I have a model and it has many construct. Most of these constructs are multidimensional ones (being multidimensional construct is not a main point to discuss in my research!) Let's say that I want to test the relationship between two multidimensional constructs. My questions are:
1- should I show that my constructs are multidimensional when drawing my model?
2- Of course I need to justify why I want to test the relationship between any two constructs! my question is: should I explain the relationship between two constructs in general or should I go into more details about each dimension of each construct?
3- I am clueless how to test the relationship between two multidimensional constructs in SPSS!!!
Please! help me with this! Thanx
Hello Saleh,
I see two possible paths:
1. If you are uncertain as to how the sets of relationships will manifest, you could use canonical correlation. Treat one multi-dimensional construct as one set of variables, and a second MDC as the other. The canonical R (for each possible root) gives a measure of the degree of association.
2. If you are proposing specific relationships (as in a structural equation model), then each MDC is, itself, a set of related latent variables (whether all in the same plane, or on multiple planes as in a second- or higher-order factor solution). So, the proposed model simply has a lot of latent variables. However, to evaluate this would require the AMOS add-on for SPSS (or, for that matter, any of a large number of other SEM-capable software packages).
Good luck with your work.
you can use SEM between structures by SPSS AMOS , by SEM you can study two types of effects : direct effect between to structures and indirect effect between independent variable and dimensions in dependent variable , best regard
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS