UT theoretical chemistry code forum

I am interested in the activation energy for oxygen migration and how it varies with doping levels in my system. I chose the higher energy configuration as initial and lower one as final state and performed the TSS calculation with 3 images. But the activation barrier from simulation (0.3-0.6 eV) seems to be very different from experimental values(around 1eV) and the trends don't match. If I use the reverse i.e lower energy config as initial and higher energy config as final states, the barrier ( 0.7-0.85 eV) seem to be slightly closer to experimental and the trends match.
I am confused whether I can analyze my result as such(I feel it is wrong) . Do we always take the higher energy config as initial state ? From Transition state theory that seems to be the case.