To determine the slope [[Linear Regression|linear]] [[Correlation|correlation]] between two [[Normal Distribution|normally distributed]] [[Random Variable|random variables]], setup a two-tailed [[Hypothesis Test|hypothesis test]] on the population correlation slope $\beta_1$ with the slope of linear regression $b$ as the test [[Statistic|statistic]].
Establish the [[Null Hypothesis|null hypothesis]] that there exists no [[Correlation|correlation]].
$
\begin{align*}
&H_0: \beta_1 = 0 \\
&H_1: \beta_1 \ne 0
\end{align*}
$
Use the [[Z-Score|standardised]] test statistic $t = \frac{r}{\sqrt{1 - r^2}/\sqrt{n - 2}}$ (using the [[Variance|variance]]/[[Standard Deviation|standard deviation]] of the correlation coefficient) on the [[t-distribution]] with $(n - 2)$ degrees of freedom (lose 2 degrees from two variables).
Reject $H_0$ if $t \ge t_{\alpha / 2}(n - 2)$ or $t \le -t_{\alpha / 2}(n - 2)$.
If the null hypothesis is rejected, then there exists a linear relationship between the two variables.
Note that $\frac{r}{\sqrt{1 - r^2}/\sqrt{n - 2}} = \frac{b_1}{s(b_1)}$, where $s(b_1)$ is the standard error of $b_1$, the variation in one variable not explained by another.