In DataDynamic you can create weight variables that can be used to make datasets more representable for certain populations. If you would like to know more about dataset weighting please have a look here.

DataDynamic supports creating weights using standard weighting (target percentages are known for all possible combinations of all dimensions) and using RIM weighting (also known as iterative raking or iterative proportional fitting).

To create a new weight variable go to the "Editing" tab in DataDynamic, select the dimension variables to use as a basis for the new weight variable and click on "Create Weight" in the top menu. You now should see the following screen:

You now have the following options:

  • You can enter the name for the new weight variable.
  • You can enter the description for the new weight variable.
  • By default RIM weighting methodology is selected, if you want to use standard weighting uncheck the checkbox.
  • When using standard weighting you can choose to create a sample correction rather then a weight. The difference is that with a standard weight you change the distribution between groups keeping the total N the same. A sample correction is created by giving target N rather then target % and as such will change the total N based on the required N per group defined.
  • In the grid you will find the list of categories for which a target % is required, the list is different based on the methodology chosen:
    • For RIM weighting the list will give each individual category of each dimension with the current distribution. The target percentages are required to add up to 100% for each dimension present.
    • For standard weighting the list will contain an entry for each category combination across all selected dimensions. The target percentage need to add up to 100% across all entries in the list.
  • After all target percentages have been entered you can click on "Create" to create the weight variable as specified.