Weights

When we have samples, we need to make sure that these samples are representative of the total population – to do this we may need to establish weights

For details of weights see:

James R. Chromy and Savitri Abeyasekera, "Statistical analysis of survey data" [Chromy2005]   formerly available from http://www.cpc.unc.edu/projects/addhealth/data/guides/weight1.pdf

---

[Chromy2005] James R. Chromy and Savitri Abeyasekera, ‘Chapter XIX: Statistical analysis of survey data’, in Household Sample Surveys in Developing and Transition Countries, New York,  N.Y.: United Nations, 2005 [Online]. Formerly available from http://www.cpc.unc.edu/projects/addhealth/data/guides/weight1.pdf

Transcript

Now, we might need to think about weights.  Why? Because we could only sample each of the populations that we wanted to some level, and therefore when we get our results from analyzing that subset, we now need to weight that to be able to fit it to our model over the entire set of data. So if these people represent ten percent of the data, then, of course, it should have ten percent of an effect upon our overall conclusions of our model. And Chromy and Abeyasekera have written this book called "Statistical analysis of survey data" that you may find interesting for more discussion about weights.