Question

Consider a population list x, with a CV= 10%. A second population list, y, has a CV= 20%. The linear regression of y on x is y =10x. What is the correlation between y and x? (Assume linearity, homoscedasticity, no outliers and normality)

Answer 1

The correlation coefficient 0.5 depicts that there is moderate positive correlation between y and x.

Explanation:

The linear regression equation is

y= a+bx

The given linear equation is

y= 10x

Here, slope=b=10 and intercept=a=0.

From above equation,

ybar=10xbar

We are given that CVx=0.1 and CVy=0.2

where CV=(standard deviation/mean)*100

We know that

b=r(Sy/Sx)

Multiplying by xbar/ybar on both sides

(xbar/ybar)b=r(Sy/Sx)(xbar/ybar)

(xbar/ybar)b=r[(Sy/ybar)/(Sx/xbar))]

(xbar/ybar)b=r[CVy/CVx]

As CVy/CVx=(Sy/ybar)*100/(Sx/xbar)*100=(Sy/ybar)/(Sx/xbar).

By putting ybar=10xbar, b=10, CVx=0.1 and CVy=0.2.

(xbar/10xbar)10=r(0.2/0.1)

1=r(0.2/0.1)

1=r(2)

r=0.5

There is moderate positive correlation between y and x.