The analysis described in Problem 1 for the posture measurement data on Shoulder Flexion (SF) may be criticized because information is lost when the 4 observations for a given subject on a given day are combined into an average score rather than being treated individually in the analysis. The data set shown below allows for an analysis t considers all 12 observations per subject. In analyzing this data set, we assume that the 4 observations for a given subject on a given day are true replicates (i.e., we assume that neither the time of day observed nor the rater used, factors that combine to provide the 4 observations for a given subject, is an important factor for predicting SF response).

a. How should the subject-specific scalar model for Problem 1 be modified to consider the data layout below? (You will need to add a second random effect to the model that reflects the interaction of Day with Subjects.)
b. Use the computer information provided below to test whether there is a significant main effect of the factor Day. (See the ANOVA table provided below.) What do you conclude?
c. Use the computer output provided below to test whether there is a significant interaction effect between Subjects and Day. If such a test is nonsignificant, why might you be concerned regarding the model that has been fit, and how might you redo the analysis?

27 2 2 4 5 5 19 24 6 27 06 22 96-s 10 18 26 20 4 1 7 4 8 45 10 30 22 2 3540 88369 20 7 30-42 17 19 9 6 2 3 6 5 2 14 30 30 10 26 8 2 52259 24 2 32 34 day-07-27 22 12 46 73 27 2 7 2 7 7 3 2 0 2 3 6 52 9-26 2 8 2 7 2 1 2 8 3 6 8-20 2 5-1 12 25 1 860 50 7 36 20 20 G 505 27 38 2 5722480413 26 2 4 8 7 19 20 28 5 12-3-1-22 the ay-17-36 2 0 5 9 46 9 3 17 2 Sh-M 5 8 6 8 7 8 7 4 2 34 27 20 15 18 10 1 2 24 8317 2 4 81 2 0832 50 29 7 6 8 2 4 8772555 20 28 33 2 le 2 2 2 2 2 2 2n 890E 234567 89