A Study on the Cognitive Potential of Pre-school Children with AR Collaborative TUI

Abstract

The most important factor in pre-school children's psychological perception is ease of learning, and the closest measure is "natural" interaction. This study aims to explore the potential of tangible user interfaces (TUI) for AR collaboration for children's cognitive development. The conceptual model is constructed by analyzing physical interaction, spatial perception and social collaboration on the usability of TUI, to explore the role of TUI in pre-school children's cognition. In the empirical study, children aged 3-6 were taken as research objects. The experimental tool is "Plugo" education application. Parents answered questionnaires after observing their children's use. Research shows that physical interaction are the most critical factor in TUI. TUI is beneficial to the cultivation of spatial ability. The results are as follows: 1. Cronbach's Alpha and KMO were 0.921 and 0.965, which were significant and passed the reliability and validity test. 2. Through confirmatory factor analysis (model fit index, combinatorial validity), we found that physical interaction were closely related to usability. 3. The path analysis of the relationship proves that usability has a significant impact on the cultivation of pre-school children's spatial ability.

1. INTRODUCTION

1.1 Background

The challenges of the globalization and knowledge-based economy in the 21st century make it imperative to develop STEM (Science, Technology, Engineering, Mathematics) skills. STEM came from the United States and was fully introduced in 2006. It aims to develop problem-solving skills and habits of thinking, such as curiosity, creativity, perseverance and collaboration. So over the past 15 years, researchers, educators, and parents have focused on early STEM development.

Augmented reality (AR) applications are conducive to the cultivation of children's STEM abilities, such as AR cards and AR picture books, which bring many benefits of three-dimensional , fun and attracting children's curiosity to explore knowledge. But the captured information is always displayed on the screen, not the actual object. Therefore, although parents accept the advantages brought by AR technology, their children are deprived of the tactile and physical environment, and they worry that it will limit their imagination and harm their physical and mental development by using screens for a long time (Fig. 1).

Fig. 1. Pre-school children use AR technology. (a) AR cards and (b) .

Montessori education promotes active learning by using contact between body and object for pre-school children aged 2-7. Jerome Bruner's Theory of Development also believes that when children learn knowledge, which the process from the concrete to the abstract is: firstly it comes hands-on “Action”, then learning with “Images” and finally children transform what they’ve learned into “Language” [1]. It seems to confirm parents' concerns that it is not enough for pre-school children to use virtual educational products, but to build knowledge with real materials and environments. The way physical objects touch each other and combine into different shapes, including materials and colors, can be used to express abstract relationships such as spatial, mathematical, and computational cognition, which provide an important foundation for pre-school children's later STEM education. Therefore, this paper studies the cognitive potential of pre-school children based on the Tangible User interface (TUI) with physical interaction.

1.2 Research Purpose and method

This study aims to meet the needs of parents for the real environment and balance the usage between digital and physical world, and cultivate the thinking ability of STEM for pre-school children in their early years. Based on distributed cognition theory and collaborative learning theory, this paper studies the cognitive potential of TUI on preschool children, according to the implicit interaction characteristics formed by physical interaction and real-time tracking provided by AR[2].

The research methods of this paper mainly using literature collection method and empirical research method. Firstly, by analyzing the system of AR collaborative TUI and the implicit features of physical interaction, and extract the elements of TUI. Then, the distributed cognition theory and computer-supported collaborative learning theory are discussed to analyze the role of TUI in preschool children's cognition, which form the supporting foundation and establish the structural equation model. Finally, the method of experiment and questionnaire survey is adopted to analyze the data and test the hypotheses. According to the results, it got the potential factors of TUI and their relationship on pre-school children's cognitive ability (Fig. 2).

Fig. 2. Research process.

2. THEORETICAL STUDY

2.1 AR environment and TUI

AR environment is to create an enhanced workspace where users can access a shared space based on the real world and real-time registration of dis-plays with physical materials [3]. This interaction mode has physical objects, so that children can feel the rich multi-modal sensory properties of the physical world, which is conducive to cognitive learning (Fig. 3).

