Velicer's MAP test and Horn's Parallel Analysis can be used for examination of dimensionality. You can use FACTOR program. It is free and user frendly.

http://psico.fcep.urv.es/utilitats/factor/Download.html

Dear Ives,

Are you sure you only have 8 variables to start with? How can it be correct to extract 8 factors?

The eigenvalue > 1 rule tends to be a bit generous. I tend to reduce the number of factors manually one at a time after an initial extraction based on this rule. Each factor requires at less 3 unique variables with loadings > 0.4 and no significant cross loadings. Please see my study guide.

You have the option to choose which option of extraction works better for your data.

Going with either eigenvalue or number of factors should abide by certain conditions in EFA such as the number of items per factors should not be less than 3, and the higher the explained variance by the model is the better.

In your questions, if you mean each variable made a factor, then this is not acceptable in EFA.



Dear Ives,

Please. Confirm me. You have 8 variables. These are scale itens (observed variables) or are constructs (latent variable) ?

If you had the factorial model before the questionary aplication. It is better use the Confirmatory Factorial Analysis (CFA)

The use of EFA with number of factors (forced rotation) will can gerenate a irreal model. There is long time this procedure is not much used. Take carreful!

I agree with Muayyad Ahmad.

Best regards

Dirceu

I agree with Professor Dirceu and suggest:

http://www.scielo.br/pdf/gp/2017nahead/en_0104-530X-gp-0104-530X3391-16.pdf