Why do we use the t-statistic to find out the significance of the predictors while we use the z-statistic to compute the confidence interval of the predictor coefficients?

linear_regression

#1

Hello,

The below image shows a regression output and the confidence intervals for the coefficients:

In the above part I have a question:

Why do we use the t-statistic to find out the significance of the predictors while we use the z-statistic to compute the confidence interval of the predictor coefficients?
Is it because t-statistic is for the sample of data whose population variance is not known and the z-statistic because we are estimating population parameters??
Can someone please help me on this??


#2

Hello,
I think in both the cases t-statistic is used. Since the degree of freedom is pretty high, the t-critical value approaches 1.96(the z-critical value) but still it isn’t spot on (t-critical = 1.9606).

Regards