Hello,
I am going to create a model to evaluate the correlation between project success criteria (dependent variables) and employee measured capabilities (independent variables)
Project success criteria are both categorical variables (questions answered with yes/no) and interval variables (cost, time).
I am considering either multivariate logistic regression or 2 regressions: a multivariate multiple linear regression (for interval dependent variables) and a multivariate logistic regression (for binary dependent variables). I have also read about canonical correlations and structural equation modeling, but I have not completely understood the difference among models.
The correlation will then be used to estimate the success of future projects according to available people.
Thank you!
Most straightforward multivariate models assumes that the outcomes are independent (ie that the errors of the models are independent). Here, they are, as I can see, different measures of the same, and errors are likely correlated. Check out gsem and cmp (user written command) in stata, to try structural equation modelling or seemingly unrelated regression With different outcome variables. These are complex models.
You can use the multiple linear regression for cost ant time as dependent variable, but must check all the assumption of least squares method.
For binary response dependent variable you can use binary logistic regression model.
Canonical correlation and sem can be used also.
Thank you both! I will start with a simple multivariate multiple linear regression, as it is the easiest to use in Excel, but I will try something more complex in R.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis