I am running an econometric analysis on how some factors (gender, income level, being a student...) might affect the degree of cooperation in a weakest-link game setting. My dependent variable (the level of cooperation chosen) can have three values (1, 2 or 3) and at first I ran a random effects multinomial logistic regression, declaring the dataset as a panel using the round number as the t variable and the player number as the id variable. However, as the cooperation levels are ranked, I thought that running a rank ordered logit might be more suitable for this matter. Furthermore, I am also considering running a bayesian multilevel model as it might allow me to control for group specific effects (the degree of cooperation of a player is likely to depend on the cooperation that the other players he/she is matched demonstrate). Do you think it is proper to use this last specification for what I am analysing?
Dear Elena,
You have posed an interesting problem. The information given by you is limited and you have mentioned applying three techniques to understand the relationships between IVs and DVs. I would suggest running all the three techniques, followed by interpreting those results. If number of records is not a constraint, you can try encoding independent variables as dummy variables.
with regards
I feel that you have confused by choosing these three techniques.
We have been experienced and practiced with English word the "level of participation" which indicates that to measure the Low, Medium and High level of Participation.
Of these three levels for which we have to select Ordered Logit Model to get accurate score value to identify the three results with ranking.
It is purely directing the researchers have to select the Econometric models because of the Dependent Variable (s) at maximum. Sometimes the additional focus on the Independent Variables are also supporting to select the Econometric models.
If the number of records are not constraint that you can encoding the Independent Variables as Dummy Variables.
- Badges
- Science topic
- Similar topics
- Cognitive Science
- Thinking