网络游戏付费群体消费行为多维解构——基于PEST、Pearson相关检验、GLM与K-Means、列联交叉模型

Multi-Dimensional Deconstruction of Customer Behavior among Online Game Players—Based on PEST, Pearson Correlation, GLM, K-Means and Contingency Cross Models

摘要: 在当今社会信息技术不断发展的情况下，网络游戏上的不同年龄段群众也逐渐增多。数字文娱产业快速扩张，网络游戏虚拟道具付费成为青年主流消费场景，圈层攀比催生非理性高额氪金问题。为系统拆解年龄、性别、职业、可支配收入多重人口特征下的付费异质性，本文构建“宏观环境–线性关联–回归识别–无监督分层–分类交互”多层实证分析框架。依托线上问卷获取有效样本542份，融合PEST行业环境分析、Pearson相关检验、广义线性模型(GLM)、K-Means聚类、列联卡方五类统计工具开展量化研究。研究结果显示：第一，职业、月可支配收入与游戏总付费仅存在微弱不显著负相关，游戏付费属于偏好驱动的非刚性消费，收入约束效应极弱；第二，职业显著区分付费游戏品类，职场人群更倾向休闲益智类轻度付费内容；第三，付费能力与年龄呈非线性正向关系，30岁以上男性为重度付费核心群体；第四，性别存在稳定消费偏好分野，男性偏向MOBA、射击竞技类游戏，女性更青睐剧情养成、休闲解谜类内容。本文研究丰富了数字消费多元统计建模范式，为游戏厂商精细化分层运营、监管机构开展青少年理性消费引导提供量化实证支撑。

Abstract: With the continuous advancement of information technology and the rapid expansion of the digital cultural and entertainment industry, online games have attracted players of all age groups. The digital entertainment industry is expanding rapidly, and purchasing virtual items in online games has become a mainstream consumption activity among young people. Peer comparison within gaming communities has led to the prevalence of irrational high-value in-game spending. To systematically dissect the heterogeneity of consumption behaviors shaped by demographic factors including age, gender, occupation and monthly disposable income, this paper constructs a multi-layer empirical analytical framework consisting of macro-environment analysis, linear correlation test, regression identification, unsupervised clustering and categorical interaction analysis. Based on 542 valid online questionnaire samples, this paper adopts five statistical methods for quantitative research: PEST analysis, Pearson correlation test, Generalized Linear Model (GLM), K-Means clustering and contingency chi-square test. The empirical results demonstrate four key conclusions: First, occupation and monthly disposable income only have weak and insignificant negative correlations with total in-game expenditure, indicating that game consumption is preference-driven non-essential spending barely restricted by income levels. Second, occupation significantly differentiates players’ preferred game genres for payment; office workers are more inclined to make low-cost purchases on casual puzzle games. Third, there is a non-linear positive correlation between age and consumption capacity, and males over 30 form the core group of high-spending players. Fourth, obvious gender gaps exist in gaming preferences: males favor MOBA and shooting competitive games, while females prefer story-based simulation and casual puzzle games. This study expands the multivariate statistical modeling paradigm for digital consumption, and provides quantitative empirical support for game enterprises to carry out refined segmented operations as well as regulatory authorities to guide adolescents to form rational consumption concepts.