A Persuasive Game to Encourage Healthy Dietary
Behaviors of Kindergarten Children
Tung-Yun Lin, Keng-hao Chang, Shih-yen Liu, Hao-hua Chu
National Taiwan University
#1 Roosevelt Rd. Sec. 4, Taipei, Taiwan 106
{r94944009, r93018}@csie.ntu.edu.tw, b90701219@ntu.edu.tw, hchu@csie.ntu.edu.tw
ABSTRACT
We have explored the design of an interactive, persuasive game to assist adults to improve dietary behaviors of kin-dergarten children. The persuasive game is played over a smart lunch tray. The smart lunch tray incorporates both context-awareness and interactive media persuasion, ena-bling the creation of a persuasive smart object.
INTRODUCTION
Studies have shown that our dietary habits are developed during the first few years of our childhood experiences with foods. Our preferences for specific foods come about mainly through the following three factors: (1) the sensory flavor of the food, (2) the post-ingestional effects of the food, and (3) the frequency of the food eaten. Among them, the frequency of the food eaten can be controlled by proper parenting. That is, a child may obtain a dislike of a certain food because he/she has never tried to taste the food or tried frequently enough to become accustomed to its taste, shape, color, texture, etc. Over time, this dislike grows into a per-manent dietary habit. Proper and smart parenting can help reverse such specific food aversions. However, most paren-tal experiences can reflect that verbal persuasions can be ineffective and overly time consuming for many young children.
To address this problem, we have designed and imple-mented an interactive game to assist parents (or educators) in encouraging healthy dietary behaviors of their young children (3-6Y). The interactive game is played over a sen-sor-augmented lunch tray that can detect a child’s dietary behaviors. In addition, a small LCD display, embedded as a tile on the lunch tray, is used to show the state of the inter-active game on the lunch tray. Our design combines the
context-awareness and the interactive media technique in persuasive computing into a persuasive smart object. There are been several work on applying computing to hu-man persuasion and behavior modification. Fogg [2] listed basic principles for designing persuasive technology. These principles are adapted in our design described later. Out [3] designed a high-tech doll that looks like a human baby to simulate how hard it is to care for a baby.
DESIGN CONCIDERATIONS
Since our target users are young children, we have identi-fied the following three design considerations. The first design consideration is that most young children are not capable of operating digital devices. Therefore, our game
design relies on using the child’s natural eating actions as inputs. The second design consideration is to minimize the change on the lunch-ware (objects) accustomed to young children during their normal eating routines. This prevents our installed sensors from adversely affecting the young children’s normal eating. The third design consideration takes into account the limited cogitative level of young children. The design of our interactive game must be simple enough for them to understand and attractive enough to draw and maintain their attention. More importantly, the game must have a persuasive ingredient in encouraging the desirable behavior - eating and finishing all the food items placed on the lunch tray. The game design adapts a com-mon strategy used in behavior modifications: applying a positive reinforcer to the desired behavior - if the young child finishes all the food items, the cartoon characters will be fully colored and look gorgeous; on the other hand, if the young child avoids eating a specific food item, the color corresponding to that food item will be missing and the cartoon character will look incomplete and unattractive.
DESIGN AND IMPLEMENTATION
The design of our system is shown in Figure 1. A child eats on an awareness-enhanced lunch tray with a media-feedback display. The smart lunch tray is made up of an array of weighing sensors embedded underneath the bottom of the tray, shown in Figure 2, which recognizes tabletop dietary behaviors of a child. The recognized behaviors are then fed as inputs to the interactive game to provide persua-sive feedbacks.
Smart Lunch Tray
Our current smart lunch tray prototype has a dining surface of 30x45 cm2, which is about the size of a small lunch tray for children. The tray is divided into a matrix of 2x3 cells, each with the size of 15x15 cm2. Except the middle top cell which is a small LCD display, each of the rest five cells
Figure 1. A child eats on an awareness-enhanced smart lunch tray with a media-feedbacks monitor.
contains a weighing sensor, and the weight sensor detects how much weight a child eats from the food container placed on the cell. To simplify our design, each food con-tainer is fixed on one and only one cell. This is done by simply gluing the food container to a corresponding cell. The mapping between the food content and each container is entered manually into our system by the child’s parents. We also assume that a child can use a fork/spoon or hands to pick up food from a food container and puts into mouth directly.
