I've got three groups A,B,C. Each has a different number of records and I have calculated the mean for each group. Now I want to know if these means are equal or not? What statistical test should I be doing?

Group   No.ofrec  Mean
A       100       10
B       200        5
C       300       20
  • 1
    $\begingroup$ Do you have the original data still, or are you able to calculate standard deviation? $\endgroup$
    – Glen_b
    Nov 26, 2013 at 5:45
  • 2
    $\begingroup$ These three means are not equal. $\endgroup$
    – Michael M
    Nov 26, 2013 at 6:44

1 Answer 1


Under the assumption that your data are continuous, you can use an ANOVA for this situation.

  • $\begingroup$ @GungIt seems that the data is categorical. May be ANOVA could be used if translate the number of records into the proportions. $\endgroup$
    – user10619
    Nov 28, 2013 at 16:10
  • $\begingroup$ @subhashc.davar, it is not very clear in the question, but I interpret No.ofrec as the $n$ for each group, & Mean as the mean of the response variable. W/ $n$s that large, even if the (w/i group) data are not very normal, you should be OK to run an ANOVA if the data are continuous. $\endgroup$ Nov 28, 2013 at 16:24
  • $\begingroup$ I have rechecked the dataset. Your response is valid. It is a valid case for ANOVA test $\endgroup$
    – user10619
    Dec 1, 2013 at 14:02

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.