The Wald W-test indicated the model would not be
improved by dropping any parameters. The LM test
indicated the model could be improved by adding a
parameter from the variable ‘the aesthetic appearance of
your office’ to the ‘satisfaction with lighting’ factor.
Satisfaction with the appearance of one’s office or workspacemight reasonably be related to the quality of lighting
available. Therefore, it was decided to add the parameter
and run a post hoc model.
This model, labelled CFA-A2, is summarised in Table 3.
Additional statistics were further examined: The frequency
distribution of the standardised residuals revealed that
most residuals (91.81%) fell between .10 and .10, which is
desirable. All estimated factor loadings were significant and
internal consistency values for each subscale were satisfactory
(privacy, a ¼ .89; lighting, a ¼ .82; and ventilation,
a ¼ .82). However, this result was not judged to be enough
of an improvement over CFA-A1 to warrant the additional
complexity of a cross-loading item (a variable contributing
to more than one latent variable). The cross-loading item
would have complicated interpretation of further analyses
using scale scores for the latent variables as dependent
variables.
加成50分，谁答下~~~急需啊~~~谢谢了~
真无语……
参考资料：你的分太少了最多帮你用金山