Fujiwara Takayoshi
1 Humanities and Language Division, 56924 Mahidol University International College , Nakhonpathom, Thailand.
Percept Mot Skills. 2018 Jun;125(3):626-644. doi: 10.1177/0031512518760881. Epub 2018 Feb 23.
The theoretical five-factor model for the 35-item Beliefs About Language Learning Inventory (BALLI) has rarely been examined empirically, despite its wide use to investigate language learning beliefs. This study examined the BALLI's dimensional structure with a sample of undergraduate students at a Thai university ( N = 537; male 43.4%, female 56.4%; age mean = 19.12 years), taking an elementary-level Japanese language course. Confirmatory factor analysis by M plus did not support Horwitz's theoretical model. While a five-factor model was identified from exploratory factor analysis, the combined items within each dimension were different from Horwitz's model. Thus, the original five-factor BALLI model was only partially supported. We were able to affirm the core aspects of language learning beliefs composed of eight BALLI items, and we offer a roadmap to develop the model.
尽管35个项目的语言学习信念量表（BALLI）的理论五因素模型在调查语言学习信念方面被广泛使用，但很少有实证研究对其进行检验。本研究以泰国一所大学的本科生为样本（N = 537；男性43.4%，女性56.4%；平均年龄 = 19.12岁），这些学生正在修读初级日语课程，研究了BALLI的维度结构。Mplus软件进行的验证性因素分析不支持霍维茨的理论模型。虽然探索性因素分析确定了一个五因素模型，但每个维度内的组合项目与霍维茨的模型不同。因此，最初的五因素BALLI模型仅得到部分支持。我们能够确认由八个BALLI项目组成的语言学习信念的核心方面，并提供了一个开发该模型的路线图。