A child’s eating activity is recognized by the Weight
Change Detector shown in Figure 2. It performs two tasks: (1) it aggregates weight samples collected from each of the 6 weighing sensors; and (2) it reports Weight-Change events of food containers by filtering out noises in the stream of weight samples. These weight change events are interpreted into the amount (weight) of food consumed by a child from these food cups on the tray. These weight change events are then fed as inputs to the Persuasive Game com-ponent described next.
Persuasive Game
To persuade children to eat all food items on the lunch tray, we take the strategy of positive, just-in-time feedbacks. The positive feedback isn't just about giving praises, but it also means rewards and incentives for trying to eat food items which they dislike. Experts say that the positive feed-back is the most powerful tool parents have to improve their children's behaviors [1]. By sensing a desirable behavior and then providing just-in-time feedback, a computer sys-tem can help a child easily learn the causal relationship be-tween a desirable behavior, e.g. finishing all food items on the lunch tray, and a positive feedback. In addition, as the targets of our work are children, we have added the ingredi-ent of fun in the persuasion, by making it an interactive game.
Based on the idea of positive feedbacks, we have designed a game to let a child play while they are eating as shown in
Figure 2. The game is to color the picture of a child’s
fa-vorite cartoon character. The positive feedback of our per-suasive game is turning their favorite character into a pretty and colorful one. Specifically, each food item corresponds to a particular crayon color, and the color will be drawn on the character when the corresponding food item is eaten. To make his/her favorite cartoon character colorful, a child is then motivated to eat and finish all food items on the table, including food items that he/she dislikes.
EVALUATION
We tested our prototype implementation on a child (Alicia) who was 3 years and 8 months old. We placed 5 cups of fruits shown in Figure 2, containing small bites of apples, bananas, papayas, wax apples (tropical fruit), and dragon fruits (tropical fruit). Alicia’s parents told us that (1) she did not like dragon fruit and banana, (2) she enjoyed coloring cartoon character very much, and (3) her favorite cartoon character was a tiger-like character shown in the upper
middle block of Figure 2. The result showed that Alicia comprehended the game and was actively trying different fruits to color her favorite character.
Based on the previous result, we extended our study to 10 NTU kindergarten students of age 4 to 5, who have been identified by their teachers as having unbalanced dietary habits. We first interviewed these 10 children to understand (1) why they disliked eating certain food, and (2) under what situations they would try the disliked food. After the interview, we let the children play our game. We observed and recorded their behavior during their play, and then asked questions on what they think about the game. Our first finding is that the game is easy and simple to un-derstand for the young children – one bite of food leads to some coloring of the cartoon character. The children found two fun aspects of the game. The first fun aspect is the novel linkage between eating food and drawing cartoon character – they have never thought that it is possible to mix the playing and eating activities together. The second fun aspect is to explore and learn the mappings between the food items and the drawn colors. This exploratory aspect creates incentive for children to try food items, which they may previously disliked, just to see what color appears where on the cartoon character. Our second finding is the important social aspect of game play. While some children were playing with the game, some of their friends were spectators watching from the side. The enthusiasm from some spectators was noticeably passed into the at-play chil-dren who initially seemed half-hearted but were later changed to become absorbed into the game.
FUTURE WORK
We will apply the social factor among young children (ac-cording to our user study) and explore more factors that may influence children’s dietary behavior. We will compare the effectiveness of a persuasive game with traditional ver-bal persuasion. Furthermore, we will incorporate persuasive games to shape other behaviors, such as brushing teeth, cleaning a room, etc.
REFERENCES
1. Sal Severe, How to Behave So Your Children Will, Too!,
Penguin USA, August 2003.
2. B.J. Fogg, Persuasive Technology: Using Computers to Change What We Think and Do, Morgan Kaufmann, Decem-ber 1, 2002, ISBN 1558606432.
3. Jennifer W. Out, Baby Think It Over: Using role-play to pre-vent teen pregnancy, Adolescence, 36(143): 571-582 (2001)
