In wikipedia's definition of a linear regression model:

$y_i = \beta_1 x_{i1} + \cdots + \beta_p x_{ip} + \varepsilon_i = \mathbf{x}^{\rm T}_i\boldsymbol\beta + \varepsilon_i, \qquad i = 1, \ldots, n, $

Then am I correct in saying the following?

  • $\varepsilon_i$ are random variables $\Omega \rightarrow \mathbb{R}$.
  • $y_i$ are random variables $\Omega \rightarrow \mathbb{R}$.
  • $\beta_i$ are constants in $\mathbb{R}$.
  • $x_{ij}$ are constants in $\mathbb{R}$.
  • $\begingroup$ The variables can be complex, but we usually deal with real space. $\endgroup$
    – Aksakal
    Commented Apr 27, 2015 at 15:08
  • $\begingroup$ $x_{ij}$ constant ? $\endgroup$ Commented Apr 27, 2015 at 15:16

1 Answer 1


In the classical linear regression framework, you are correct. The only thing I would like to add is that the true, unknown $\beta_i$ are constants but the estimated $\hat{\beta}_i$ are random variables as they depend on the random variables $y_i$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.