This is an idealized response to a stimulant (the stimulant in our case is exposure to an air pollutant, but the D-R curve can be applied to any type of stimulant, such as the presence of a gas that has an odor, if the concentration is high enough). The "binary" nature means that at some single threshold intensity, there is no response below the threshold and there is 100% response above the threshold (i.e., all or nothing...only two possible states for the whole population). It's easy to set up pollution regulations for this type of response, since the maximum safe level would be just below the threshold level.

Real-world response look more like this sigmoid ("S-shaped") curve. In this case, not everyone responds to the same intensity of the stimulus, depending on the sensitivity of different individuals. Subjects exhibiting a response below a threshold (not a true, binary-type threshold; this value usually corresponds to a 50% response rate) are considered "sensitive", while those who do not exhibit a response until the stimulus intensity is greater than the threshold are considered "insensitive".

Response curves like this make regulation difficult, because now we must perform a risk assessment, where we decide how much of the population we accept as being potentially affected by pollution levels near our regulatory threshold (0%? 25%? 50%?). Particular bad are the situations where the slanted portion of the curve is closer to horizontal than in this diagram, which draws out the sensitive/insensitive differences over a very large range of stimulus intensities (so the curve is "flatter" and less like a binary response curve). This means that minimizing the exposure risk might mean setting the maximum pollution levels at unrealistically low values, where the cost of control could be prohibitive or excessive.