Thanks, i have to correlate one normal distributed variable (antibody concentration) and other non-normal variable (percentage of killed bacteria), thus spearman correlation is an option because my data are not completely normal and some data is repeated (tie data). A question, i need to transform my raw data variable into ranks before the test? or the the raw data is transformed by default, if i understand well, when you say that both of your variables are at the ordinary scale that means that you are not using your raw data, your previously transformed into ranks?.
Hi there, Spearman will take care of the rank transformation for you. If you happen to have a small sample, you might want to look into Kendall Tau correlation instead: ruclips.net/video/DleIHPdDw_o/видео.html Best Regards, Björn.
Perfect as usual, thanks a lot.
Thanks for the positive feedback!
Best Regards, Björn.
Thanks, i have to correlate one normal distributed variable (antibody concentration) and other non-normal variable (percentage of killed bacteria), thus spearman correlation is an option because my data are not completely normal and some data is repeated (tie data). A question, i need to transform my raw data variable into ranks before the test? or the the raw data is transformed by default, if i understand well, when you say that both of your variables are at the ordinary scale that means that you are not using your raw data, your previously transformed into ranks?.
Hi there, Spearman will take care of the rank transformation for you. If you happen to have a small sample, you might want to look into Kendall Tau correlation instead: ruclips.net/video/DleIHPdDw_o/видео.html
Best Regards, Björn.
@@statorialsThank you