Hagger Martin S, Chatzisarantis Nikos L D
Department of Psychology, University of Essex, UK.
Br J Soc Psychol. 2005 Dec;44(Pt 4):513-35. doi: 10.1348/014466604X16219.
The present study examined the validity of an augmented version of Ajzen's (1991) theory of planned behaviour (TPB), adopting second-order latent factors to explain relationships between the differentiated components of the theory constructs. This model was tested on data from two independent samples, for exercise and dieting behaviours, respectively. The affective and instrumental attitude, injunctive and descriptive norms, and perceived controllability and self-efficacy constructs all exhibited discriminant validity in the first instance. First- and second-order confirmatory factor analyses exhibited satisfactory fit with the observed data from both samples. There were no substantial differences in the fit indices across the first- and second-order models, and the second-order models exhibited the most optimal parsimony-corrected fit indices. The higher-order models could therefore not be rejected on the grounds of inferior fit or parsimony. First- and second-order structural equation models accounted for significant variance in intentions and behaviour. These results corroborate Ajzen's (2002a) and Bagozzi, Lee, and van Loo's (2001) premise that recent augmentations of the TPB that differentiate the model components can be subsumed by global, higher-order factors while still making the distinction at the subordinate level.
本研究检验了阿杰恩(1991)计划行为理论(TPB)增强版的有效性,采用二阶潜在因子来解释该理论结构中不同成分之间的关系。该模型分别在来自两个独立样本的运动和节食行为数据上进行了检验。情感态度和工具性态度、指令性规范和描述性规范,以及感知可控性和自我效能感结构在最初都表现出了区分效度。一阶和二阶验证性因子分析与两个样本的观测数据拟合良好。一阶和二阶模型的拟合指数没有实质性差异,二阶模型表现出最优化的简约校正拟合指数。因此,高阶模型不能以拟合不佳或简约性不足为由被拒绝。一阶和二阶结构方程模型解释了意图和行为中的显著方差。这些结果证实了阿杰恩(2002a)以及巴戈齐、李和范洛(2001)的前提,即最近对TPB进行的增强,通过区分模型成分,可以被全局高阶因子所涵盖,同时仍在从属层面进行区分。