t ratios

PRO TIP: You will rarely ever reference the t Ratio for the Intercept.

t Ratio = 20.5313 / 4.16456
t Ratio = 4.93

t Ratios are simply the t-scores for each predictor variable's significance.

t Ratio = b_k / SE(b_k)

b_k is the change in the response with each unit +/- of the predictor "k".
SE(b_k) is the standard error of b_k.

x-bar is the sample mean.
µ is the population mean.
s / √[n] is the standard error.

The equivalent of population mean (µ) would be the actual change in response from predictor variable (β_k).

And in regression... it is assumed that β_k is zero. Because unless proved otherwise through a hypothesis test, the predictor variable is assumed to not have a significant relationship with the response variable.

t Ratio = (b_k - β_k) / SE(b_k)
t Ratio = (b_k - 0) / SE(b_k)
t Ratio = b_k / SE(b_k)