In: Statistics and Probability
Research into the relationship between hours of study and grades shows widely different conclusions. A recent survey of graduates who wrote the Graduate Management Admissions Test (GMAT) had the following results.
Hours Studied ( Midpoint) | Average Score |
40 | 220 |
50 | 310 |
65 | 350 |
75 | 440 |
85 | 560 |
105 | 670 |
95 | 700 |
a) Run the regression analysis in Excel on this data. Include your output with your answer. (Note: You may calculate by hand if you prefer).
b) What is the regression equation for this relationship?
c) Use the regression equation to predict the average score for each category of hours studied.
d) Plot the original data and the regression line on a scatter gram. (You may use Excel).
e) How accurate is this regression at predicting GMAT scores based on hours studied? Explain.
f) Use the t statistic to determine whether the Correlation Coefficient is “significant” at the 95% confidence level.
a)
x | y | (x-x̅)² | (y-ȳ)² | (x-x̅)(y-ȳ) |
40 | 220 | 1127.04 | 59675.51 | 8201.02 |
50 | 310 | 555.61 | 23804.08 | 3636.73 |
65 | 350 | 73.47 | 13061.22 | 979.59 |
75 | 440 | 2.04 | 589.80 | -34.69 |
85 | 560 | 130.61 | 9161.22 | 1093.88 |
105 | 670 | 987.76 | 42318.37 | 6465.31 |
95 | 700 | 459.18 | 55561.22 | 5051.02 |
ΣX | ΣY | Σ(x-x̅)² | Σ(y-ȳ)² | Σ(x-x̅)(y-ȳ) | |
total sum | 515 | 3250 | 3335.714286 | 204171.4 | 25392.86 |
mean | 73.57 | 464.29 | SSxx | SSyy | SSxy |
sample size , n = 7
here, x̅ = Σx / n= 73.57 ,
ȳ = Σy/n = 464.29
SSxx = Σ(x-x̅)² = 3335.7143
SSxy= Σ(x-x̅)(y-ȳ) = 25392.9
estimated slope , ß1 = SSxy/SSxx = 25392.9
/ 3335.714 = 7.6124
intercept, ß0 = y̅-ß1* x̄ =
-95.7709
b) so, regression line is Ŷ =
-95.7709 + 7.6124
*x
c)
Ŷ |
208.726 |
284.850 |
399.036 |
475.161 |
551.285 |
703.533 |
627.409 |
d)
e)
SSE= (SSxx * SSyy - SS²xy)/SSxx =
10870.343
std error ,Se = √(SSE/(n-2)) =
46.627
correlation coefficient , r = Sxy/√(Sx.Sy)
= 0.9730
R² = (Sxy)²/(Sx.Sy) =
0.9468
Very accurate model as R square is high.
f)
Ho: ρ = 0
Ha: ρ ╪ 0
n= 7
alpha,α = 0.05
correlation , r= 0.9730
t-test statistic = r*√(n-2)/√(1-r²) =
9.429
critical t-value =
2.5706
Decison: t value >t critical , So, Reject
Ho
Hence R is significant.
Please revert in case of any doubt.
Please upvote. Thanks in advance