I think Kendall's Tau correlation for a dataset is exactly the same as Spearman (or Pearson !) - only difference is the p values are not exactly the same ... ?
Hi, Thank you for the video, can you please suggest something? I have interviewed 6 people on likert scale about the use of washing powder giving them different situations. The sample size is very small and the questions asked from each person are 120. Please suggest me what kind of tests i should use? to compare all the interviewees and the questions too which i suppose do not relate with each othe and can not be correlate. Thank you
kendalls tau-a is used when there is a large number of tied scores and b is used when there is not a large number of tied scores. C I have never heard of sorry.
viniciusfdavid Search for information on "Kendall's Coefficient of Concordance." It is distinct from Kendall's Tau, and I believe it can be used to produce one measure of agreement when there are 3 or more raters.
Todd Grande Thanks. I found some things about it, but it`s a little strange. I have 4 raters with high kendall`s tau between them (all pairs above 0,8), but the coefficient is very low (0,03). I guess Im doing something wrong, I will try to figure it out.
+Todd Grande Answering myself, when calculating Kendall`s tau in SPSS you have to "restructure" the data, raters at the lines and patients at columns. That makes sense, after all we are comparing the ranks of the patients acros raters. Thanks.