Dear all,
I'am using mlogit to test IIA hypothesis. I'get the test statistique who
is negative, a p-value of 1, and a text saying that the IIA hypothesis is
rejeted. Does a negative value mean that constraint model is more efficient that
that the full model ? what to do in this case ?
Sincerly.
Justin BEM
BP 1917 Yaoundé
Tél (237) 76043774
[[alternative HTML version deleted]]