Weigthed Mean

[Mila]

Also btw....the misisng values issue is clearly delineated in the question "Grand Advisor" @Reeza😛

[Rick_SAS · Accepted solution]

There are two ideas being discussed here.

1. The OP asked about a "weighted mean" of VARIABLES, but it seems what she really wants is a linear combination of six variables, where the coefficients change if one or more variables have missing values.

2. There were a few comments about using PROC EXPAND, which computes a moving average of OBSERVATIONS. I see no evidence that this is what the OP wants.

The reason I think that the OP wants (1) is because she mentions principal components.

Mila: The SUM function handles missing values, whereas the '+' operator propagates missing values. That is why someone suggested using SUM. 

If you want to standardize the weights, just divide the sum by the sum of the weights, For example

y = sum(A1*x1, A2*x2, ..., A6*x6) / sum(A1, A2, ..., A6);

See the article "Compute a weighted mean in SAS."

[Mila]

Thank you very much, this helps clear things up tremendously - best wishes