a. Scatterplot Res vs. Fit for the complete bee sting data, including the outlier. (If yo did the decomposition as part of Exericse D. 14, you already have the residuals. To fitted values, note that for a CB design, Fit = Block average + Treatment effect.) Once you have the plot, circle the two points for block 9, which contains the outlier, and 5. Res vs. Fit #1 get fit the data draw a "balloon" around the other points, as in Example 7.21. b. Pick one: Apart from the two outlying points: (1) The plot shows about what you would expect if the model and Fisher assumptions are appropriate. (II) The plot shows a curved old lationship between residuals and fitted values, suggesting that the model does not larger fitted values, suggesting that the assumption of same SDs is not appropriate for the in the present scale. (III) The plot is fatter toward the right : larger residuals tend to go with data in the present scale. (IV) The plot shows a balloon with a slight tilt. Block Stung Fresh Average 30 1 2 3 27 9 33 33 4 22* 4 4 5 6 7 8 9 33 9 21 15 6 16 19 15 10 9 27 24 5 19 20 24 21* 33 70 28 Average 40 22 16 Dia 4 Con Dia 36 48 28 33 36 39 39 49 36 46 Con Dia 39 39 39 39 39 39 39 39 39 39 Con -4 -4 -4 -4 -4 4 4 Con Dia 1 5 -7 -10 1 -4 4 6 1 3 + + 4 4 4 Block Avg SS 1935 Grand Avg 1521 1 Betw eff 160 1 Block eff 254 8 DF 10 FIGURE 7.21 CR decomposition of the dog (= block) averages. The decomposition is based on the CR structure used to assign operations to dogs in the SP/RM design. a. Scatterplot Res vs. Fit for the complete bee sting data, including the outlier. (If yo did the decomposition as part of Exericse D. 14, you already have the residuals. To fitted values, note that for a CB design, Fit = Block average + Treatment effect.) Once you have the plot, circle the two points for block 9, which contains the outlier, and 5. Res vs. Fit #1 get fit the data draw a "balloon" around the other points, as in Example 7.21. b. Pick one: Apart from the two outlying points: (1) The plot shows about what you would expect if the model and Fisher assumptions are appropriate. (II) The plot shows a curved old lationship between residuals and fitted values, suggesting that the model does not larger fitted values, suggesting that the assumption of same SDs is not appropriate for the in the present scale. (III) The plot is fatter toward the right : larger residuals tend to go with data in the present scale. (IV) The plot shows a balloon with a slight tilt. Block Stung Fresh Average 30 1 2 3 27 9 33 33 4 22* 4 4 5 6 7 8 9 33 9 21 15 6 16 19 15 10 9 27 24 5 19 20 24 21* 33 70 28 Average 40 22 16 Dia 4 Con Dia 36 48 28 33 36 39 39 49 36 46 Con Dia 39 39 39 39 39 39 39 39 39 39 Con -4 -4 -4 -4 -4 4 4 Con Dia 1 5 -7 -10 1 -4 4 6 1 3 + + 4 4 4 Block Avg SS 1935 Grand Avg 1521 1 Betw eff 160 1 Block eff 254 8 DF 10 FIGURE 7.21 CR decomposition of the dog (= block) averages. The decomposition is based on the CR structure used to assign operations to dogs in the SP/RM design
