To increase the sample size.

To prevent the carry-over effect of the first treatment.

To ensure the participants are blinded.

To reduce non-response bias.