Method deriving weight of second level for mtobit model

[Bhuban Dha]

I would like to analyse my data in the multi-level Tobit method (metobit) with applying two-level weights. In multi-level linear mixed model we get the first level weight [pweight= IPW] by calculating a propensity weight and the Stata automatically estimates the weight. "size" for cluster level (2nd level). Stata command for the linear mixed model is "mixed unemp i.year inc edu [pweight= IPW]|| year:, pwscale(size)", My data of the unemployment variable is censored in distribution. My data would better fit in metobit method. But metobit does not automatically estimate the "size" weight. Stata command for mtobit is "metobit unemp inc edu [pweight=IPW] || year:, pweight(wvar2) ll(0)". The wvar2 in pweight(wvar2) is the 2nd level weight similar to "size". I do not know how to derive the weight data from my first-level estimate. Stata manual explained the meaning of the weight but not explained the calculation method. Can you please help me by describing a) the 2nd level weight-deriving method as Stata calculates the weight "size" b) providing reading materials to calculate the weight?

[Bhuban Dha]

Clarification and follow up of my last post dated 27Nov {N=1691107]:
Based on my understanding from the multi-level mixed model manual of Stata, when users provide weight data for the first level (pweight), the Stata automatically estimates second level weight "pwscale" as "size". Technically how to calculate My understanding can be wrong. Please advise me whether it is write?. The pwscale(size) command does not work with "metobit" estimation method. I may need to estimate it and provided as pweight=clusterWt). I calculated the first level weight by transforming the inverse propensity weight of individual years into long format of data. My problem can be beyond the the Stata software related. I would greatly appreciate the help of the experts to resolve the problem.