In: Statistics and Probability
The PulseRates data set has pulse, height and weight for 90 patients who had an echocardiogram test in one of the clinic in Ottawa in the month of September. (a)[2] Fit a L-S regression line using height to predict the pulse rate (pulse rate is called the response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient)
(b)[2] Fit a L-S regression line using weight to predict the pulse rate (response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient) .
(c)[2] Fit a L-S regression line using height to predict the weight (response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient) .
(d) What is the predicted value for weight for a patient whose height is 68.7 inches?[1] .
For 71.5 inches?[1] . For 83 inches ?[1] . Which of these seem to make sense?[1] . Can you predict the height if the weight is 156 lb
Pulse | Height | Weight |
64 | 66 | 140 |
58 | 72 | 145 |
62 | 73.5 | 160 |
66 | 73 | 190 |
64 | 69 | 155 |
74 | 73 | 165 |
84 | 72 | 150 |
68 | 74 | 190 |
62 | 72 | 195 |
76 | 71 | 138 |
90 | 74 | 160 |
80 | 72 | 155 |
92 | 70 | 153 |
68 | 67 | 145 |
60 | 71 | 170 |
62 | 72 | 175 |
66 | 69 | 175 |
70 | 73 | 170 |
68 | 74 | 180 |
72 | 66 | 135 |
70 | 71 | 170 |
74 | 70 | 157 |
66 | 70 | 130 |
70 | 75 | 185 |
96 | 61 | 140 |
62 | 66 | 120 |
78 | 68 | 130 |
82 | 68 | 138 |
100 | 63 | 121 |
68 | 70 | 125 |
96 | 68 | 116 |
78 | 69 | 145 |
88 | 69 | 150 |
62 | 62.75 | 112 |
80 | 68 | 125 |
62 | 74 | 190 |
60 | 71 | 155 |
72 | 69 | 170 |
62 | 70 | 155 |
76 | 72 | 215 |
68 | 67 | 150 |
54 | 69 | 145 |
74 | 73 | 155 |
74 | 73 | 155 |
68 | 71 | 150 |
72 | 68 | 155 |
68 | 69.5 | 150 |
82 | 73 | 180 |
64 | 75 | 160 |
58 | 66 | 135 |
54 | 69 | 160 |
70 | 66 | 130 |
62 | 73 | 155 |
76 | 74 | 148 |
88 | 73.5 | 155 |
70 | 70 | 150 |
90 | 67 | 140 |
78 | 72 | 180 |
70 | 75 | 190 |
90 | 68 | 145 |
92 | 69 | 150 |
60 | 71.5 | 164 |
72 | 71 | 140 |
68 | 72 | 142 |
84 | 69 | 136 |
74 | 67 | 123 |
68 | 68 | 155 |
84 | 66 | 130 |
61 | 65.5 | 120 |
64 | 66 | 130 |
94 | 62 | 131 |
60 | 62 | 120 |
72 | 63 | 118 |
58 | 67 | 125 |
88 | 65 | 135 |
66 | 66 | 125 |
84 | 65 | 118 |
62 | 65 | 122 |
66 | 65 | 115 |
80 | 64 | 102 |
78 | 67 | 115 |
68 | 69 | 150 |
72 | 68 | 110 |
82 | 63 | 116 |
76 | 62 | 108 |
87 | 63 | 95 |
90 | 64 | 125 |
78 | 68 | 133 |
68 | 62 | 110 |
86 | 67 | 150 |
(a)[2] Fit a L-S regression line using height to predict the pulse rate (pulse rate is called the response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient)
Ans:
The equation of the L-S regression line is
Pulse = 18.69 -0.6624 Height
Predictor Coef SE Coef T P
Constant 118.69 21.3600 5.56 0.000
Height -0.6624 0.3100 -2.14 0.035
S = 10.5969 R-Sq = 4.9% R-Sq(adj) = 3.9%
The estimated p-value for height is 0.035. Hence, it makes sense to use this line because the Height has significant effect on pulse rate at 0.05 level of significance. The correlation coefficient value is -0.222.
(b)[2] Fit a L-S regression line using weight to predict the pulse rate (response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient) .
Ans: The equation of the L-S regression line is
The regression equation is
Pulse = 86.475 - 0.0918 Weight
Predictor Coef SE Coef T P
Constant 86.4750 7.027 12.31 0.000
Weight -0.0918 0.0477 -1.93 0.057
S = 10.6463 R-Sq = 4.0% R-Sq(adj) = 3.0%
The estimated p-value for covariate weight is 0.057. Hence, it does not make sense to use this line because the Weight has insignificant effect on pulse rate at 0.05 level of significance. The correlation coefficient value is -0.201.
(c)[2] Fit a L-S regression line using height to predict the weight (response variable). What is the equation of the L-S regression line? . Does it make sense to use this line? Why? (Check the correlation coefficient) .
Ans: The equation of the L-S regression line is
Weight = - 204.52 + 5.0875 Height
Predictor Coef SE Coef T P
Constant -204.52 30.09 -6.80 0.000
Height 5.0875 0.4367 11.65 0.000
S = 14.9304 R-Sq = 60.7% R-Sq(adj) = 60.2%
The estimated p-value for covariate weight is 0.000. Hence, it makes sense to use this line because the Height has significant effect on Weight at 0.05 level of significance. The correlation coefficient value is 0.779.
(d) What is the predicted value for weight for a patient whose height is 68.7 inches?[1] .
For 71.5 inches?[1] . For 83 inches ?[1] . Which of these seem to make sense?[1] . Can you predict the height if the weight is 156 lb
Ans:
The predicted value for weight for a patient whose height is 68.7 inches
Weight = - 204.52 + 5.0875 *68.7=144.99 lb
The predicted value for weight for a patient whose height is 71.5 inches
Weight = - 204.52 + 5.0875 *71.5=159.23 lb
The predicted value for weight for a patient whose height is 83 inches
Weight = - 204.52 + 5.0875 *83=217.74 lb
All make sense.
For predicting the height if the weight is 156 lb required the equation of the L-S regression line, that is,
Height = 51.452 + 0.1192 Weight
Predictor Coef SE Coef T P
Constant 51.452 1.509 34.10 0.000
Weight 0.1192 0.0102 11.65 0.000
S = 2.28576 R-Sq = 60.7% R-Sq(adj) = 60.2%
The predicted height if the weight is 156 lb is
Height = 51.452 + 0.1192*156 = 70.0472