Hi, everyone,
I am using the confirmatory factor analysis (CFA) to test the validity of a revised scale.
Seventy-five students filled the scale 4 times (day1, day2, day3, day4). May I analyze the total 75 * 4 scale data? If not, which one is better? Data in the first time?
Thanks!
Yes, you can use models of longitudinal CFA for that purpose. Your sample size is fairly small. However, having more than one time point may actually help compensate for the relatively small sample. This is because the additional time points provide additional information without necessarily increasing the number of parameters to be estimated dramatically. That is, you may be able to set at least some of the measurement-related parameters (loadings, intercepts, etc.) to be invariant across time without a loss in model fit. This is something that should be tested though by comparing nested models with different levels of longitudinal measurement equivalence/invariance. See, for example
Geiser, C. (2021). Longitudinal structural equation modeling with Mplus: A latent state-trait perspective. New York: Guilford Press.
Little, T. D. (2013). Longitudinal structural equation modeling. New York: Guilford Press.
Chapter Latent Variable Modeling of Person-Situation Data
Chapter Longitudinal structural equation modeling of personality data
Chapter Latent State-Trait Models
It doesn't seem possible, but if needed maybe you can try.
A little explanation about CFA:
Confirmatory factor analysis is a factor analysis technique where a priori based on theories and concepts that are known to be understood or determined in advance, several factors will be formed, as well as what variables are included in each of the formed factors and the purpose is definitely. The establishment of confirmatory factors (CFA) is deliberately based on theories and concepts, to obtain new variables or factors that represent several items or sub-variables, which are observed variables or observer variables.
The objectives of confirmatory factor analysis are: first, to identify the relationship between variables by conducting a correlation test. The second objective is to test the validity and reliability of the instrument. In testing the validity and reliability of the instrument or questionnaire to obtain valid and reliable research data with confirmatory factor analysis.
