0

I am trying to analyze the trend of cesarean section rates in several countries for 1980-2014. Given that this outcome variable can only adopt values between 0 and 1 I am wondering which is the best model for this data. I thought of a Tobit model but I am not sure if there are any other better alternatives.

Thanks!

Mercedes
  • 13
  • I forgot to mention I the data has many missing values, so the other alternative I was considering was a GEE model, is it possible to add restrictions? – Mercedes Jul 10 '15 at 16:58
  • You might look at total values instead of rates, and then apply content from the following: http://www.analyticsvidhya.com/blog/2015/03/framework-application-build-arima-model If you can provide some dummy form of the data, then I could make an answer that goes step by step, has source code, and works with your input data format. – EngrStudent Jul 10 '15 at 17:17
  • Thanks a lot! I do not have the absolute values of the number of c-sections. I could estimate them by using the rates and number of births per year. The problem with this approach is that it won´t answer my question, since the the increase in the number of c-sections could be due to an increase in the number of births. That is why I was going to use rates. I am analyzing the change in the number of c-sections relative to natural delivery through time. – Mercedes Jul 10 '15 at 20:08
  • Why not predict the number of c-sections then divide by the expected population? – EngrStudent Jul 10 '15 at 20:10

0 Answers0