Randomized Block Design In chapter on paired t-tests, we learned to “match” subjects on variables that: influence performance but are not of interest. Matching gives a more sensitive test of H0 because it removes sources of variance that inflate 2. Lecture 17
Randomized Block Design In the analysis of variance, the matched subjects design is called the Randomized Block Design. subjects are first put into blocks a block is a group matched on some variable subjects in a block are then randomly assigned to treatments for p treatments, you need p subjects per block Lecture 17
Question: does SSB come from SSE or from SST? Sums of squares In the RBD, we compute SST as before. Compute SSB (SS for Blocks) analogously: Compute deviations of block means from grand mean. Square deviations, then add them up. Question: does SSB come from SSE or from SST? Lecture 17
Where does SSB come from? SST SSB SSTotal SSE Residual SSE Lecture 17
Conceptual Formulas SST = Σb(XTi – XG)2 p-1 SSB = Σp(XBi – XG)2 b-1 SSTotal = Σ(Xi – XG)2 n-1 SSE = SSTotal – SST – SSB (b-1)(p-1) = n-b-p+1 MST = SST/(p-1) MSB = SSB/(b-1) MSE = SSE/(b-1)(p-1) = SSE/(n-b-p+1) Lecture 17
Summary table Source df SS MS F Treat p-1 SST SST/(p-1) MST/MSE Blocks b-1 SSB SSB/(b-1) MSB/MSE Error n-p-b+1 SSE SSE/(n-b-p+1) Total n-1 SSTotal Lecture 17
Computational Formulas CM = (ΣX)2/n SSTotal = ΣX2 – CM SST = ΣTi2/b – CM SSB = ΣBi2/p – CM SSE = SSTotal – SST – SSB p = # of samples b = # of blocks Ti = Total for ith treatment Bi = Total for ith block Lecture 17
Randomized Block Design – Example 1 H0: 1 = 2 = 3 HA: At least two differ significantly Statistical test: F = MST/MSE Rej. region: Fobt > F(2, 8, .05) = 4.46 Lecture 17
Randomized Block Design – Example 1 CM = 104834.4 SSTotal = ΣX2 – CM = 782 + 812 + … + 942 – 104834.4 = 105198 – 104834.4 = 363.6 Lecture 17
Randomized Block Design – Example 1 SST = Σ(Ti2)/b – CM = 4012/5 + 4212/5 + 4322/5 – 104834.4 = 104933.2 – 104834.4 = 98.8 SSB = 2442/3 + … + 2712/3 – 104834.4 = 105075.33 – 104834.4 = 240.93 Note how much SSB is removing from SSE here. Lecture 17
Randomized Block Design – Example 1 SSE = SSTotal – SST – SSB = 363.6 – 98.8 – 240.93 = 23.87 Lecture 17
Randomized Block Design – Example 1 Source df SS MS F Treat 2 98.8 49.4 16.55 Blocks 4 240.93 60.23 20.18 Error 8 23.87 2.98 Total 14 363.6 Lecture 17
Randomized Block Design – Example 1 H0: SN = GO HA: SN ≠ GO (Note: this is a post-hoc test. We’ll do N-K.) Statistical test: Q = Xi – Xj √MSE/n Lecture 17
Randomized Block Design – Example 1 Rank order sample means: Sleepy Sneezy Grumpy Dopey Goofy 90.3 86 81.3 80.6 79.67 Qcrit = Q(4, 8, .05) = 4.53 Lecture 17
Randomized Block Design – Example 1 Qobt: 86 – 79.67 = 6.33 = 6.35 √(2.984)/3 0.997 Reject H0. Goofy & Sneezy differ significantly. Lecture 17