1
$\begingroup$

I'm trying to quantify the difference between two sample means. I'm not sure if I can use the standard two sample t-test.

  • The first sample comes from the difference of two time series (inflation rate & 5Y-breakeven rate over time)

  • The second sample also comes from the difference of two time series (inflation rate & 7Y-breakeven rate over time)

I assume the two samples are independent, however, I'm not sure what I can say about the populations they come from, since they're derived from time series'.

The distributions of both samples are pretty far from normal and skewed, and I can't say much about the normality of the population.

Can I use the two- sample means test or should I investigate into non-parametric tests?

$\endgroup$
2
  • 2
    $\begingroup$ Time series over overlapping time periods might well be correlated. Switching from a t test for comparing two independent sample to a nonparametric test for doing the same does not make the issue of independence disappear. I'm not saying your two datasets are correlated--maybe so, maybe not. I guess my first move would be to investigate independence. $\endgroup$
    – BruceET
    Sep 14 at 1:58
  • $\begingroup$ How can I reasonably go about exploring independence. F-statistic from regressing the samples to one another? What test of mean comparison can I do otherwise? Matched pairs test? $\endgroup$
    – somnaik09
    Sep 16 at 20:13

Your Answer

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

Browse other questions tagged or ask your own question.