Compare proportions with survey weights

[Helani Kottage]

Hi,

I used the following command to generate the output given below.
Code:
svy: tabulate Recoded c_urban_rural, tab( d_all_annual_amount) column
Here, Recoded is a nominal variables with 21 categories, c_urban_rural is a binary variable and d_all_annual_amount is a continuous variable. Following is the output table.

Now, I need to test whether the proportion of urban is equal to that of rural for each category of 'Recoded'. For example, when Recoded = gardens, the proportion of urban (0.0724) is similar or different to the proportion of rural (0.3672). I hope this is clear enough. Any support is kindly appreciated.

Thank you in advance.

[Felix Bittmann]

Here is a discussion of potential solutions: https://www.statalist.org/forums/for...726-prtest-svy

[Helani Kottage]

Thank You Felix. But, I cannot use the methods discussed in this link. In my case, The proportion is calculated using a third variable which is continuous. Therefore, I cannot find a way to use test or lincom commands. In fact, I cannot find the names of the cells because, `coeflegend' command doesn't work with the `svy:tabulate v1 v2, tab(v3)' command.
Thanks again.

[Leonardo Guizzetti]

You could do this, but why do you want to? It seems like a pretty plain conclusion that land use/categorization is different between rural and urban environments. How does a hypothesis test help?