In: Statistics and Probability
2.40 Stretching a scatterplot. Changing the units of
measurement can greatly alter the appearance of a scat-
terplot. Consider the following data: STRETCH
x −4 −4 −3 3 4 4
y 0.5 −0.6 −0.5 0.5 0.5 −0.6
(a) Draw x and y axes each extending from −6 to 6. Plot
the data on these axes.
(b) Calculate the values of new variables x* = x/10 and
y* = 10y, starting from the values of x and y. Plot y*
against x*
on the same axes using a different plotting
symbol. The two plots are very different in appearance.
(c) Find the correlation between x and y. Then find
the correlation between x*
and y*
. How are the two
correlations related? Explain why this isn’t surprising.
solution:
a)
Scatter plot of x and y:
b).
Scatter plot of x*and y*:
c).
For x and y:
X | Y | XY | X² | Y² |
-4 | 0.5 | -2 | 16 | 0.25 |
-4 | -0.6 | 2.4 | 16 | 0.36 |
-3 | -0.5 | 1.5 | 9 | 0.25 |
3 | 0.5 | 1.5 | 9 | 0.25 |
4 | 0.5 | 2 | 16 | 0.25 |
4 | -0.6 | -2.4 | 16 | 0.36 |
Ʃx = | Ʃy = | Ʃxy = | Ʃx² = | Ʃy² = |
0 | -0.2 | 3 | 82 | 1.72 |
Sample size, n = 6
SSxx = Ʃx² - (Ʃx)²/n = 82 - (0)²/6 = 82
SSyy = Ʃy² - (Ʃy)²/n = 1.72 - (-0.2)²/6 = 1.71333333
SSxy = Ʃxy - (Ʃx)(Ʃy)/n = 3 - (0)(-0.2)/6 = 3
Correlation coefficient, r = SSxy/√(SSxx*SSyy) = 3/√(82*1.71333) = 0.2531
-----------------
For x* and y*:
X | Y | XY | X² | Y² |
-0.4 | 5 | -2 | 0.16 | 25 |
-0.4 | -6 | 2.4 | 0.16 | 36 |
-0.3 | -5 | 1.5 | 0.09 | 25 |
0.3 | 5 | 1.5 | 0.09 | 25 |
0.4 | 5 | 2 | 0.16 | 25 |
0.4 | -6 | -2.4 | 0.16 | 36 |
Ʃx = | Ʃy = | Ʃxy = | Ʃx² = | Ʃy² = |
0 | -2 | 3 | 0.82 | 172 |
Sample size, n = 6
SSxx = Ʃx² - (Ʃx)²/n = 0.82 - (0)²/6 = 0.82
SSyy = Ʃy² - (Ʃy)²/n = 172 - (-2)²/6 = 171.333333
SSxy = Ʃxy - (Ʃx)(Ʃy)/n = 3 - (0)(-2)/6 = 3
Correlation coefficient, r = SSxy/√(SSxx*SSyy) = 3/√(0.82*171.33333) = 0.2531
please give me thumb up