Cognitive Load Model for Child-Computer Interaction: Aggregating Child and Design Factors Shuli Gilutz, John B. Black [email protected]

Department of Human Development, Teacher’s College, Columbia University, New York, NY, USA 10027

Preschoolers’ comprehension of novel computer interfaces was studied with the goal of identifying both child and design factors affecting interface comprehension. 117 children ages 3-5 were presented with four interfaces differing in levels of complexity and familiarity. Parental questionnaires assessed children’s previous technology experience. Comprehension of the interface was measured by children’s ability to recognize the actions needed to play the game. Findings revealed a significant three-way interaction between age, technology-experience, and complexity; with age and experience moderating the effect of complexity on children’s comprehension and creating three significant interface-comprehension groups. Familiarity had an overall positive main effect. These findings suggest that both child factors (age, technology experience) and design factors (complexity, familiarity) play a significant role in children’s human-computer interaction and should be taken into consideration when designing interfaces for children, assigning software to children, and conducting research in this field. The cognitive load theory model for child-computer interaction is presented, defining the elements contributing to the mental load of the interface and the mental capacity limits of the child during the process of comprehension of an interface.

Each participant and a researcher sat together at a child-height desk in a separate room from the main classroom. Sessions lasted 25 minutes on average. The researcher presented the participant with a randomly assigned interface (within age-groups), and asked: “Can you show me how to play this game?”. Each participant played the game on their own for up to 5 minutes, and were then prompted to look for advanced features, for example: “Can you find another drawing to color?”.

The data was analyzed using a 2 x 2 x 2 x 3 factorial ANOVA, with comprehension as the outcome. The model was found statistically significant F(23,116) = 3.27, p < 0.01, r2 = 0.447. All four independent variables were found significant as well: age F(1,116) = 14.58, p < 0.01, r2 = 0.136, technology experience F(2,116) = 4.26, p = 0.017, r2 = 0.084, complexity F(1,116) = 6.75, p = 0.011, r2 = 0.068, and familiarity F(1,116) = 8.46, p < 0.01, r2 = 0.083. Additionally, one 3-way interaction was found significant: age, technologyexperience, and complexity F(2,116) = 4.56, p = 0.013, r2 = 0.089

b. familiar complex

d .unfamiliar complex

Distractions

Intrinsic Extraneous

Mental load (design) Cognitive Load

Motivation

Comprehension

Germane Mental effort capacity (child)

Age

Domain specific

Domain general

Metaphor familiarity

Experience

• This model is based on Cognitive Load Theory (CLT) and assumes that the load on working memory will have a significant impact on comprehension in general, and more specifically, the child’s comprehension of a novel interface. This theory assumes a limited capacity to working memory. If the mental load exceeds the amount that working memory can handle (i.e., a child’s mental capacity limit), comprehension will be hindered [13,14]. • The CLT for Child-Computer Interaction (CCI) model contains three types of elements: child-based elements (age, domain specific experience, domain general experience), design-based elements (navigation, distractions, motivation/content), and the outcome: comprehension. In order to explain the relationships between these elements, an additional layer was added that groups the elements in terms of cognitive load theory (mental load, metal effort limit, sum cognitive load).

Discussion

Model of three-way interaction of age and complexity by experience (a) experiences moderates the interaction of age and design (b) the complexity levels create a ceiling for each age group, (c) statistically significant different achievement groups: younger/no experience/complex, older/high experience/simple

Methodology • Four novel interfaces were designed a. familiar simple varying in levels of familiarity and :the Paint Pad interface (a., b.) represents a familiar coloring book environment. • The Monster Maker interface (c., d.) is a novel interface, an unfamiliar environment in which participants have no prior knowledge regarding how to c. unfamiliar simple play the game. • Both games had been designed in two versions so that each reflected two levels of complexity: simple and complex. The two versions differ in the number of actions that are available on the screen for exploring and clicking.

