2
$\begingroup$

I have two sets of data and a metric associated with each of them, I want to see if one category of data (e.g. males) have a tendency to have ahigher metric associated with it than the other category (e.g. females). The metric is continuous, however first category has over 2 million samples and the other only 200, is there a statistical test I can use?

Improve this question
$\endgroup$
2
  • $\begingroup$ The way this question is phrased and the tags you have applied suggest there is something special about the numbers 2 million and 200 in the context of a t-test for you. What would it be? In what way are you supposing this question differs from one where the numbers might instead be, say, 10 and 20? $\endgroup$
    – whuber ♦
    Commented Aug 10, 2015 at 14:24
  • $\begingroup$ It is mainly due to articles like these: blog.minitab.com/blog/statistics-and-quality-data-analysis/… where they emphasize the t-test giving huge significance to even the smallest of differences when the sample is too big. Also, I am not sure if such uneven sample size poses a problem as well. $\endgroup$
    – zukkmet
    Commented Aug 11, 2015 at 12:17

0

Reset to default

Your Answer

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