02-16-2019 12:17 PM
As explained here: https://www.statisticshowto.datasciencecentral.com/kendalls-tau/, Kendall's Tau, or Kendall Rank Correlation Coefficient, is a non-parametric measure of relationships between columns of ranked data. The Tau correlation coefficient returns a value of 0 to 1, where:
A quirk of this test is that it can also produce negative values (i.e. from -1 to 0). Unlike a linear graph, a negative relationship doesn’t mean much with ranked columns (other than you perhaps switched the columns around), so just remove the negative sign when you’re interpreting Tau.
Several version’s of Tau exist.
Most statistical packages have Tau-B built in, but you can use the following formula to calculate it by hand:
Kendall’s Tau = (C – D / C + D)
Pay attention to the setup of this one, you need to make sure that you start with a sorted, ranked column of values. The measure in Power BI looks like this:
Kendall's Tau = VAR __table = 'Data' VAR __table1 = ADDCOLUMNS(__table,"__Concordant",COUNTROWS(FILTER(__table,[Index]>EARLIER([Index])&&[Interviewer2]>EARLIER([Interviewer2])))) VAR __table2 = ADDCOLUMNS(__table1,"__Discordant",COUNTROWS(FILTER(__table,[Index]>EARLIER([Index])&&[Interviewer2]<EARLIER([Interviewer2])))) VAR __C = SUMX(__table2,[__Concordant]) VAR __D = SUMX(__table2,[__Discordant]) RETURN ABS(DIVIDE(__C - __D , __C + __D,0))
This is really fantastic!
Do you have any work around the Mann Kendall P value, I want to calcualte the Mann Kendall P value by multiple groups like (Region, Country, Department, Product).