Fig. 3. AR environmental framework

Inspired by research into AR environments, Ishii's team first proposed the concept of TUI in 1997. Enhance computing to the physical environment to achieve a high degree of readability and seamless information. TUI has three key concepts: interactive surfaces, real-time computing of crawlable objects , and awareness of environmental context[4]. Ishii advocates using grabable physical objects and environmental media as inputs, rather than by considering pure visual enhancement. Four elements were extracted from TUI features for analysis (Table 1).

Table 1. Interaction elements for TUI.

2.2 Physical interaction

TUI make the interface "natural" to use, especially for pre-school children that they can easy to use. The natural transition between ubiquitous computing and the physical interface creates an implicit style of interaction [5]. It has four dimensions: simple physics (gravity, friction, persistence of objects, relative scale); Body awareness (proprioception, range of motion, hand coordination, whole-body interaction); Spatial perception (navigation, depth perception, distance between objects) and social awareness and skills (the presence of others, verbal and non-verbal communication). These four interactive features meet the requirements of hands-on learning, real environment and social collaboration, so this study will evaluate the cognitive potential of TUI for pre-school children based on physical interaction.

2.3 Differences between TUI and GUI

TUI designers try to increase the capacity of working memory by using visual, auditory, and tactile channels to distribute information processing among the three sensory channels. Graphical user interfaces (GUI) require users to interact within screen space (Fig. 4). Manches and Malley proposed the value of tactile, visual, and auditory information processing for long-term interactions with physical operations [6]. For example, when children are facing a computer, abstract actions such as touching the screen and clicking can not be as helpful as TUI touching objects, even if they add vivid and lifelike vision. Marshal demonstrated that children maintained more control over physical objects than digital objects accessed through multi-touch desktops [7]. Jacob compared several interactive techniques for organizing information in space, showing how TUI can be more effective than GUI [8]. Verhaegh says TUI is more effectiv for 5- to 7-year-olds than clicking on a game's graphical interface that involves manipulating objects in space [9].

Fig. 4. TUI versus GUI interaction patterns (a) and (b) [8].

2.4 Distributed cognition theory

Distributed cognition as the theoretical basis of human-computer interaction research has been developed over the past 12 years. It is tailored specifically for understanding the interaction between people and technology, providing a rich framework for designing and evaluating human-computer interaction.

It has three principles. 1. Social organization itself is a form of cognition structure. Children live in different social organizations, such as family, school and society. They transfer information in the cognition process. The social environment is a cognition form of children. 2. It's a process of embodied cognition. Andy Clark believes that taking the body and the world seriously is a successful way of dealing with the complex interaction between the inside and the outside. Pre-school children's cognition thinking is the internal model of understanding the external world. It is inseparable from external resources, such as good physical materials, the body and so on. They are integrated into children's cognition and control activities. 3. Cognition and culture are inseparable. Culture and cognition are based on the process of historical updating. Children are accompanied by the accumulation of culture in the cognition process, which is the development form of pre-school children's cognition.

Therefore, children's cognition not only needs to understand the information processing of the brain, but also needs to deal with how information is arranged in the physical and social world (Fig. 5). Distributed cognition theory plays a special role in understanding children's interactions with technology because it focuses on the whole environment, constantly creating external scaffolding to simplify children's cognition tasks. This is quite different from traditional information processing psychology, which requires symbolic interaction.

Fig. 5. Distributed cognition framework

2.5 Computer supported collaborative learning theory

In Collaborative Learning theory, learning is a natural social behavior that occurs through conversation, trying to solve problems, and seeking to understand the world. Vygotsky's zone of proximal development (Fig. 6), in which children engage in conversations with "more knowledgeable others" (teachers or parents). Over time, through social interaction, it can develop the ability to solve problems and complete certain tasks independently. In TUI systems, collaborative learning with a com-puter can be formulated by designing both physical objects and learning activity instructions. Guided learning helps children show better comprehension.

