On the evaluation of procedural systems in laboratory experiments