I am running a non-parametric correlational analysis, and I am not sure about if I should use Spearman or Kendall. Kendall is recommened if there is a small data set with large number of tied ranks, but what does a "large number" mean?
I have a population of about 50 individuals. In one of my datasets, about 30 % of the ranks are tied, and the tied ranks occur with the frequency of 2-3 with the same value. In the other datasets, there is a higher percent of tied ranks, maybe 60%-70%, where these tied ranks occur with a frequency from 2-25 with the same value.
Hope this makes sence, a bit complicated to explain!
Does this sound like a large number of tied ranks? How can I know?
I have a population of about 50 individuals. In one of my datasets, about 30 % of the ranks are tied, and the tied ranks occur with the frequency of 2-3 with the same value. In the other datasets, there is a higher percent of tied ranks, maybe 60%-70%, where these tied ranks occur with a frequency from 2-25 with the same value.
Hope this makes sence, a bit complicated to explain!
Does this sound like a large number of tied ranks? How can I know?
Last edited: