In a simple linear regression $Y_i = \beta_0 + \beta_1 X_i + \epsilon_i$, suppose you scale the predictor by a factor $k > 0$, such that $X_i^* = kX_i$. How does the new least-squares slope estimator $\hat{\beta}_1^*$ compare to the original slope estimator $\hat{\beta}_1$?