confidence interval... I am doing empirical researach, usually perform differences-in-differences, instrumental variable method. my most often used models are: probit, logistic, and tobit regression. I of course report confidence interval in my result, but I barely really use it in my final analysis. I often use 95% confidence interval level, instead of 99%. But I think it's a quick way to use confidence interval to reassure the significance of the coefficient. If zero is within the 95% confidence interval, your coefficient is NOT statistically significant. It's more intuitive than p-value and it's quicker than carrying out t-test.
|