Creating weights
Weighting data can be done for a number of different reasons, but often to account for sampling bias.
Weighting can vary from the simple to the complex. You should take care to check weights carefully if you're:
- weighting sub-samples differently from each other
- recomputing weights each time period (wave)
- changing weighting definitions over time
Simple weights can be constructed in the web app, but combining weights will require scripting.
Note: pay attention to whether the weights are being derived automatically or if they are being hard-coded ('materialised') in the data. Weighting can be a tricky part of tracking studies, as it often relies on data collection to be complete before it is implemented.
In the web app
Using R
Crunch Automation