High t-ratio(0) for WTP estimates
Posted: 22 Oct 2020, 23:37
I have estimated two mixed logit models in WTP space. In the first model, beta_cost is fixed while all other parameters are specified as random with normal distributions. The second model is identical to the first except that beta_cost takes on a negative exponential distribution. WTP estimates are nearly identical but the t-ratios for WTP estimates in the second model are very high, ranging from 50 to 350. Both models seem to have converged properly with cc<1. Should I be worried about these high t-ratios (very small standard errors)? Is so, what could be causing the second model to behave this way?