Navigation elements

Results

Theoretical Background • According to User Centered Design, interfaces should be designed according to users’ capabilities, needs, and expectations, using usability testing to undercover problems and an iterative design cycle [12]. • In the same manner, children’s unique characteristics play an important role in creating a successful user experience for them [1, 8, 9]. However, child factors such as age and technology-experience have not been incorporated consistently in children’s interface research, so we do not know how they interact with each other, or with other factors such as the design elements themselves [7,10]. • The variable of age, therefore, encompasses in it many critical developmental differences in children’s ability to interact with technology. Studies have shown significant differences in the way children interact with interfaces even with one year age difference [5,6] • Research with adults has shown significant differences between experts’ and novices’ learning [2,3,11]. Similarly, there have been studies that show that children’s previous experience with technology significantly affected their learning of new interfaces [4]

Cognitive Load Model for Child-Computer Interaction Complexity

Abstract

The results of the study can be aggregated into three key findings: design-, age-, and technology experience-related effects. Each of these may be explained using the construct of cognitive load. •Age effects: Younger children with no technology experience are the most vulnerable to complexity and familiarity issues. Also, older children with high technology experience present ‘expert’ behavior in the simple condition. •Design effects: The complexity level of the interface defined a limit to the success of each age group, and to the extent of experience’s ability to help them reach that limit. •Technology experience effects: When cognitive load is high, technology experience has more impact than age or design. Experience moderates the interaction of age and design. Another way to summarize all three effects on comprehension is to say that age sets the minimum, design sets the maximum, and technology experience sets the distribution between them.

Similarly to the other two cognitive models in this field, this model can be used to help in the research and design process of children’s technology. When considering different interfaces for a specific mental capacity limit, designers can analyze their level of mental load by teasing apart their intrinsic and extraneous loads and concluding which interface would have the largest cognitive load for the user. This can be the basis for iterative testing with users of the same age and technology experience, to fine-tune the appropriate load needed for ease of comprehension. Similarly, educators can use this model for assessing which of their students would best benefit from it. Since the mental load is fixed, the interface can be tested with a small group which has the same mental effort capacity to fine-tune ease of comprehension. Additionally, younger users or those with less exposure to technology (perhaps due to the digital divide) could be assigned to a few training sessions that will give them enough advantage to learn at ease like their peers.

References 1. Brouwer-Janse, M. D., Suri, J. F., Yawitz, M., Vries, G. d., Fozard, J. L., & Coleman, R. (1997). User interfaces for young and old. interactions, 4(2), 34-46. 2. Chase, W., & Simon, H. (1973). Perception in chess. Cognitive psychology, 4(1), 55-81. 3. Chi, M., & Glaser, R. (1985). Problem-Solving Ability. In R. Sternberg (Ed.), Human Abilities: An iformation processing approach (pp. 227-250). San Francisco: Freeman. 4. Donker, A., & Reitsma, P. (2004). Usability testing with young children. Paper presented at the Proceedings of the 2004 conference on Interaction design and children: building a community. 5. Druin, A. (2002). Age matters. ACM SIGCHI Bulletin - a supplement to Interactions, 2002, 5-5. 6. Egloff, T. H. (2004). Edutainment: a case study of interactive cd-rom playsets. Comput. Entertain., 2(1), 13-13. 7. Glaubke, C. R. (2007). The effects of interactive media on preschoolers’ learning. Oakland, CA. 8. Hanna, L. (2007). Designing Electronic Media for Children. In R. Lueder & V. Rice (Eds.), Ergonomics for children: designing products and places for toddlers to teens. New York, NY: Taylor & Francis. 9. Haugland, S. (1992). The effect of computer software on preschool children's developmental gains. Journal of Computing in Childhood Education, 3(1), 15-30. 10. Hourcade, J. P. (2008). Interaction Design and Children. Found. Trends Hum.-Comput. Interact., 1(4), 277-392. 11. Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23-31. 12. Norman, D. A., & Draper, S. W. (1986). User centered system design: Erlbaum Hillsdale, NJ. 13. Sweller, J. (1994). Cognitive load theory, learning difficulty, and instructional design. Learning and instruction, 4(4), 295-312. 14. Sweller, J. (2006). How the human cognitive system deals with complexity Handling Complexity in Learning Environments: Theory and Research (pp. 13).