engaged in stories with similarly sized characters, in the study design; however, we did not account for gender differences as our presented characters had one male and one female character, both of whom were presented in a similar fashion in each of the body shape conditions. As previously reviewed, the traditional gender representation in media is more likely to have women being represented as underweight and men presented as muscular (
In addition, we have only presented the character’s body shape in an initial segment of a long story, most of the effects observed here are based on the beginning status of the character body shape. Will there be a differential response when the characters’ body shapes start to increase or decrease in accordance with children’s exercise participation over different time intervals (
It is also interesting to observe that the athletic slim body type of cartoon characters did not appeal to children in either weight group with the 75th BMI percentile as the cutoff point. While dividing children’s weight groups in a much more fine-tuned fashion may allow us to identify some children with less BMI percentile to engage with the athletic slim condition more, further examination of the data among the slim children proved otherwise; we could not identify a significant number of children who gave this condition higher ratings. This may have something to do with potentially visual and aesthetic fatigue with respect to children’s long-term exposure to the stereotyped media characters’ body shape presentation (
This research project is not without limitations. The participants only watched 15 min of the narrative and did not watch the full six-season animated series over time. Thus, it is unclear if the results would be altered should the children have been exposed to more episodes and more viewing sessions. In addition, due to space and time constraints, we have only measured participants’ response through survey questionnaires. Children did not actually play the AVG after watching the episodes. Thus, exercise motivation was determined based on self-report instead of objectively measured PA outcomes. More research is needed regarding the actual increase in AVG playtime instead of hypothetical change reported on paper. Our access to children was also constrained by the schedule of different afterschool activities in different classes, resulting in relatively limited sample size, which could contribute the borderline and lack of significant results. Similarly, although our participant base was racially diverse and relatively equally divided between boys and girls, the small sample size may preclude us from in-depth examination of specific racial and sex characteristics and differences.
Despite these limitations, this project found that narrative cartoon characters that mirror the target participant’s body shape can increase NI, WI, and NE. NE in turn mediates the effect of character body shape manipulation on AVG and PA motivation. Such findings strongly suggests the importance of having relatable characters to increase children engagement with narrative health interventions. Future health narrative content producers should identify optimal strategies in character body shape design to encourage children of different weight status to participate in PA with engaging stories to maximize health narratives’ persuasive potentials.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The studies involving human participants were reviewed and approved by Northeastern University IRB. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.
DA contributed to data analysis and wrote the first draft of the manuscript. CS participated in data analysis and results interpretation. AL designed the study, supervised the data collection, participated in data analysis, results interpretation, and supervised the study. All the authors contributed to the manuscript writing, read and approved the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors would like to thank the Health Technology Laboratory’s Active Video Game Project Team for their help and effort with the data collection and data preparation.
The Supplementary Material for this article can be found online at:
Shapiro–Wilk and Levene’s tests results.