In: Statistics and Probability
The results of the blood pressure measurements taken of the same patient by fourteen different second-year medical students are listed in the table below.
Systolic Diastolic
138 82
130 91
135 100
140 100
120 80
125 90
120 80
130 80
130 80
144 98
143 105
140 85
130 70
150 100
a. Use a 5% level of significance to test whether there is a significant linear correlation between the two variables {use hypothesis test for the slope}
b. For a given systolic blood pressure of 125, identify the 95% prediction interval estimate for the slope (diastolic blood pressure). Write a statement interpreting your interval in the context of this example.
x | y | (x-x̅)² | (y-ȳ)² | (x-x̅)(y-ȳ) |
138 | 82 | 16.58 | 44.13 | -27.05 |
130 | 91 | 15.43 | 5.56 | -9.26 |
135 | 100 | 1.15 | 128.98 | 12.17 |
140 | 100 | 36.86 | 128.98 | 68.95 |
120 | 80 | 194.01 | 74.70 | 120.38 |
125 | 90 | 79.72 | 1.84 | -12.12 |
120 | 80 | 194.01 | 74.69898 | 120.3827 |
130 | 80 | 15.43 | 74.69898 | 33.954 |
130 | 80 | 15.43 | 74.70 | 33.95 |
144 | 98 | 101.43 | 87.56 | 94.24 |
143 | 105 | 82.29 | 267.56 | 148.38 |
140 | 85 | 36.86 | 13.27 | -22.12 |
130 | 70 | 15.43 | 347.56 | 73.24 |
150 | 100 | 258.29 | 128.98 | 182.53 |
ΣX | ΣY | Σ(x-x̅)² | Σ(y-ȳ)² | Σ(x-x̅)(y-ȳ) | |
total sum | 1875.00 | 1241.00 | 1062.93 | 1453.21 | 817.6 |
mean | 133.93 | 88.64 | SSxx | SSyy | SSxy |
a)
sample size , n = 14
here, x̅ = Σx / n= 133.929 ,
ȳ = Σy/n = 88.643
SSxx = Σ(x-x̅)² = 1062.9286
SSxy= Σ(x-x̅)(y-ȳ) = 817.6
estimated slope , ß1 = SSxy/SSxx = 817.6
/ 1062.929 = 0.76924
intercept, ß0 = y̅-ß1* x̄ =
-14.37981
so, regression line is Ŷ =
-14.380 + 0.769
*x
SSE= (SSxx * SSyy - SS²xy)/SSxx =
824.2540
std error ,Se = √(SSE/(n-2)) =
8.2878
b)
X Value= 125
Confidence Level= 95%
Sample Size , n= 14
Degrees of Freedom,df=n-2 = 12
critical t Value=tα/2 = 2.179 [excel
function: =t.inv.2t(α/2,df) ]
X̅ = 133.93
Σ(x-x̅)² =Sxx 1062.93
Standard Error of the Estimate,Se= 8.2878
Predicted Y at X= 125 is
Ŷ = -14.3798 +
0.7692 *125= 81.7747
For Individual Response Y
standard error, S(ŷ)=Se*√(1+1/n+(X-X̅)²/Sxx) =
8.874
margin of error,E=t*std error=t*S(ŷ)=
2.179 * 8.874 =
19.3345
Prediction Interval Lower Limit=Ŷ -E =
81.775 - 19.335 =
62.4402
Prediction Interval Upper Limit=Ŷ +E =
81.775 + 19.335 =
101.1092