Dear all,
I used those two types of regression to estimate effects of environment factors on hospital efficiency.
However, they gave me opposite results. For instance, in FR percentage of people above 65 was negatively associated with DEA score, whereas in DBTR it was positively associated. The same results was for turnover rate.
I appreciate if you could explain me such contradictory.
McDonald, (2009) suggested that bootstrap truncated regression was not suitable and inappropriate in the later stage of DEA. According to that study, TE is fractional data, which is not generated by a censoring procedure. The study revealed the appropriateness of Ordinary Least Square (OLS) as a suitable method.
Reference: Article Using least squares and tobit in second stage DEA efficiency analyses
Aleksandar P. Medarevic Double-Bootstrapped Truncated Regression please see
Article Estimation and Inference in Two-Stage, Semi Parametric Model...
Article Two-stage DEA: caveat emptor
Dear Ilyas and Nandy,
thank you for your comments. However, my question is refereed to "fractional regression". Using frm R package I get output (attached). Does it means that increasing in v15 leads to greater or lower efficient?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression