Dear Hadeer,

In response to your questions:

Have a look at my guide: https://www.researchgate.net/publication/304490328_Advice_on_Exploratory_Factor_Analysis

Working Paper Advice on Exploratory Factor Analysis

How did you decide to examine 3 factors? If you used Eigenvalues greater than one as the criteria, that is a very weak way to make a judgement. A better approach would be to examine a "scree" plot to determine whether the smaller factors are really do deserve attention. Even better would be to do a parallel analysis.

Also, you did not tell use what extraction method you used or whether you examined a rotated solution. I would recommend an oblique rotation (correlated factors), which means you should avoid principal components analysis and use either principal factors or maximum likelihood instead.

Dearest Dr. Peter and Dr. David,

Thank you so much for your reply and your help.

I decided on the three factors based on parallel analysis. For the extraction method, i used principal component. For the rotation method, i used promax. 

i have followed the approach of removing the cross-loading items and re-running the analysis.  The pattern matrix shows one factor with loadings from the relevant items. yet, the other two factors still have irrelevant items. 

My question: Can i interpret the factor based on the relevant items since there are double the number of irrelevant items and they have high loading?. for example, factor 1 has 8 relevant items (From Competence dimension) but 4 irrelevant ones(from autonomy dimension) while factor 3 has 6 items for autonomy and 3 for competence. so does it make sense to say that factor 1 is competence and factor 3 is autonomy? 

kindly find the output attached. 

Thanks in advance. 

Promax forces the factors to be uncorrelated. Do you truly believe the Autonomy has no relationship what so ever to Competence?

Dearest Dr. David,

The definitions of competence and autonomy are as follows:

Competence: refers to the extent to which individuals are able to impact their environment, to have opportunities to express their capabilities, and to handle the challenges in their lives.

Autonomy: refers to the degree to which individuals experience free choice, volition, and responsibility for their own actions and behaviors.

Theoretically these two dimensions relate to one’s self-determination. Yet, I am not sure I have enough arguments and support to argue that competence and autonomy can load on the same variable. This is why I am wondering if I can interpret factor 1 as competence and factor 3 as autonomy.

The reason I am suggesting that you do an oblique rotation is that it may give you a different picture of the loadings, in comparison to an orthogonal rotation, which forces the factors to be uncorrelated.

As far as arguments about what to name the factors the factors in terms of what they do or do not measure, that raises substantive questions that no factor analysis can address.

Hello Hadeer,

Promax rotation is an oblique method; it allows extracted factors/components to be correlated.  I agree with David Morgan that a different extraction method (e.g., principal axis, maximum likelihood) would be preferable to principal components (which makes very strong presumptions about the nature of common variance).

Good luck with your work!

Sorry about the confusion on Promax.

Dear Hadeer, 

I think that you have an ordinal-type data. Therefore, it is recommended to do EFA on a mtrix of Polychoric correlations instead of using Pearson correlations. You can do this using the free program "FACTOR", which can be downloaded from:

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

Good Luck