Fig. 6. Zone of proximal development for children.

3. EMPIRICAL RESEARCH

3.1 Problem Description

Based on the implicit interaction characteristics of physical interaction and AR environment, this study attempts to study the potential of TUI for pre-school children's cognitive development. The study involved pre-school children aged 3 to 6 and the experimental tool is an AR education application named . The research method is experiment and questionnaire survey.

3.2 Research design and procedures

In order to verify the research problem, the following experimental procedures were used to carry out the research（Table 2）: 1. Preliminary experiment. To ensure the effectiveness of the subjects, it was necessary to ensure that all pre-school children used , requiring parents to install gaming software and participate in their children's games. The experiment was conducted from February 20 to 27, 2022. 2. Secondary experiment. A closed questionnaire was conducted for parents who met the age requirements of pre-school children. SQ questionnaire and research topic questionnaire were set up (Table 3). The survey was carried out after the end of the fisrt experiment using 5-level Likert scale. A total of 267 questionnaires were issued and 245 valid questionnaires were recovered. The effective rate of the questionnaire was 91.76%. In order to verify the feasibility of the extracted factors, reliability test and exploratory factor analysis were performed using SPSS26.0. AMOS23.0 software was used to conduct confirmatory factor analysis, model fitting degree and path analysis for structural equation model. The purpose is to explore the factors and relationship between TUI and pre-school children's cognitive development potential.

Table 2. The experimental process.

Table 3. The questionnaire.

3.3 Outcome analysis

3.3.1 Exploratory factoranalysis

In the reliability test (Table 4), Cronbach coefficient method was used to test the survey data in this study, and the results showed that four effective factors were extracted from the questionnaire. Cronbach's Alpha values were respectively 0.859, 0.814, 0.849 and 0.762, which was greater than the standard value of 0.7. The Cronbach's alpha coefficient of the whole scale was 0.921. The closer the data is to 1, the higher the consistency of the questionnaire is and the more reliable it is. Therefore, the reliability of this scale is good. Meanwhile, the KMO value of the sample is 0.965, the Chi-square test value of Bartlett sphericity test is 2126.868, the degree of freedom is 105, and the significance is 0.000. It shows that the correlation between non-identity matrix and statistics is very significant and the validity is good, so it is suitable for factor analysis.

Table 4. Reliability and validity test.

As can be seen from Table 5, the four effective factors extracted from the questionnaire contents are "social collaboration", "Physical interaction", "spatial perception" and "usability", which are consistent with the theoretical perspective.

Table 5. Factors results.

3.3.2 Confirmatory factor analysis

In order to further refine the items in each dimension and improve the accuracy of latent variables in the model analysis, this study conducted a confirmatory factor analysis on the convergence validity of the measurement model. Further evaluation of the structural model can only be performed if the fit of the measurement model reaches an acceptable standard. From the results of factor analysis, it can be seen that the standardized factor loads of all potential variables are above 0.7 (STD in Table 6). The P value is significant. The coefficient of composite reliability (C.R.) of each latent variable is 0.857, 0.801, 0.848 and 0.746. They are all higher than the standard value of 0.7. The AVE (Average Variance Extracted) value was 0.560, 0.516, 0.582 and 0.490. All above or close to the standard value of 0.5. It indicates that each latent variable has good discrimination and inner fit, so it has good validity.

3.3.3 Structural equation model and path analysis

The structural equation model and path coefficient are shown in Table 7. In addition, the main adaptation indexes obtained from structural model inspection are listed in detail. The CMIN/DF coefficient of the model fitting index was 1.263, the RMSEA was 0.033, the GFI was 0.945, the AGFI was 0.923. All meet the standard value above 0.7. It shows that the model fits very well.

