Hello Everyone,
Following review of literature, I have designed following alternate hypothesis for confirming the validity of a model.
I have used a Likert based survey to collect data for my Constructs- A, B, C, D
Some of the constructs have been re-operationalized
The Cronbach alpha for scale is 7
Following are the alternate hypothesis.
H1a: There is a positive effect of A on B
H1b: There is a positive effect of C on B
H2a: There is a positive effect of type of D on B
H2b: The effect of A on B is dependent on the type of D such that the B is stronger in presence of certain types of D.
H2c: The effect of C on B is dependent on the type of D such that the B is stronger in presence of certain types of D
I am proceeding to use CFA for my purpose. My questions are as follows
1. Since I have used 5 point Likert scale to collect data for various constructs. I was advised that I need to apply z transformation to proceed further. But can we apply z transformation on OPINION based data. It is acceptable? If not what is the alternative.
2. Should I also state alternate hypothesis for various variables of the constructs or its it fine if I state the alternate hypothesis for checking the effect of constructs on each other.
3. Are the alternate hypotheses for checking moderating behaviour of construct D stately correctly?
Thanks and regards,
Rekha Mishra