0
$\begingroup$

Can someone explain to me what a Wald test is please. I have seen it twice today and cannot figure out what it does (other than test for sig difference like a t-test or chi-square). What would you use it for, and what makes it the right test to chose? Thanks in advance Cat

$\endgroup$
0
$\begingroup$

A Wald test is any statistical test that imposes a restriction (null hypothesis, say, $H_{0}=0$ for t-test test of significance of an estimator) and then tests the validity of this restriction to decide wether to keep a variable/retain the alternative (in the example of the t-test again, $H_{a}\neq 0$). It is one of three classes of statistical tests (see [enter link description here][this paper]’s introduction + more if you’re interested). The three types of tests are classified as: Wald tests, Lagrange Multiplier (LM) tests, and Likelihood Ratio (LR) tests; all have a slightly different approach to testing.

You would likely use Wald tests for deciding wether or not to remove a variable from a regression model. It is perhaps its most used application. Or to decide wether or not to include a quadratic term of a variable in a regression model.

$\endgroup$

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