Table 6. Reliability and validity analysis of measurement model.

As can be seen from Fig. 7, factor loads of potential variables in social collaboration, physical interaction, spatial perception and usability are all greater than 0.7, indicating that potential variables have a very good explanatory ability to observation variables and high reliability. According to the path coefficients among the potential variables of social collaboration, physical interaction, spatial perception and usability, physical interaction have a significant impact on usability, with the maximum path coefficient of 0.95, indicating that physical interaction are the key factor for parents to approve the usability of TUI. At the same time, physical interaction are also closely related to spatial perception and social collaboration, and the path coefficient values are 0.94 and 091, indicating that physical interaction are a key factor in providing children with learning potential. Other factors that affect usability are social collaboration, with a path coefficient of 0.60, indicating that TUI helps children develop a sense of collaboration and explore knowledge with parents. The coefficient of spatial perception to usability is -0.38, which proves that the usability of TUI is beneficial to the cultivation of children's environmental space perception and ability.

Table 7. Model fit coefficient.

Fig. 7. Structural equation model.

3.4 Results

This study investigates the cognitive potential of TUI based on physical interaction, distributed cognition theory and cooperative learning theory. First of all, based on the theoretical investigation, the characteristics of physical interaction and children's cognitive characteristics are analyzed. Secondly, empirical analysis is carried out through experiment and questionnaire survey. The reliability and validity of the questionnaire met the standard value of 0.7, so it was suitable for exploratory factor analysis. Four effective factors were extracted, named social collaboration, physical interaction, spatial perception and usability. In order to improve the accuracy of each potential variable in the model analysis, confirmatory factor analysis was carried out, and the standardized factor loads of all potential variables were above 0.7, with significant P values. The component reliability of each potential variable (C.R) and the extraction volume of mean variance (AVE) both met the standard value, indicating good validity. CMIN/ DF, RMSEA, GFI, AGFI and so on all meet the standard range, so the model fitting degree is good. Through path analysis of potential variables, it is found that the path coefficient of physical interaction on usability is the largest, and then social collaboration and spatial perception , which proves that physical interaction are the most important factor in TUI cognitive potential. At the same time, social collaboration affects usability, and usability affects spatial perception, suggesting that the cognitive potential of TUI plays an important role in cultivating children's spatial ability. Spatial ability is one of the important STEM skills, so the cognitive potential of TUI is conducive to the development of STEM ability in early pre-school children.

4. CONCLUSION AND LIMITATIONS OF THE STUDY

4.1 Conclusion

Studies have shown that the rich properties of physical interaction play an important role in fostering social collaboration and environmental spatial ability. TUI supported by AR is conducive to the development of children's spatial ability. Specific analysis is as follows:

(1) Physical interaction are the most critical element in TUI.

Tactile, visual and body movements of physical materials can help children perceive the real environment, enhance their sense of collaboration, build knowledge and enhance memory.

(2) TUI is beneficial to pre-school children's spatial ability cultivation.

The awareness of social collaboration and the interaction of physical materials are conducive to pre-school children's understanding of abstract things. Their mental rotation ability also are improved. These are the potential obtained by using TUI. At the same time, all these lay an important foundation for the development of children's spatial ability.

4.2 Research limitations and future research

This study analyzes the potential of TUI and AR collaboration on pre-school children's cognitive development, which is embodied in four aspects: social collaboration ability, spatial ability, usability and physical interaction.

The limitation of this study is that only preschool children aged 3-6 years were selected as subjects, so generalization is limited for all TUI users. It is predicted that further studies of other age groups will be more meaningful results. In addition, based on the dimensions of natural interaction, this study limited the cognitive potential of TUI to four categories for analysis, so the research results may be one-sided. If the cognitive potential of TUI interaction is divided according to the category of cognitive development on the basis of natural interaction research. It can be ensured more objective certainty. TUI's role in spatial ability training will be further studied.