I have a dataset of patients (n=730) who attended an outreach service and completed a survey, looking at ownership of assets and dwelling characteristics. If we would now like to follow-up a sub-set from the sample, to see if the self-reported responses are identical to what we observe at the household level, I have been told the kappa inter-rater agreement is a good measure. My main indicator is disability... i.e. this is descriptive analysis to analyse poverty, wealth and disability relationship. My question is.. How do I calculate the sub-sample size from original sample?