You can but you need to insert the individual variables from the group into the metrics. From there, you can weigh it.

Hi jbk!

Thanks for your response.

But I do not understand that completely: When I insert the individual variables from the group into the metrics, then I cannot display an overall mean. And this is what I need: an overall mean for the weighted variables.

Hi ​@HenrikeS ,

You can’t apply weights to measure groups. I think what you want is create a custom metrics (on which weights can be applied in the custom metric definition instead of inside the widget).

@vgayraud

I think this method leads to wrong results when there are missings in one or both of the variables.
Somewhere I have to work in logics:

• IF NPS1 has a value AND NPS2 has no value → set NPS Custom Metric to weighted NPS1
• IF NPS1 has no value AND NPS2 has a value → set NPS Custom Metric to weighted NPS2
• IF NPS1 has a value AND NPS2 has a value → set NPS Custom Metric to (weighted NPS1 + weighted NPS2)/2

Unfortunately I cannot see how to do this.

Absolutely, you’ll get the mean of the weighted means. If statement 1 has 1 answer and statement 2 had 10 answers, the 1 answer from statement 1 will weight as much as the 10 from statement 2.

One way to go around this, although it’ll require manual edits in your dataset if you already collected data, is to use scoring in your survey flow, and pipe in the weighted mean of the scoring category at the survey level.

Something like this:

I see. But indeed, I already collected data.

It won’t help you for your current project, but there is a product idea on this that you can upvote on Evolve. I can’t link directly, but do a search on “weighting measure group” to find it.

For others who may be interested in a similar solution search on Evolve using EV-4077. Tip: type the id (EV-4077) into the tiny search bar next to ‘Sync now’ for a quicker result. Thanks ​@vgayraud for suggesting this!