Hi,

Confirmatory Factor Analysis:  It is a model a prior

1.  We can compare number of factors

2.  In this analysis, Which items load on each factor

3.  A model supported by theory or previous research

4.  Its a Error explicitly

Exploratory Factor Analysis:

1.  To determines the factor structure or model 

2.  To Explains a maximum amount of variance

For your question, how to measure composite reliability, convergent validity, discriminant validity and AVE in SPSS, kindly see the following link:

http://www.youtube.com/watch?v=DqN_mpqGLtA

Hi,

Confirmatory Factor Analysis:  It is a model a prior

1.  We can compare number of factors

2.  In this analysis, Which items load on each factor

3.  A model supported by theory or previous research

4.  Its a Error explicitly

Exploratory Factor Analysis:

1.  To determines the factor structure or model 

2.  To Explains a maximum amount of variance

For your question, how to measure composite reliability, convergent validity, discriminant validity and AVE in SPSS, kindly see the following link:

http://www.youtube.com/watch?v=DqN_mpqGLtA

Very good video Senthilvel, In-fact James Gaskin always makes wonders to his audience.

@ SUhail, It is always advised that to perform EFA first then CFA. 

IBM's SPSS can help you in EFA but for CFA you have to use IBM's AMOS.

EFA explores or reduces the huge no. of factors into few, by showing which factor loads on which latent factor. To find hw many latent factors, it depends some thumb rule like Eigen value greater than 1, screen plots, factor loading greaterthan 0.55 and cross loadings less than 0.45.

Better to use IBM's AMOS for CFA, where it confirms the factor structure extracted by EFA and is a theory oriented model.It is a model which tests  to which extent the relations are valid. To be more confident and certain about the hypothesised structural model, one need to look at composite reliability  and average variance extracted and also to  perform advanced test if validity like convergent and discriminant validity tests. All these are not conducted in EFA, it is just a priliminary analysis. one can also exclude the factor loading which are not meeting the above mentionec criteria.

Thank you  Mr Rahul Pratap Singh Kaurav and Ms Sania Khan 

Thanks a lot Mr. Senthilvel Vasudevan

Hi Suhail

adding to these wonderful contributions by our circle in this field, note that there are 2 variants of SEM:

1. Covariance based, e.g AMOS, LISREL

2. PLS based, SmartPLS, WarpPLS

*PLS = partial least square

In both variants, researchers normally use a 2-step approach suggested by Anderson and Gerbing. These are:

1. Measurement model assessment (this is CFA)

2. Structural model assessment (vary based on which SEM variant, e.g. Covariance SEM uses fit indices such as RMSEA, TLI. While PLS-SEM uses Q-square, R-square)

Your question on composite reliability, AVE, discriminant and convergent validity lies in the factor loadings. And these are stats at CFA in both SEM variants.

These links give you what to expect for CFA. although written for PLS-SEM, the measurement model (CFA) assessment stage remains similar in both SEM approaches.

https://www.researchgate.net/publication/267871651_PLS_analysis__Guidelines_to_evaluate_measures?ev=prf_pub

https://www.researchgate.net/publication/267815037_Partial_least_square_in_a_nutshell__Saiyidi_MAT_RONI_2_0_1_4

Best wishes

Saiyidi

Article PLS analysis – Guidelines to evaluate measures

Chapter Partial least square in a nutshell | Saiyidi MAT RONI 2 0 1 4

Thanks Mr. Saiyidi Mat Roni 

For EFA use SPSS for CFA use AMOS. CFA tests and compares models which is what you do with CFA  

 how can I conduct CFA in SPSS?

Can anyone help?

Thanking in advance!

" You should only do an EFA if your instrument has never been explored before"-  does this mean that we can skip EFA  and directly go for  CFA  

If only a new mediator  inserted which is rarely studied before, should EFC be done?

Dear Mariya Razzaghian

SmartPLS for PLS-SEM , AMOS , R and MPlus software used for CFA and Structural equation modeling (SEM) analysis