To do total score, you need to follow the following steps

in SPSS, Tansform, compute, give name for factor1 under Target variable {you need to do these steps for the other 2 factors with different names}

then the items that you got from factor analysis for factor one need to be selected one by one and entered to [the window titled Numeric Expression]. you have to be careful in entering these items...so first select the items, click the arrow in the middle of the screen {left to numeric expression} then click the add icon. then do the same for the 2nd items, until you finish all items of factor one. Then, click ok. you will get in the data sheet the new variable which corresponds to Factor one total score.

Hi Khalil,

You can simply run correlation syntax  for F1 & F2, F1 & F3 using SPSS.

(I hope you have taken care of other assumptions while doing any kind of multivariate statistical analysis such as sample size requirements etc...).

If you are further going for path modelling or SEM, you can have other measures to check latent correlations.

Regards,

Abhay

 Hello Abhay,

Thank you for your reply. I did take care of other assumptions as suggested, thank you for the reminder.

Regarding your answer, I am aware that I need to do a correlation syntax for F1 & F2, and F1 & F3.

The problem is, when I go to SPSS, Analyze, Correlate, Bivariate, I can only add questions to the variables, and not factors - that is, I will have to add all the questions attributed to F1 and F2 to the variables section on the right of the opened window. I don't think this is how it should be, right?

Any other way to do this?

Thank you in advance.

Hello Khalil,

If you computed and saved factor scores for F1-F3, you can calculate correlations among the factors using those scores.  Of course, if you used orthogonal rotation to produce the factors, the scores will be uncorrelated.  However, if you used oblique rotation, there may be some degree of correlation among each of the three pairs of scores.

Regards,

David

First, you need to do total score for each of the 3 factores.

second, in SPSS, Analyze, correlate, bivariate correlations;; then enter the 3 factors total scores. [keep the default setting to do Pearson's correlation]

Hello David,

Thank you for enlightening me on this. I did conduct an oblique rotation analysis and the results were not encouraging so I conducted an orthogonal analysis and the results were good. I thought I could perform a correlation analysis now.

So, based on what you said, there is no way there is a correlation between the factors of my survey? Even if it doesn't make sense that the first factor is not correlated to the other factors?

Thank you.

Regards,

Khalil

To do total score, you need to follow the following steps

in SPSS, Tansform, compute, give name for factor1 under Target variable {you need to do these steps for the other 2 factors with different names}

then the items that you got from factor analysis for factor one need to be selected one by one and entered to [the window titled Numeric Expression]. you have to be careful in entering these items...so first select the items, click the arrow in the middle of the screen {left to numeric expression} then click the add icon. then do the same for the 2nd items, until you finish all items of factor one. Then, click ok. you will get in the data sheet the new variable which corresponds to Factor one total score.

Kindly find below an update on my work and let me know if this makes sense to you:

1) I conducted an oblique rotation analysis and found a weak correlation between the three factors - yet, there is a correlation. Please refer to picture titled "Point 1".

2) As the oblique rotation analysis performed the analysis on the three factors altogether, and my aim is to perform an analysis between F1 and 2, and F1 and 3. I decided to name the factors by total scoring them in the compute variable section. After defining F1, F2, and F3, I conducted a correlation analysis between F1 and F2, and F1 and F3. I finally obtained better results. Please find attached the results in the two attached pictures titled "Point 2 - F1 and F2" and "Point 2 - F1 and F3".

Is this acceptable?

I noticed that the only significant correlation is between F1 & F3.

you need to create table to inlcude the three variables together.

Do you mean this one?

yes, but you need to show only half of the rectangle

don’t repeat the correlation

use only the r value, and add a star over the significant one. Use foot note to indicate that the star means p < .05

Hello Khalil,

I would like to go right back to basics.  It seems that you conducted factor analysis on data from only 40 people.  One of the basic principles of factor analysis, which is admittedly broken under some circumstances, is that there should be at least 10 times the number of respondents as there are items.  Sometimes people say that there can be as few as 100 people in a factor analysis, but I think that might be risky.

So, I would ask how many items you had that you subjected to factor analysis.  If you don't have enough respondents, the factor analysis is likely to be quite misleading and even perhaps totally meaningless.

Could you therefore let us know how many items you are working with please?  It might (or might not) be possible to pick up the pieces from there - particularly including the advice that Muayyad has given you above.

Robert.

Hello Robert,

Thank you for your message.

I actually had 29 items. You are totally right that the number of respondents should be 10 times that of the items but since I was only conducting a pilot study, I could not have that many respondents.

Hello Khalil,

I wrote a long response to your post immediately above, but it seems to have disappeared.  Please let me know if you didn't get it and I'll try to revive it!

Robt.

Hi Robert,

I unfortunately did not receive it. I would love if you could revive it.

Thank you.