You implement that like any other between-subjects condition, i.e., set up for <expt>s (one per condition), or use the <variables> element (not generally recommended), or use conditional <include> elements (one per condition). If you're in the lab, you obviously have direct control over how you assign participants to conditions. If on the web, you have several options:
- generate a random id for assignment (without replacement)
- use sequential id generation, or
- pass the id in via a query parameter if you want direct control.