Parameter Estimates

The Cubic model states that the expected proportion of infected e-mails is equal to -0.123 + 0.088*hours since detection - 0.004*hours since detection squared + 0.00004399*hours since detection cubed. The positive value for the linear term suggests that the dependent variable increases initially with time. The negative value for the squared term says that past a certain point, the number of infected e-mails decreases. The positive value for the cubic term actually suggests that the number rises again!