If you have data that was collected with the intent all observations be used in analysis because each is representative, even in absence, how can you force Stata to incorporate the empty observations?

For example,
Code:
svy: regress score female age
is producing an error
"Note: Missing standard errors because of stratum with single sampling unit."
But what it is doing is looking at the 5,000 observations that are 'complete' with entries for both female and age variables and ignoring the 125 observations that are empty except the strata. But the point is that the stratum would NOT be missing if the empty observations were not being dropped - they are assigned to the strata and weights. The options for
Code:
singleunit(scaled) or singleunit(centered)
are known but are not right for this because using that option still ignore the 125 observation that were meant for use. The data is meant to be with all observations. In other software there is a way to make a command stop dropping missiving or empty values. Can Stata do this too?