WLS = reg + aweights?

[Christian Danner]

I would like to understand whether I apply the regression technique "weighted least squares (WLS)" by adding weights to the standard "reg" command.

My dependent variable is price per kg (price_kg). My observations are orders, each order is defined by the volume ordered (kg) for a specific price per kg. I attach weights to each data point to give prices a weighting that corresponds to the quantity (kg) ordered for that price. I include several explanatory variables driving price changes (x1, x2, x3). Standard errors are heteroscedasticity-robust.
I run the following regression:

reg price_kg x1 x2 x3 aweight = kg, robust

1) Is this command an application of the regression technique "weighted least squares (WLS)" or if not what is the difference?
2) Do you agree that the command is right given the context provided?

Thanks for your response and help in advance!!!

Christian

[Carlo Lazzaro]

Christian.
see the detailed explanation under Example 8, -regress- entry, Stata .pdsf manual.