Sei sulla pagina 1di 1

each pair of numbers x; y be represented by a dot in a diagram with the xs on

the horizontal axis, and we have the scatter diagram shown in Figure 8.1. The dots
do not fall perfectly on a straight line, but scatter around a line, very typical for
statistical relationships. However, a straight line seems to t very well. Generally,
the 12 dots go from upper left to lower right, and we have a negative association.
As shown in Example 2.8, we obtained a Pearsons cor- relation coecient of r
0:946, indicating a very strong negative association.
TABLE 8.1
x y x 2 y2 xy
112 63 12,544 3,969 7,056
119 52 14,161 2,704 6,188
120 6,561 14,400 9,720 84
72 11,236 5,184 7,632 103

111 66 12,321 4,356 7,326 107 72 11,449 5,184 7,704


92 75 8,464 5,625 6,900 80 118 6,400 13,924 9,440 81
114 7,056 12,996 9,576 118 42 13,924 1,764 4,956 106
90 10,609 8,100 9,270 94 91 8,836 8,281 8,554

1,207 975 123,561 86,487 94,322


Figure 8.1 Graphical display for data in Example 8.1.
286 CORRELATION AND REGRESSION
Applying the formulas, we obtain estimates for the slope and intercept as follows:
b1
94;3221207975=12 123;56112072=12 1:74 x 1207 12 100:6 y
975 12 81:3 b0 81:31:74100:6 256:3
For example, if the birth weight is 95 oz, it is predicted that the increase be- tween
days 70 and 100 of life would be
^ yy 256:31:7495 90:1% of birth weight
Note: An SAS program would include these instructions:
DATA; INPUT WEIGHT GAIN; DATALINES; 112 63 111 66 ... 103 90 94 91 ; PROC REG;
MODEL = WEIGHT; PLOT GAINWEIGHT;
for which we will get the analysis as well as the scatter diagram.
Example 8.2 In Table 8.2 the rst two columns give the values for age (x, in years)
and systolic blood pressure (y, in mmHg) for

Potrebbero piacerti anche