Logistic regression for pesticide toxicity data. A environmental agency is testing the effects of a pesticide that can cause acute poisoning in bees, the world's most important pollinator of food crops. The environmental agency collects data on exposure to dierent levels of the pesticide in parts per million (ppm). The agency also identies collapsed beehives, which they expect could be due to acute pesticide poisoning. In the data they collect, each observation is pair (mi, 3),), where 9:; represents the dosage of the pollutant and :9, represents whether or not the hive survived. Take y,- = 1 means that the beehive has collapsed from poisoning and y, = 0 means the beehive survived. The agency collects data at several different sites, each of which was exposed to a different dosages. The resulting data can be seen below: it