在21世纪的社会，信息与通信技术（ICT）已经成为教育改革和创新的重要推动力，对于提高教育质量和促进教育公平具有重要的意义。ICT已经在学校教学和学生学习环节中成为不可缺失的部分。

本文基于国际学生评估项目（PISA）2022年中国香港、澳门地区数据，构建因果效应估计模型，利用倾向得分匹配与因果森林估计在校ICT的学术及娱乐使用对学生数学成绩的平均处理效应、个体处理效应以及变量重要性排序。此外，本文根据PISA2022的结果报告及因果森林变量重要性排序结果选取部分变量作为分类变量，讨论对于不同性别、移民背景、家庭资本、家庭教育期望及数字能力自我效能感群体的处理效应异质性。

研究主要得到以下结论：（1）对于香港、澳门学生而言，在校ICT的学术使用对学生数学成绩均产生显著的正向影响，在校ICT的娱乐使用对学生数学成绩的影响不显著甚至是负向影响；（2）学生的数字能力自我效能感、家庭社会经济地位、家庭教育期望对因果效应的影响程度较大；（3）在香港学生中，男生、家庭资本匮乏、家庭教育期望低、数字能力自我效能感低、第一代移民等群体利用学校ICT资源开展学习活动对其数学成绩的提升效应显著高于其他群体；（4）在澳门学生中，男生、家庭资本匮乏、家庭教育期望低、数字能力自我效能感高、第一代移民等群体利用学校ICT资源开展学习活动对其数学成绩的提升效应显著高于其他群体。

In the society of the 21st century, information and communication technology (ICT) has become an important driving force for educational reform and innovation, and is of great significance for improving the quality and equity of education.ICT has become an indispensable part of school teaching and student learning.

This paper builds a causal effect estimation model based on the 2022 Program for International Student Assessment (PISA) data from Hong Kong and Macao, China, and uses propensity score matching and causal forest to estimate the average treatment effect, individual treatment effect and importance ranking of variables of academic and recreational use of ICT in schools on students' mathematical literacy. In addition, this paper selects some variables as categorical variables based on the PISA2022 results report and the causal forest variable importance ranking results to discuss the heterogeneity of treatment effect for different gender, immigration background, family capital, family education expectations and digital ability self-efficacy groups. 

The study mainly draws the following conclusions: (1) For students in Hong Kong and Macao,the academic use of ICT in school has a significant positive impact on students’ mathematical literacy, while the recreational use of ICT in school has an insignificant or even negative impact on students’ mathematical literacy; (2) Students Digital ability self-efficacy, ESCS and family education expectations have a greater impact on the causal effect; (3) Among Hong Kong students, lack of family capital, low family education expectations, low digital ability self-efficacy, first-generation immigrants and the male group's use of school ICT resources to carry out learning activities has a significantly higher effect on improving their mathematics scores than other groups; (4) Among Macao students, lack of family capital, low family education expectations, high digital ability self-efficacy, first-generation immigrants and the male group's use of school ICT resources to carry out learning activities has a significantly higher effect on improving their mathematics scores than other groups.