Please enable JavaScript.
Coggle requires JavaScript to display documents.
Correlation & Agreement - Coggle Diagram
Correlation & Agreement
-
Agreement
-
-
Generally this measures inter-rate reliability. Two different measuring schemes (could be a person or a test) measuring the same type of outcome. Common example is body fat %. DEXA vs Calipers - how much agreement.
Correlation Tests
Spearman's Rank
-
-
-
Measures the strength and direction of the monotonic relationship between two variables. two variables only
Usually first check if each variable is normally distributed. If not then use Spearmans. If normal Pearsons could apply.
Pearson's Correlation
-
-
-
-
-
Is there a linear relationship between variable 1 and 2? Null hypothesis says no. Alt hypothesis says yes. p-value will tell us if we should reject the null. If p-value < 0.05 we reject.
0 may be that there is no linear relationship but could still be coreelated or realtionship in other ways.
Kendall's Rank/Tau
-
-
Manual figuring out
-
Concordant pairs are the number of observed ranks below a particular rank which are larger in value than a particular rank
-
-
-
Agreement Tests
Cohen's Kappa
-
-
Helps describe actual agreement of tests vs random agreement (broken clock and working clock will agree at least 2x a day).
Interpretation
-
-
x <0 = less agreement than what is expected at random
x = 0 agreement that is entirely due to due to random
x > 0 more agreement than just random
x = 1 perfect agreement
0.9 = almost perfect
0.8 - 0.9 = strong
0.6 - 0.79 = Moderate
0.4 - 0.59 = weak
0.21 - 0.39 = minimal
0 - 0.2 = none
-
-
-
-
