Hi,
Can we test endogeneity in SPSS/AMOS to check the simultaneity (direction of causality)? If yes, can someone explain how? The data are cross-sectional.
Thanks in advance.
Endogeneity occurs because of omitted variables and it biases the cause-effect relationships in cross-sectional data.Therefore find out the instrumental variable that is related to independent variables but unrelated to the disturbance term (observed Y-expected Y). Finding an instrument is a difficult task and one needs to theoretically justify it, and. then, empirically verify it. D Suar
You can choose for this analysis R, Stata or SAS - to my knowledge.
You can use just about any package, as long as you have a model (e.g., instrumental variables) that allows you specify such an effect. However, I personally have doubts about the meaning of assessing such models in cross-sectional data. For example, are you expecting both of the causal coefficients to be positive? If so, then you are predicting unlimited growth. Or, if you think they are both negative, this will drive them both to zero. So, unless you are predicting a well-balanced pair of positive and negative coefficients, your system will be inherently unstable.
Endogeneity occurs because of omitted variables and it biases the cause-effect relationships in cross-sectional data.Therefore find out the instrumental variable that is related to independent variables but unrelated to the disturbance term (observed Y-expected Y). Finding an instrument is a difficult task and one needs to theoretically justify it, and. then, empirically verify it. D Suar
you can specify the model and use the IVreg command in stata it will give you the wu-hausman test or used the ivendog command in stata after the ivreg command to know if your suspected variable is actually endogenous or exogenous
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS