US Cars Japanese Cars

20 24
16 27
20 27
15 26
21 33
17 34
16 17
16 19
16 29
23 24
23 17
16 20
17 22
26 18
- 20
- 17
- 23

n 14 17
mean 18.71429 23.352941176
Standard Deviation 3.451389 5.3960770064

t-value -2.897086
n-1 16
t-critical 2.119905
p-value 0.010504 since the p-value or probability that the change in mean

2- sample t-test: compares two normally distributed but independent groups. Since MPG is
stastical comparison in MPG

1 sample test: compares the mean to the null hypothesis. This test compare single data se

paired sample: compares the difference between paired observations (before and after) an
dependant i.e: same group at different times. However, in this case the samples are in
robability that the change in mean is due to noise is <0.05, chance that the effect due country of manufacture h

ndependent groups. Since MPG is compared for two different manufacturer that differ in location, this test is su

. This test compare single data set to standard value; therefore, this type of test is not suitable in evaluating tw

observations (before and after) and then perform 1 sample test on the difference. to perform this test the two s
ver, in this case the samples are independent as MPG are compared between two different manufactuerer in diff
ountry of manufacture has significant effect on MPG

location, this test is suitable to perform

uitable in evaluating two different samples

form this test the two samples need to be

nt manufactuerer in different countries
 x y 𝑦 ̂ 〖 (𝑦−𝑦 ̂) 〗 ^2
〖 (𝑦 ̂−𝑦 ̅) 〗 ^2
100 16.7 10.4396 39.19261
100 13.1 10.4396 7.077728
100 8.4 10.4396 4.159968
100 6.6 10.4396 14.74253
200 16.3 16.5996 0.08976
200 12.5 16.5996 16.80672
200 11.3 16.5996 28.08576
200 22.1 16.5996 30.2544
400 27.6 28.9196 1.741344
400 30.9 28.9196 3.921984
500 36.6 35.0796 2.311616
500 34.2 35.0796 0.773696

sum 149.1581
mean 250 19.69167
n data 12
a (#regression coefficent) 2
Predicted R^2 0.8731
Adjusted R^2 0.844886

The predicted R^2 indicates that ~87.3% of the variation in the response variable y is due
The adjusted R^2 account for the number of predictors and the sample size hence providin
measurement to the model's performance

a 0.061648
b 4.27963
SSE 149.1581 1024.533
SSR 1024.533 14.91581
SST 1175.289 68.68771
R2_predicted) 0.871728
A 1 linear regression
n 12
R2_adjusted 0.860397
Georgia 33 214 1 83.5
Idaho 44.5 116 0 114
Illinois 40 124 0 89.5
Indiana 40.2 128 0 86.2
Iowa 42.2 128 0 93.8
Kansas 38.5 166 0 98.5
Kentucky 37.8 147 0 85
Louisiana 31.2 190 1 91.8
Maine 45.2 117 1 69
Maryland 39 162 1 76.5
Massachusetts 42.2 143 1 71.8
Michigan 43.5 117 0 84.5
Minnesota 46 116 0 94.5
Mississippi 32.8 207 1 90
Missouri 38.5 131 0 92
Montana 47 109 0 110.5
Nebraska 41.5 122 0 99.5
Nevada 39 191 0 117
NewHampshire 43.8 129 1 71.5
NewJersey 40.2 159 1 74.5
NewMexico 35 141 0 106
MewYork 43 152 1 75.5
NorthCarolina 35.5 199 1 79.5
NorthDakota 47.5 115 0 100.5
Ohio 40.2 131 0 82.8
Oklahoma 35.5 182 0 97.2
Oregon 44 136 1 120.5
Pennsylvania 40.8 132 0 77.8
RhodeIsland 41.8 137 1 71.5
SouthCarolina 33.8 178 1 81
SouthDakota 44.8 86 0 100
Tennessee 36 186 0 86.2
Texas 31.5 229 1 98
Utah 39.5 142 0 111.5
Vermont 44 153 1 72.5
Virginia 37.5 166 1 78.5
Washington 47.5 117 1 121
WestVirginia 38.8 136 0 80.8
Wisconsin 44.5 110 0 90.2
Wyoming 43 134 0 107.5
−𝑦 ̂) 〗 ^2 〖 (𝑦−𝑦 ̅) 〗 ^2
〖 (𝑦 ̂−𝑦 ̅) 〗 ^2
85.60074 8.950069
85.60074 43.45007
85.60074 127.5017 40

85.60074 171.3917
35
9.560876 11.5034 f(x) = 0.0616
9.560876 51.72007 R² = 0.87309
30
9.560876 70.42007
9.560876 5.800069 25
85.15475 62.54174
85.15475 125.6267 20

y
236.7885 285.8917
15
236.7885 210.4917
10
1024.533 1175.289
5

0
50 100 1

e response variable y is due to predictor variable x

e sample size hence providing more precise ANOVA TABLE
Source of error
Model
Error
Total

p-value<5% hence chance that the

 y vs. x

f(x) = 0.0616481481481481 x + 4.27962962962963

R² = 0.873090176467998

0 100 150 200 250 300 350 400 450 5

SS df MSE f val p val

1024.533 1 1024.533
149.1581 10 14.91581 68.68771 8.628553E-06
1175.289 11

% hence chance that the effect is due to noise and the x main effect is statistically significant
450 500 550
Values of "Prob>F" less than 5% indicates model terms are significant. In this case, A-mix,
On the other hand, "Prob >F" greater than 10% are insignificant (C-Lab, AC &AD)
Values of "Prob>F" less than 5% indicates model terms are significant. In this case, A-A, B-
nt. In this case, A-mix, B-mixing time,D-temperature, E-derying time, and AB are signifacant model terms
Lab, AC &AD)
nt. In this case, A-A, B-B, and AB are signifacant model terms
cant model terms