Suppose you are analyzing a two-way model with main effects only (yink = μ + αi + βj + oijk) with a = 4 levels for factor A and b = 3 levels for factor B and with r replicates for each treatment combination. However, by the time you receive the data you notice that all the observations {y11k : 1 ≤ k ≤ r} and {y12k : 1 ≤ k ≤ r} have been lost. Compute the least square estimates for the contrasts α1 - α2 and β1 - (β2 + β3)/2.