Predictive model for Turnover - Correlation or Survival Analysis

[srivabs]

Question 1 - I a working on developing a model to predict employee turnover. The variables that I have are age, tenure, job satisfaction, role clarity etc. Through research papers I have found the correlation coefficient for these variables but I am not sure on how to develop a model around it?? Some leads on how to go ahead with this??

Question 2 - Can these variable be used in survival analysis models like Cox??

Which technique would be better.

I am new to statistics but ready to learn

 

[Miner]

Survival analysis is a time to an event analysis. If you are analyzing time till turnover, they would be appropriate and a proportional hazards model would allow you to include variables such as those you mentioned.

 

[srivabs]

Can you please tell some way around for question 1 as well..

I have correlation values of various factors with turnover but how to use them in a predictive model??

 

[Miner]

I would use the Cox proportional hazards analysis to create your predictive model.

 

[Cope_LA2]

I found the following link really useful for my first run through a Cox Proportional Hazards model if you need a point in the right direction

http://www.amstat.org/chapters/northeasternillinois/pastevents/presentations/summer05_Ibrahim_J.pdf