Why are the standard errors/confidence intervals in my AskCHIS estimates different compared to those I find using the Public Use Files?
There could be a few reasons for this difference. First, the variable you’re using in the Public Use Files may be a different variable than the one being used by AskCHIS. Second, AskCHIS does not use the replication method; it employs the Taylor series linearization method for standard error calculation. The main reason AskCHIS uses the Taylor series method is because it is computationally faster and easier on our servers. This may produce differences in standard errors or confidence intervals.
Related Questions
- The user’s manual states that in order to compare estimates we need to use correct standard errors, which can be generated using replicate weights. What and where are the replicate weights and how do I use them?
- Why are the standard errors/confidence intervals in my AskCHIS estimates different compared to those I find using the Public Use Files?
- How does one obtain standard errors and t-ratios for the parameter estimates in an interactive model?
