A research questionnaire is divided into four parts with 12, 15,12, 15 statements in each part. Is it justified to apply the factor analysis separately on each part?
Usually Factor Analysis would be applied to the entire dataset.
Think about:
Does each part contain sub-parts?
What are you trying to achieve by running the factor analysis?
It might be useful to run it all together, followed up separate analyses.
It is always good to get to know your data as much as you can.
Depends entirelt on your domain separation. Are they clearly and theorized as orthogonal?
Matt
Since you already have the items which you think measure what is intended, I would perform confirmatory factor analysis (CFA). As each of the 4 parts is intended as a sub-scale, I would perform CFA on each. If the whole questionnaire is intended to measure a common concept, I would then try a second order CFA-model with the 4 latent factors as sub-factors
Best,
Generally it is applied to entire data sheet as mentioned by Samantha Curle, sometimes it happens that few items of your present variables can be listed in forming new factors in your study, This happened in my research this can vary as variables you have taken supports that sample perception while applying factor analysis to entire data set can yield to some different variable which will show uniqueness in your study on that sample.
- Badges
- Science method