My (previously unexamined) assumption was that the slower time exactly offsets the lower acceleration in terms of speed gained, such that in the reference frame of the accelerating ship, experiments would continue to show 1G acceleration.