s**********y 发帖数: 38 | 1 I'd like to know the agreement of 1st rating and 2nd rating.
the p value of McNemar's Test is 0.0455, but the Kappa is 0.6602.
How should I explain? Thanks.
Kappa is 0.66. Does that mean moderate agreement? Should I only use kappa in stead of McNemar's Test to explain agreement? Thanks.
1st rating 2nd rating
Frequency N Y Total
N 67 12 79
Y 4 26 30
Total 71 38 109
Frequency Missing = 134
McNemar's Test
Statistic (S) 4.0000
DF 1
Pr > S 0.0455
Simple Kappa Coefficient
Kappa 0.6602
ASE 0.0768
95% Lower Conf Limit 0.5097
95% Upper Conf Limit 0.8107 | s**********y 发帖数: 38 | 2 McNemar's null hypothesis is
The null hypothesis of marginal homogeneity states that the two marginal
probabilities for each outcome are the same
for one variable: p value of McNemar's=0.2, which means the two ratings are
the same. But the kappa is 0.1966, which is poor agreement.
Isn't this conflict?
Hope someone can help me with this. |
|