1
$\begingroup$

I have a set of data from two nights worth of monitoring.

The monitoring picks up how busy a server is in terms of the number of threads processing.

A change has been made and we want to analyse if this has had an impact of reducing the threads.

How can I compare the two data sets given that the load on the server is going to be a variable (but comparable).

I have been thinking of ANOVA, is this a good method? Any others people can suggest?

The data seems to display a positively skewed normal distribution. Is ANOVA still valid?

As an example (dummy data)

                          Thread Count
        Night Before Change      Night After Change
00:00          341                    223
00:05          365                    321
00:10          465                    445
.
.
.
08:50          512                    243
08:55          314                    556
09:00          256                    354
Improve this question
$\endgroup$
1
  • $\begingroup$ If you do a paired t-test, it will be the normality of the pair differences that matter, rather than the raw values. $\endgroup$
    – Glen_b
    Commented Jul 9, 2014 at 9:12

1 Answer 1

Reset to default
0
$\begingroup$

I assume you want to compare the results of Night Before Change vs. Night After Change to see whether there is an significant difference or not. If this is the case you need to use t-test.

The t-test is used when determining if two averages or means are the same or different. The ANOVA is preferred if comparing three or more averages or means.

Improve this answer
$\endgroup$

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.