6
$\begingroup$

While most of the tests for normality help for continous variables, is there a way to test the normality assumptions for binary variables.

From what i've read on wiki, the K-S test can be applied for continous variables. How do we perform normality tests for binary (or even categorical variables)?

Improve this question
$\endgroup$
1
  • $\begingroup$ vote to close. (but I'm below 500r) $\endgroup$
    – M. Tibbits
    Commented Oct 12, 2010 at 13:47

1 Answer 1

Reset to default
13
$\begingroup$

There is no such thing as normality of categorical variable. Normal distribution is a continuous distribution so in assumption don't cover categorical output.

Improve this answer
$\endgroup$
3
  • $\begingroup$ Ok, let me correct my question. How do you know if a binary variable follows the binomial distribution? $\endgroup$
    – Roger
    Commented Oct 12, 2010 at 12:42
  • 1
    $\begingroup$ See en.wikipedia.org/wiki/Binomial_test . Talking as a moderator now: so can you please make a new, good question? You'll get a better answer and the wiki spirit of this site will be held. $\endgroup$
    – user88
    Commented Oct 12, 2010 at 12:49
  • $\begingroup$ Sure. Let me do that. $\endgroup$
    – Roger
    Commented Oct 12, 2010 at 13:14

Your Answer

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