I have recently compared results form polo plus 2.0 and spss 27.0. whit the same dataset results are quite different. has anyone experience on this? many thanks in advance
It would good to look at whether it refers to the coefficients or marginal effects. The marginal effects can be computed in different ways (average ME, ME at the mean, etc.). Take also into account how binary variables are computed. In some software you might need to tell the software they are indeed dummy variables, but it affects to ME, as far as I know.
I think there are papers documenting small differences in the results using different packages. I assume that being a ML estimator, where the estimator is reached following an iterative process, there could be differences...
This is also likely to happen when you compare the results from stata and SPSS. A small difference would emerge as usual in terms of parameters coefficient.
I checked this with my data and there isn't much difference. I think SPSS doesn't consider the control treatment.
Hi
Respectfully, the difference in the amount obtained should not be much based on my experience
Be careful to check the amount LC50
Probit is correct if it is to analyze a parameter.
The Polo program is mostly used for the amount of LD50
But it is still the correct software for calculation.
The basic premise is Probit in SPSS software as far as I know.
Good luck
- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts