Im quite confuse if I should use a weighted or unweighted survey data when doing categorical data analysis such as Chi-square, G2, concordant or discordant analysis, etc.

Basically, these analyses use the numbers in a cross tab table. My take is to use weighted data because its assumed its a better approximate of the population value. However, using the weighted (and expanded) data may influence the p-value making most of the results significant.

Any idea is greatly appreciated. Thank you.

  • 1
    $\begingroup$ It depends on the survey and the meanings of the weights. Could you explain? $\endgroup$ – whuber Oct 7 at 20:04
  • $\begingroup$ It is a community survey and the survey weights estimates the number of people in a population. $\endgroup$ – dixi Oct 7 at 20:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.