In user research, between-groups designs reduce learning effects; repeated-measures designs require fewer participants and minimize the random noise.