Would the criteria for admitting individuals into the analysis tend to increase or decrease larynx cancer rates among the exposed? Were these criteria relatively strict or liberal?