Prevalence , Risk factors



I’m new to the NHIS data and currently working on

  1. Prevlance of a condition

  2. Risk facors for the outcome of mortality among these subjects with this condition

My questions:

  1. As for the prevalence, should I use the person weight as it is adjusted for age, race/ethnicity and sex

  2. Which weight should I use for risk factor analysis

Thanks for your time,



The general rule of thumb is to use the sampling weight associated with the level of analysis. That is, if you are performing person level analysis (e.g. each observation is an individual person) you should use the person level weights. On the other hand, if you are performing household level analysis (e.g. each observation is a household) you should use the household level weights. So, in both of your cases, it sounds like you should be using person level sampling weights. More information and links to existing documentation about sampling weights can be found on this page.