I want to use Spearman's rho test on proportional data for one variable and continuous data for the second variable (e.g. % survival of fish related to concentration of arsenic)? Further I want to know that "Does it violate an assumption of the Spearman's rho test to use on data that when plotted have no monotonic relationship?"

  • $\begingroup$ When you say "have no monotonic relationship" do you mean you can't discern any relationship, or you can but it's clearly not monotonic. $\endgroup$ – Glen_b Apr 23 '16 at 7:41
  • 1
    $\begingroup$ Say rather than Spearman's rank correlation quantifies how far you have a monotonic relationship. If it were an assumption behind the method that relationships are monotonic, then any failure to get 1 or $-$1 would be regarded as a failure of the method, but not so. The ranks don't know where they came from: the ranks for proportions would be just the same as for absolute amounts before you divide by a total. $\endgroup$ – Nick Cox Apr 23 '16 at 13:47
  • $\begingroup$ @statsnorth your 2nd variable is well defined.The definition of Ist variable may facilitate the response from the community. $\endgroup$ – Subhash C. Davar Apr 23 '16 at 14:27
  • $\begingroup$ thank-you. the relationship appears visually on a plot as a straight vertical line because the % survival data is mostly 100% and then the concentration varies small changes in water quality concentration (eg ug/L of antimony), so the data appears as a vertical line. That seems like a non-monotonic relationship that would violate the assumption of running the Spearman test. $\endgroup$ – Statsnorth Apr 25 '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.