You need three variables to make an analysis. Variables newprov and race have been used before, so you should know what those variables are called. If you do not know a variable name, an actual housing rent in each Province for example, then use lookfor option.
lookfor rent
163. rent_act int %9.0g 9b:monthly rent (actual) 164. rent_mkt int %9.0g 9c:monthly rent (mkt)
You want to use actual rent instead of market rent price, so use rent_act.
Also, it is worth checking to make sure the variable rent_act does not have any negative value, which is an invalid answer. In general, when you want to check an existence of an invalid number, use codebook varname and see range of the value. If there's a negative number that does not make sense to you, check the survey. In this case:
codebook rent_act
rent_act --------------------------------------------- 9b:monthly rent (actual)
type: numeric (int)
range: [-4,2200] units: 1
unique values: 90 coded missing: 0 / 5284
mean: 25.6673
std. dev: 158.387
percentiles: 10% 25% 50% 75% 90%
-2 -2 -2 -2 30
Since there are negative value of the rent which does not make sense, you should exclude invalid answers by using option if rent_act>0.
tab newprov race, sum(rent_act) mean, if rent_act>0