Svy commands for Kappa and Pearson's Correlation Coefficient

[Elizabeth Duthie]

Hello,
I am new to using survey commands. I have a survey data set and need to calculate Kappa statistics and Pearson correlation coefficients. Is there a way to do this using survey commands? If not, how should I approach this analysis?
Thank you kindly,
Elizabeth

[Rich Goldstein]

for Pearson correlation, you can always take advantage of the fact that the correlation is equal to the square root of R-squared from an OLS regression (and the sign is the same sign as on the coefficient); thus, you can use -svy estimate- with the command -regress-

[daniel klein]

kappaetc (SSC, SJ) estimates weighted Cohen's Kappa (and other agreement coefficients); you would need to rescale your sampling weights so they sum to N. You could probably get standard errors by bootstrapping.

If you only have two raters (variables), you could alternatively use tabulate with aweights and calculate Kappa based on the weighted cell frequencies. Again, you would need to think about getting the standard error.

Best
Daniel