Integrated cases: reading a full regression printout

In short

• From R² and the sign of the slope get the correlation: $r = \pm\sqrt{R^2}$, with the sign of the slope. 'Multiple R' on a printout is $|r|$, not $R^2$.
• Two 'standard errors' sit on a printout: the SEE ($\sqrt{\text{MSE}}$, in Y units) in the regression statistics, and the standard error of each coefficient in the coefficient table.
• With one predictor, $F = t^2$ for the slope, and 'Significance F' equals the slope's p-value. Check that SSR + SSE = SST and the mean squares are SS ÷ df.
• Test the slope against any hypothesised value with $t = (\hat b_1 - B_1)/s_{b_1}$; the printout's t-statistic tests against zero only.
• In log-lin and log-log forms, convert slopes exactly: $\%\Delta Y = e^{b_1\,\Delta X} - 1$ in a log-lin model; the approximation $b_1\Delta X$ is good only when it is small.