How to weight data?

Hi! I’m a graduate student working on a thesis. This is my first time working on research and my first time encountering data weighing. I’ve read the page regarding how to weight data, and I’ve gone through the forum looking for a basic explanation. It’s still really confusing, and I want to make sure I weight the data correctly. Can anyone offer a simplified explanation of how to weight the NHIS data prior to 2018 when taken over a 16 year period using IBM SPSS?

Thank you so much in advance!

I am not able to provide direct guidance about how to incorporate the weights into your research without knowing additional information about your intended analysis, specifically what samples and variables you intend to use in your work. Generally speaking, to weigh the IPUMS NHIS data in 2018-prior samples, you will most likely want to use PERWEIGHT or SAMPWEIGHT when conducting a person-level analysis (e.g., association between age and self-reported health) and HHWEIGHT when conducting a household-level analysis (e.g., estimating number of families in the household). To figure out which weights to use for your analysis, you should review the Weights tab for each variable that you plan to include in your analysis (e.g. see the Weights tab for AGE). Researchers should use the weight for the most restrictive subsample in their analysis. This means using SAMPWEIGHT when conducting a person-level analysis that includes any variable that requires SAMPWEIGHT and only using PERWEIGHT if all of the variables in your analysis state that PERWEIGHT is the correct weight to use. This is because variables that use SAMPWEIGHT correspond to questions that are only asked of sample persons (one randomly selected person per household) rather than all persons.

You may find this description of how to incorporate weights in your analysis or this description of the SPSS “WEIGHT” command helpful. Please note that I am not able to provide analytical or syntax support, but I can provide further guidance on which weight seems most appropriate for your analysis.

I hope this helps. Please let me know if you have any questions.