I have collected data (n=1126) to measure Materialism (8 items), Religious Activities (3 items), Religious Influence (5 items) and "Locus of control (LOC)" Rotter's Scale short version (13 items). Afterward, Internal and external locus of control are calculated (value ranging from 0 to 13). Materialism, Religious Activities and Religious Influence are collected on Likert Scale (7 point). Age, Education and Gender related information are also collected.
The hypotheses are as such:
1. There is a relationship between Materialism and LOC
2. There is a relationship between Religious Influence and LOC
3. There is a relationship between Religious Activities and LOC
4. Age and Gender can be considered as mediating variable (Optional)
Preferably, I like to use SEM. The question is Likert scale seem to be Ordinal rather than Interval. Hence, is it possible? I already have developed a model (roughly). Please comment whether is it properly drown. The indices implies a moderate fit of the model.
Please consider the attachment.
Dear Muhammad
SEM analysis seams to be perfect in this case. Forsure you can use Likert scale for measurement your items. in this case LOC would be consider as dependent variable (indogenous), and the rest could be exogenous (independent variable). Also you may inter the genter into your model as moderator between DV and IV.
Regarding your attachment, I am not agree with this model. As we know that Independent variable must be on left and DV must be right of your screen. Accordingly, the items of each construct would be on their left or right.
Hope this could help you.
If the Likert-scale has 7 steps we use it in psychology as an interval scale. See Tabachinik and Fidell.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling