2
$\begingroup$

I am trying to determine if a certain conversion on my site is statistically significant. I remembered doing this type of stuff in school but I can't seem to remember how to do it now.

For 1st set: n = 7297 and conversion was 2.618% For 2nd set: n = 6107 and conversion was 2.669%

Any tips on how to do this?

$\endgroup$
  • 4
    $\begingroup$ With large enough samples, even trivial differences may be statistically significant. Is a difference in the third significant figure of any practical importance (how much difference could it make to you, really)? If there's no practical difference, there's no point testing for statistical significance. $\endgroup$ – Glen_b -Reinstate Monica Oct 12 '13 at 0:13
3
$\begingroup$

Percents are a proportion. The traditional way to test differences between proportions is the chi-square test. Based on the information you have given me (7106 and 191 [2.62%] in one half and 5944 and 163 [2.67%] in the other), the chi-square test results in a non-significant value of .88 (p value). Your proportions are 2.67 and 2.62, so it is no surprise that these are not statistically significant, despite your large sample.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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