I have done a cluster analysis for customer data using compareGroups, but I don't understand what p.overall means. Does it show each that variable in clusters are the same, or the value each variable from clusters are different?

Here are some of the results

data1 = data.frame(x,cluster = g1$cluster)
group<-compareGroups(cluster~.,data= data1)

--------Summary descriptives table by 'cluster'---------

                  1             2            3            4             5       p.overall 
                N=233          N=9          N=37         N=94         N=67                

¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ Channel:

1            229 (98.3%)    1 (11.1%)    2 (5.41%)    6 (6.38%)    60 (89.6%)     <0.001     
Fresh        11236 (9863) 20696 (15256) 8817 (18414) 7815 (7349)  21120 (17473)   <0.001
Milk         2001 (1344)  37557 (20033) 16097 (7131) 7545 (3067)   6587 (3667)    <0.001  



1 Answer 1


Here, p.overall simply lists the p-values for t-tests (or Mann-Whitney U tests, if your data are non-normal) performed between each of your clusters. Looking at the documentation for compareGroups, we see that it apparently corrects for multiple comparisons when necessary.

In this case, if you choose your significance level to be alpha = 0.05, then your reported p-values would indicate that the groups differ from one another in both the "Fresh" and "Milk" variables.

This question is not really appropriate for StackOverflow, but I figured I'd help you anyway. If you have other questions about statistical analysis, try: https://stats.stackexchange.com/

  • $\begingroup$ Ouuuh... i see thank u for the answer, the reason i asked here... Because kinda difficult to get an answer there. $\endgroup$ Apr 21, 2018 at 1:33

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.