Expanding on Dr Klosowska's very good comments, you need to think about what you are trying to do with your analysis.

Exploratory Factor Analysis (EFA), of the type you get in SPSS Factor, is used to discover what items go together and that can be summarised by a small number of factors.

If you already have a theory that certain items should go together then you do not need to discover the factors. Instead, you want to check that the items do indeed conform to your theory. For that, you use Confirmatory Factor Analysis (CFA), which is not available in basic SPSS. The SPSS AMOS module is very good for CFA.

In the meantime, you can do some quick but crude analyses using SPSS Factor. Take only the items that are used to measure "job satisfaction" and apply EFA to them. Confine the factor solution to one factor only. (Obviously, rotation will not occur). Check the communalities and eigenvalues to see if these are acceptable to you.

And here is another possibility...

Consider that many measures of topics such as "job satisfaction" and "organisational commitment" were created by middle-aged upper-middle-class white males in Mid-Western American universities, for use with other American white upper-middle-class employees about 30 years ago.

Is it possible that these same items don't mean the same thing with your sample? My colleagues in Australia have certainly found this.

Just to underline the very good statements above, here are some references that might help. I would suggest, based on your research aim, to use AMOS.

Byrne, B. M. (2013). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Routledge.

Blunch, N. (2012). Introduction to structural equation modeling using IBM SPSS statistics and AMOS. Sage.

Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A., & King, J. (2006). Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of educational research, 99(6), 323-338.

https://books.google.de/books?hl=de&lr=&id=0cHbAAAAQBAJ&oi=fnd&pg=PP1&dq=AMOS+CFA&ots=_-UAkEgvhT&sig=KlXVY9PnwwLPZpMY0-F8Qe8A8t8#v=onepage&q=AMOS%20CFA&f=false

https://books.google.de/books?hl=de&lr=&id=dmDez8-ygc4C&oi=fnd&pg=PP2&dq=Blunch,+N.+(2012).+Introduction+to+structural+equation+modeling+using+IBM+SPSS+statistics+and+AMOS.+Sage.&ots=48p3vPiokj&sig=dJt1K0eyROOYZbuIVc663KuY4P8#v=onepage&q&f=false

http://www.tandfonline.com/doi/abs/10.3200/JOER.99.6.323-338

http://search.proquest.com/openview/b0ebfd2ea853bb8fa909400604d51238/1?pq-origsite=gscholar

https://www.youtube.com/watch?v=JkZGWUUjdLg

My aim was to check the validity of question items through EFA by finding items of each construct on same related factor. However, if EFA through PCA and Varimax rotation is not a solution then should I conduct CFA. My aim is not to find the relationship among variables rather to validate the instrument that may confirm the measurement what the instruments claim to be. 

It is important to clear here that Instruments are already validated in various previous researches, so is there need each time to find the validity of already validated instruments whenever to use? So, kindly give suggestion about the matter, please. 

I would suggest that we should always consider (and if possible, test) validity and reliability of our measurement instruments. As was pointed out above, the question would be if the questionnaire has been validated in a sample that is almost identical to your sample (Age, Gender, Education, Industry ect.)? Otherwise, it might indeed be problematic to just claim that your measurement instrument is valid. Further, your results suggest that the proposed structure is not what you have expected. And that you definitely need to consider.

You don't indicate how you decided on the number of factors to rotate. You have 4 scales it looks like, but you rotated 7 factors. You might have too many factors, and that can result in factors breaking in two. If your purpose is merely to verify that items of the four scales load together, I would rotate 4 and see if the item loadings match the original scales. That said, current practice in the journals is to use a CFA to verify that scales match factors. An advantage of CFA is that it provides overall fit statistics that indicate how well the data fit the hypothesized structure, although you cannot see the magnitude of cross-loadings llke you can in EFA because in CFA they are fixed to zero. If you are thinking to publish this, I would probably use CFA. If you decide to stick with EFA, many reviewers will complain if you report a principal components analysis--I would switch to a common factor model. Results will likely be very similar, but there are theoretical reasons many people prefer the common factor model.

The following paper provides a nice discussion of CFA vs. EFA.

Hurley, A. E., Scandura, T. A., Schriesheim, C. A., Brannick, M. T., Seers, A., Vandenberg, R. J., & Williams, L. J. (1997). Exploratory and confirmatory factor analysis: Guidelines, issues, and alternatives. Journal of Organizational Behavior, 18(6), 667-683.

Thanks to everybody who contributed in an interpretation and gave nice suggestions which ultimately have given a direction and solved the problem for this time as well as further help in future. 

M. Imran

With Exploratory Factor Analysis (EFA), researchers can decide the number of factors by examining output from a principal components analysis (i.e., eigenvalues are used).

To assess the EFA, factor loadings with less than .40 to be excluded from the construct. Therefore, the three factors (1, 5 & 7) loaded in job satisfaction can be taken if each of their loading is above .40.

You may refer to the following link in Exploratory Factor Analysis.

https://www.youtube.com/watch?v=-6tw7ebr860

Mr. Khan

I suggest you must go for separate factor analysis for each variable then there will not be issue of loading on different factors of other variables. Moreover, as suggest by Judeh the loading values above .40 should be considered and the statements having values higher than .4 on two separate factors but their difference is lower than .25 should also be excluded from further analysis.

For analysis Principal component analysis with varimax rotation can be used and in SPSS >> Factor analysis>> under 'options'>> choose   'Sorted by size' 

This will arrange the statements according to loading values in ascending order as well as most dominant factor will appear first.

I hope this will help.

Thanks, Ms Garima Mathur and Mr. Mahfuz Judeh for kind help. However, some researchers neglect to conduct EFA on the basis that there is no need to conduct EFA for already validated instrument.  So, should I leave EFA or it is necessary due to differences in cultures between where the instrument was validated and the another one.

M. Imran