Using svyset and switching between two datasets with two-stage design

[Daniel Jensen]

I have two datasets of a sample that were created using Probability Proportional to Size (PPS) sampling in a two-stage design with strata. The sample is of childcare centers:

1. A sample of childcare centers was selected with PPS, stratified by provider model
2. Within each center, 1-2 classrooms were randomly selected with equal probabilities

I have a dataset of director surveys (one per center) and a dataset of caregiver surveys (1-2 per center). Since PPS was used, each dataset has probability weights, so I am using svyset and the svy prefix for commands.

In the do file I'm working on, I have to frequently switch between the two datasets. In both datasets, the weight variable has the same name, although the value differs. For example, it might look like this in the director survey dataset:

Provider	Center_ID	Weight
1	101	7.6014171
1	102	4.0412597
2	201	2.6494566
3	301	3.7340294
3	302	9.8233698

And this in the caregiver dataset:

Provider	Center_ID	Clasroom_ID	Weight
1	101	10101	17.422834
1	101	10102	17.422834
1	102	10201	11.333779
2	201	20101	6.7823697
3	301	30101	10.430736
3	301	30102	10.430736
3	302	30201	21.57423

I set up the weights like this:

svyset Centro_ID [pw=Weight], str(Provider)

I understand that this command is correct for both datasets, although including Center_ID in the command for the director survey dataset is unnecessary.

Do I need to reset svyset every time I switch between datasets, or is it sufficient to just do this at the begninng of the do file?