中文摘要：

双因子模型是～种既确全局因子又有局部因子的模型，近年来有了许多应用。本文讨论了双因子模型和高阶因子模型在数学模型、参数之间的关系，概念上和应用上的差异；概述了双因子模型在信度研究、平衡量表、探索性因子分析和项目反应理论中的应用。作为例子，在Rosenberg自尊量表结构的研究中，通过双因子模型分析了自尊特质效应与项目表述方法效应。

英文摘要：

Empirical psychological, educational and social scientists are often confronted with multifaceted constructs which are com-prised of several related dimensions. The traditional methods to measure the muhifaeeted constructs include the total score approach, theindividual score approach and the higher - order factor analysis. The total score approach creates a composite score （ a total or averagescore of all the items） based on the individual facets; this approach is easily to be understood, but ignores the differences among the di-mensions. The individual score approach separately analyzes each facet of the construct ; this approach examines each dimension＇s indi-vidual effect, but does not take the ~ommonality of the dimensions into account. The higher - order factor analysis simultaneously con-siders the common and individual eftects, but the higher - order factors are mysterious and incomprehensible because they are definednot by tests, but by first - order factors. In summary, all of these approaches cannot solve the bandwidth - fidelity dilemma in multidi-mensional testing. Bi -factor models often provide better analysis for multidimensional constructs. Bi -factor models are potentially applicable when（a） there is a general factor that is presumed to account for the commonality of all items; （b） there are several domain specific factors,each is presumed to account for the unique influence of one specific domain. A bifactor model represents a special multidimensionalstructure, characterizing both the unidimensional and hierarchical structures. Bifactor models are widely applied in various fields, inclu-ding cognitive ability, disposional hope, emotional intelligence, psychological reactance, vocational interests, psychopathology, well -being and so on. In this article, the relationship between bi - factor models and higher - order factor models is discussed from the perspective of theconcepts, mathematical models, parameters, and practical applications. Both bi -factor and higher- order factor mo