Chinese and American Children’s Internet Attitudes: A Multi-group Confirmatory Factor Analysis

• In Event: AI and Youth Learning Across the Globe: Insights from Four Large-Scale Surveys

Abstract

Objective
The internet is a key source of information for children, and their search behaviors are increasingly shaped by AI as platforms like Google integrate AI-driven tools to support users’ information-seeking activities. To understand children’s attitudes and beliefs in such a process, we introduced the Children's Internet Attitudes Scale (CIAS). In this study, we established the internal validity of a model of children's internet attitudes and investigated how internet attitudes vary between Chinese and American children using multi-group confirmatory factor analysis and measurement invariance testing.

Theoretical Framework
The latent themes in this analysis were grounded in a tri-part theoretical framework: (1) beliefs about internet accuracy are shaped by selective trust in digital informants (e.g., Girouard-Hallam & Danovitch, 2022), (2) evaluations of breadth of information are rooted in cognitive process (e.g., Hermes et al., 2018) and educational experience (e.g., Eskela-Haapanen & Kiili, 2019), and (3) comfort with independent internet use is a result of self-efficacy in digital navigation (e.g., Duarte-Torres et al., 2014).
Previous research demonstrated that Chinese and American children's trust in internet-based sources emerges around age 7, when they begin preferring digital informants over human sources for factual information (Danovitch et al., 2016; Wang et al., 2019). Cross-cultural developmental theory informed comparisons between American and Chinese children, acknowledging different parental mediation practices (Rideout & Robb, 2020; The Central Committee of the Communist Youth, 2021) and governmental internet policies that may shape children's digital attitudes.

Methods
The study employed a three-phase validation approach across 856 children aged 6-10 years from the United States and China. After an exploratory factor analysis, a multi-group confirmatory factor analysis (mg-CFA) with 439 participants validated the factor structure across cultural groups. Measurement invariance testing was conducted to ensure validity. Structural equation modeling (SEM) examined age, country, and experience as predictors of internet attitudes.
Participants included Chinese children from a primary school in Wuhan and American children from elementary schools across the continental US and university laboratory settings. The CIAS consisted of 16 initial items exploring children's internet attitudes using a four-point Likert scale (NO!-no-yes-YES!), along with six items about children’s internet experiences.

Results
The final validated CIAS comprised 8 items across three factors: scope, comfort, and accuracy. The predictors of interest (age, country, and experience) all related to children’s attitudes. Country significantly influenced scope beliefs (p=.002) and accuracy perceptions (p=.028), with American children perceiving broader internet scope while Chinese children viewed the internet as more accurate. Age positively predicted comfort (p<.001) but negatively predicted accuracy beliefs (p<.001). Experience with diverse internet activities enhanced both comfort and scope perceptions independently of age effects (ps<.001).

Significance
This project provides a developmentally appropriate, cross-culturally validated measure of children's internet attitudes, addressing a critical gap in digital literacy assessment tools. The results of this survey provided insight into Chinese and American children’s internet attitudes.

Authors

• Lauren Girouard-Hallam, University of Michigan

• Yu Tong, Jianghan University

• Fuxing Wang, Central China Normal University

• Judith H Danovitch, University of Louisville