I am doing some mixed effects modeling where we would like to test the hypothesis that the dependent variable (Dependent.var) as a function of time depends on the value of the covariate 6 months to a year earlier. The difficulty is that not all subjects have measurements every 6 months (as shown below for the 1st three subjects). We would like to derive the new datasets for the 6 month lag, 1 year lag and so on using efficient code. For the 6 month lag the Time zero row could get thrown out and the covariate column would move one cell down (and the covariate value at year 3 would be taken away as well). Is there some easy way to make these datasets Subj.ID TIME Dependent.Var Covariate 1 1 0.0 -0.02720222 0.05055147 2 1 0.5 -0.95878074 0.07402854 3 1 1.0 0.05178116 0.15676033 4 1 1.5 -0.90101795 0.05398925 5 1 2.0 0.05873045 0.35908569 6 1 2.5 -0.20770872 0.37844074 7 1 3.0 0.60331684 -1.05727034 8 2 0.0 -0.35234081 -0.43684877 9 2 0.5 -3.41076246 0.34309430 10 2 1.0 3.12190327 0.83680466 11 2 1.5 -0.04691748 0.59614786 12 2 2.0 -0.59465266 -0.09676054 13 2 2.5 -2.30089870 -0.68813666 14 2 3.0 0.07287830 -0.72876164 15 3 0.0 -0.94644384 -1.62981406 16 3 0.5 -1.06351763 0.19774138 17 3 1.0 -0.46948937 -0.59600225 18 3 1.5 1.80077781 0.46992426 19 3 2.0 -0.27466530 1.22268195 20 3 2.5 -0.87696707 -2.29894151 21 3 3.0 0.45811417 1.76158843 [[alternative HTML version deleted]]