In: Statistics and Probability
The following data is representative of that reported in an article with x = burner-area liberation rate (MBtu/hr-ft2) and y = NOx emission rate (ppm):
x | 100 | 125 | 125 | 150 | 150 | 200 | 200 | 250 | 250 | 300 | 300 | 350 | 400 | 400 |
y | 150 | 150 | 180 | 210 | 180 | 310 | 280 | 410 | 440 | 430 | 400 | 610 | 610 | 680 |
(a) Does the simple linear regression model specify a useful
relationship between the two rates? Use the appropriate test
procedure to obtain information about the P-value, and
then reach a conclusion at significance level 0.01.
State the appropriate null and alternative hypotheses.
H0: β1 ≠ 0
Ha: β1 =
0H0: β1 = 0
Ha: β1 ≠
0 H0:
β1 = 0
Ha: β1 <
0H0: β1 = 0
Ha: β1 > 0
Calculate the test statistic and determine the P-value.
(Round your test statistic to two decimal places and your
P-value to three decimal places.)
t | = | |
P-value | = |
State the conclusion in the problem context.
Fail to reject H0. There is evidence that the model is useful.Reject H0. There is no evidence that the model is useful. Reject H0. There is evidence that the model is useful.Fail to reject H0. There is no evidence that the model is useful.
X | Y | XY | X² | Y² |
100 | 150 | 15000 | 10000 | 22500 |
125 | 150 | 18750 | 15625 | 22500 |
125 | 180 | 22500 | 15625 | 32400 |
150 | 210 | 31500 | 22500 | 44100 |
150 | 180 | 27000 | 22500 | 32400 |
200 | 310 | 62000 | 40000 | 96100 |
200 | 280 | 56000 | 40000 | 78400 |
250 | 410 | 102500 | 62500 | 168100 |
250 | 440 | 110000 | 62500 | 193600 |
300 | 430 | 129000 | 90000 | 184900 |
300 | 400 | 120000 | 90000 | 160000 |
350 | 610 | 213500 | 122500 | 372100 |
400 | 610 | 244000 | 160000 | 372100 |
400 | 680 | 272000 | 160000 | 462400 |
Ʃx = | 3300 |
Ʃy = | 5040 |
Ʃxy = | 1423750 |
Ʃx² = | 913750 |
Ʃy² = | 2241600 |
Sample size, n = | 14 |
x̅ = Ʃx/n = 3300/14 = | 235.7142857 |
y̅ = Ʃy/n = 5040/14 = | 360 |
SSxx = Ʃx² - (Ʃx)²/n = 913750 - (3300)²/14 = | 135892.8571 |
SSyy = Ʃy² - (Ʃy)²/n = 2241600 - (5040)²/14 = | 427200 |
SSxy = Ʃxy - (Ʃx)(Ʃy)/n = 1423750 - (3300)(5040)/14 = | 235750 |
Slope, b = SSxy/SSxx = 235750/135892.85714 = 1.734822602
y-intercept, a = y̅ -b* x̅ = 360 - (1.73482)*235.71429 =
-48.92247043
Sum of Square error, SSE = SSyy -SSxy²/SSxx = 427200 -
(235750)²/135892.85714 = 18215.5716
Standard error, se = √(SSE/(n-2)) = √(18215.57162/(14-2)) = 38.96106
Null and alternative hypothesis:
Ho: β₁ = 0
Ha: β₁ ≠ 0
n = 14
Test statistic:
t = b/(se/√SSxx) = 16.41
df = n-2 = 12
p-value = T.DIST.2T(ABS(16.4143), 12) = 0.000
Conclusion:
Reject H0. There is evidence that the model is useful.