What I understand is that in this scenario,

the data is partitioned into Z weighting classes on the basis of variables observed for respondents and nonrespondents. $n_z$ represents the sample size, $r_z$ represents respondents in weighting class Z

The estimate of response probability $\phi _z = r_z/n_z$ , the estimate of weights for responding units $w_i = r(\pi_i\phi_i)^-1/\sum_{z=1}^{r}(\pi_z\phi_z)^-1$

If this is my test dataset

  Age       n         r            mean      Sd
  20-29     23        19           219       32
  30-39     37        29           222       37
  40-49     30        17           245       43
  50-59     15        6            270       41

How are the weights $w_i$ calculated ? I know $\phi_z = {19/23, 29/37, 17/30, 6/15}$ I am struggling a bit to understand how the weights are estimated here specifically the $\pi_i$ component.

Sorry if this is a moon shot but this reflects the gap in my understanding of these concepts. Any help or suggestion is appreciated.


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.