I am using Data Envelopment Analysis (DEA) and Stochastic Frontier Analysis (SFA) to measure the eco-efficiency of dairy farms. I would like to use the "percentage of grasslands in the agricultural area" as an input variable. This variable is expressed in % (from 0 to 100%). Can I include variables expressed as percentage in DEA and SFA models? If not, how to deal with such variables?
I suggest you to read this article
https://link.springer.com/chapter/10.1007/978-0-387-71607-7_17
Hi Pauline,
Magnus Richter is totally right and thanks for the EJOR source! The rule "The more similar the number format, the more resilient the results of the DEA efficiency calculation are" worked well for me.
Good luck with your upcoming research,
Dominic
In conjunction with the above answers, take a look at the following:
Article Selective convexity in DEA models
Article DEA models for ratio data: Convexity consideration
Agreed with the abovementioned. However, you should be cautious with the interpretaition of your results (slacks and targets in DEA and elasticities in SFA). If you decide to employ a percentage variable as an output one in the SFA context, things become more complicated
agreed with the above answers and Magnus Richter but the most simple is to convert it into log. the coefficient will be automatically in percent.
- Badges
- Science topic
- Similar topics
- Correction