ERIC Number: EJ1115616
Record Type: Journal
Publication Date: 2016
Causal Indicators Can Help to Interpret Factors
Bentler, Peter M.
Measurement: Interdisciplinary Research and Perspectives, v14 n3 p98-100 2016
The latent factor in a causal indicator model is no more than the latent factor of the factor part of the model. However, if the causal indicator variables are well-understood and help to improve the prediction of individuals' factor scores, they can help to interpret the meaning of the latent factor. Aguirre-Urreta, Rönkkö, and Marakas (2016) have provided an important theoretical discussion as well as empirical demonstration that in so-called causal indicator models, the meaning or interpretation of the latent variable should be based on the variables that define the common factor and not on the variables that predict the common factor. Figure 2 of Aguirre-Urreta et al., a mimic model, illustrates the point well. In this model, exogenous variables x[subscript 1] and x[subscript 2] imperfectly predict a latent factor ? that in turn predicts endogenous variables y[subscript 1] and y[subscript 2], and the issue at hand is whether the meaning of ? is derived from the x or the y variables. They conclude that although the x variables may help to identify the model, the latent factor ? must be interpreted via the endogenous y[subscript 1] and y[subscript 2] variables and not via the exogenous x[subscript 1] and x[subscript 2] variables. In this article, the author argues that such a model can help improve one's understanding of what a latent variable might actually be; that is, it may shed light on the meaning or interpretation of a latent variable.
Descriptors: Causal Models, Factor Analysis, Prediction, Scores, Equations (Mathematics), Data Interpretation
Psychology Press. Available from: Taylor & Francis, Ltd. 325 Chestnut Street Suite 800, Philadelphia, PA 19106. Tel: 800-354-1420; Fax: 215-625-2940; Web site: http://www.tandf.co.uk/journals
Publication Type: Journal Articles; Opinion Papers; Reports - Evaluative
Education Level: N/A
Authoring Institution: N/A