In: Statistics and Probability
A random sample of 24 companies from the Forbes 500 list was selected, and the relationship between their annual sales (, in hundreds of thousands of dollars) and annual profits (, in hundreds of thousands of dollars) was investigated by the researchers. [Note: if x = 200, that means the company had annual sales of $20 million; similarly for y values]. The following simple linear regression model was used: = + + .
Answer the following questions:
(a) In order for the researchers to validate their choice for simple linear regression, list 3 assumptions that the error terms, , must satisfy? List these assumptions one by one, followed by a short description of how to check that particular assumption.
(b) You are given the following summary data: ∑x= 6,600 ∑x2= 1,818,000 ∑y= 4,008 ∑y2= 673,868 ∑xy= 1,104,920 SSE = 1,905.075
(i) Calculate the sample regression equation. Provide a clear interpretation of the slope coefficient in the context of the question.
(ii) Calculate an interval (confidence or prediction) for the profit of one randomly selected Forbes 500 company whose sales are 270.5 hundred thousand dollars. Use a level of 95% and provide an interpretation of your interval in the context of the question.
a)
1)No or little multicollinearity
2)No auto-correlation
3)Homoscedasticity
b)
sample size , n = 24
1)
here, x̅ =Σx/n = 275.0000 ,
ȳ = Σy/n = 167
SSxx = Σx² - (Σx)²/n =
3000.000
SSxy= Σxy - (Σx*Σy)/n =
2720.000
SSyy = Σy²-(Σy)²/n = 4532.000
estimated slope , ß1 = SSxy/SSxx = 2720.000
/ 3000.000 = 0.9067
intercept, ß0 = y̅-ß1* x̄ =
-82.3333
so, regression line is Ŷ =
-82.33 + 0.91 *x
It means if we incraese the sales by 1 million then profit will increase by .91 milliuon.
2)
X Value= 270.5
Confidence Level= 95%
Sample Size , n= 24
Degrees of Freedom,df=n-2 = 22
critical t Value=tα/2 = 2.074
Predicted Y at X= 270.5 is
Ŷ = -82.333 +
0.907 * 270.5 =
162.920
standard error, S(ŷ)=Se*√(1/n+(X-X̅)²/Sxx) =
2.132
margin of error,E=t*Std error=t* S(ŷ) =
2.0739 * 2.1322 =
4.4220
Confidence Lower Limit=Ŷ +E = 162.920
- 4.4220 = 158.498
Confidence Upper Limit=Ŷ +E = 162.920
+ 4.4220 = 167.342
thanks
revert back for doubt
please upvote