Specific tasks give the participant a defined goal and scenario (e.g., 'find and purchase a blue shirt in size medium'), which helps measure whether users can complete a known workflow. Open-ended tasks give the participant broader freedom (e.g., 'explore the site and tell us what stands out'), which helps surface unscripted usability issues and first impressions. Researchers choose between the two, or mix them, based on the study's research questions.