I am working on Scale development, about efa should i use PCA with varimax or PAF with oblique. i have set of statements not dimensions so i dont know how to see the concept of correlated or uncorrelated dimensions. Further should i first resolve the outliers issue and then proceed towards efa?
You must remember that oblique rotation fits more to correlated factors, but if you do not have a theoretical model for dimensions and consider unifactorial you can perform both analysis if it fits your evaluation can be one-dimensional and then if you would have to perform confirmatory factor analysis for a more accurate analysis of evidence of validity. It is also recommended to adjust the atypical data.
How have you constructed your scale? Did you construct it to measure a single construct, or are there multiple facets/dimensions to the area that you are looking at? From your statement "I have a set of statements, not dimensions, so I don't know ... " I am assuming you have not had formal training in psychometrics. A psychometrics (or related) course might be useful before you begin constructing your scale.
On basis of literature review i have generated set of statements employed a deductive approach. how researchers look at it types have lead to statements which are further being tested to measure single test.
- Badges
- Science topic
- Similar topics
- Psychology