New answers tagged

3

Standard error is not sd but se. Standard error is related to sd but they are not the same $\text{se}=\frac{sd}{\sqrt{n}}$. You can also confirm here that in the first case $1.41\approx 4.46/\sqrt{10}$ Confidence interval is, for 95% confidence and using t-statistics, $\pm t^* \cdot\text{se}$, not $\pm 1.96\cdot se$. The critical value of $t^*$ at 95% level ...


2

The Neymen-Rubin potential outcomes terminology is is not typically used in economics outside policy evaluation where your policy will be binary. This being said there are still counterfactuals. For example, if you are regressing interested rate on car purchase, if at time $t$ interest rate $i$ was 6% and associated car sales $s$ were 500 of cars at ...


Top 50 recent answers are included