3. (20pts) Table 2 shows the data from a study to compute the causal effect of antiretroviral therapy A on death Y in subjects infected with HIV. Individuals were classified as treated (A = 1) if they received antiretroviral therapy, and as untreated (A = 0) otherwise. Death is coded as Y = 1. The variable L represents CD4 count (1:low; 0:high). Assume that this is a conditional randomized experiment given L. This is the hint formula for these three question. L=0 L= 1 Y=1 Y=0 Y=1 Y=0 a) ELY(1)-Y(0) | L=0]= E[Y(1)-Y(0) | L=1]= A=1 20 30 108 252 A=0 40 10 24 16 bjE[Y(1)-Y(0)]= Table 2: HIV data c) E[Y(1)-Y(0) | A=1]= (a) (6pts) Compute the conditional causal risk difference, given L = 0 and L = 1. (b) (6pts) Compute the marginal causal risk difference. (c) (8pts) Can you compute the causal risk difference for those who actually received the treatment