You are on page 1of 9

Bouchra Tannir

260356824
Assignment 5

8.32 Cooling method for gas turbines (check which ones


is H0)

a. x=11066 s=1594 n=47 α=0.05

H₀: μ≤10000

Ha: μa>10000

Z= x-10000sn= 11066-10000159467=5.47

5.47>1.96=Zα2

Z>Zα2

Thus we reject the null hypothesis H₀ at the significance level α


= 0.05. We conclude that the mean heat rate of gas turbines
augmented with high pressure inlet fogging exceeds
10000kJ/kWh.

b. We rejected H₀ at the significance level α = 0.05. However,


we know that this hypothesis testing will lead to this conclusion
incorrectly (type I error) only 5% of the time when H₀ is true,
meaning when μ is lower than 10000kJ/kWh.

If the test statistics does not fall into the rejection region, we do
not reject H₀. However, we do not conclude that the null
hypothesis is true because we do not know the probability β
that our test procedure will lead to an incorrect acceptance of
H₀, meaning a type II error.

8.68 Testing a mosquito repellent

a. H₀: μ0≥0.95 Ha: µa <0.95 α=0.1

t = x-µ₀ s/√n=0.83-0.950.15/√5= -1.79

- tα = -1.53

-1.79 < -1.53 thus –t<-tα


Thus we reject H₀ with a significance level α=0.10. We then
conclude that the mean repellency of the new mosquito
repellent is less than 95% (with a significance level α=0.10).

b. The assumptions required are that the random sample is


selected from the target population and that the population
from which the sample is selected has a distribution that is
approximately normal (the sample is smaller than 30).

8.72 Lengths of great White Sharks

x=24+20+223=22 S=2 α=0.05 tα2=4.303

H0: μ0≤21

Ha: μa>21

t=x-μ0sn=22-2123=0.866 <4.303= tα2

Thus, we don’t reject H₀. The following data doesn’t support the
marine biologist’s claim that the great white shark grows much
longer than 21 feet. However, we reserve judgment on which
hypothesis is true.

8.86 Verbs and double-object datives

H₀: p₀≥ 13

Ha: p< 13

σp= 1323√35=0.08

p= 1035=0.29

Z= 0.29-(13)0.08=-0.54

Zα=z0.05=0.0199

-0.54< -0.0199

Thus we reject H₀ with a significance level α=0.05. We then


conclude that the true fraction of sentences with the verb “buy”
that are DODs is less tha 1/3 (with a significance level α=0.05).
8.152 NCAA March Madness

a. 1 vs. 16 seed:
H0: p > p₀
HA: p ≤ p₀ p₀ = 0.5 p= 5252

Z = (^p – p₀)/√([ p₀ (1- p₀)/n)


Z = (1 – 0.5)/√([0.5 (1- 0.5)/52)
≈ 7.21

Z < -zα
Z < -z0.05
7.21 > -1.644 thus H₀ is not rejected.

2 vs. 15 seed:
H₀: p > p₀
Ha: p ≤ p₀ p₀ = 0.5 p= 4952

Z = (^p – p₀)/√([ p₀ (1- p₀)/n)


Z = (0.9423 – 0.5)/(0.06934)
≈ 6.38

Z < -zα
Z < -z0.05
6.38 > -1.644 thus H₀ is not rejected.

8 vs. 9 seed:
H0: p > p₀
HA: p ≤ p₀ p₀ = 0.5 p= 2252

Z = (^p – p₀)/√([ p₀ (1- p₀)/n)


Z = (0.423 – 0.5)/(0.06934)
≈ -1.11

Z < -zα
Z < -z0.05
-1.11 > -1.644 thus H₀ is not rejected.

There is not enough evidence to reject the perception that the


higher seeded team has a better than 5252 chances of winning.
The trend is that as p decreases, the test statistic also
decreases.

b. 1 vs. 16 seed:
H0: μ > μ₀
HA: μ ≤ μ₀ μ₀ = 10; x = 22.9; s = 12.4; n = 52

Z = (x – μ₀)/(s/√n)
= (22.9 – 10)/(12.4/√52)
≈ 7.502

Z < -zα
Z < -z0.05
7.502 > -1.644 thus H₀ is not rejected.

4 vs. 13 seed:
H0: μ > μ₀
HA: μ ≤ μ₀ μ₀ = 10; x = 10; s = 12.5; n = 52

Z = (x – μ₀)/(s/√n)
= (10 – 10)/(s/√n)
=0

Z < -zα
Z < -z0.05
0 >-1.644 thus H₀ is not rejected.

There is no evidence to reject the claim that a 1-, 2-, 3- or 4-


seeded team will win by an average of more than 10 points.

c. 5 vs. 12 seed:
H0: μ < μ₀
HA: μ ≥ μ₀ μ₀ = 5; x = 5.3; s = 10.4; n = 52

Z = (x – μ₀)/(s/√n)
= (5.3 – 8)/(10.4/√52)
≈ -1.87

Z > zα
Z > z0.05
-1.87 < 1.644 thus H₀ is not rejected.

8 vs. 9 seed:
H0: μ < μ₀
HA: μ ≥ μ₀ μ₀ = 5; x = -2.1; s = 11.0; n = 52

Z = (x – μ₀)/(s/√n)
= (-2.1 – 8)/(11/√52)
≈ -6.62

Z > zα
Z > z0.05
-6.62 < 1.644 thus H₀ is not rejected.

There is not enough evidence to reject claims that 5-, 6-, 7- or


8- seeded teams will win by an average of less than 5 points.

e. H₀: μd= μ₀
Ha: μd ≠ μ₀ xd= 0.7; s d = 11.3; n = 360; μ₀ = 0

Z = (xd – μ₀)/(s d /_n)


= (0.7 – 0)/(11.3/√360)
= (0.7)(√360)/11.3
≈ 1.175

Z>Zα/2
Z>Z0.05/2
1.175 < 1.96 thus H₀ is not rejected, the true main difference
μdcan equal μ₀ = 0. Therefore, the point spread, on average,
can be considered a good predictor of the outcome of the
game.

9.22 Pig castration study


S1=S2=0.9 α=0.05 tα2=t0.025

Spooled= n1-1S12+n2-1S22n1+n2-2=230.92+240.9247=0.9

t=x1-x2- D0Sp21n1+1n2= 0.74-0.70-0√(0.81)(124+125)=0.156

2.00<t0.025;47<2.021

0.156<tα/2

Thus, we do not reject H₀ and conclude that the population


mean number of high frequency vocal responses doesn’t differ
for piglet castration by the two methods.

9.48 Homophone Confusion to Alzheimer’s patients.

In this case, the two observations are paired (it’s paired


difference experiment). The two tests are dependant.
95% confidence interval for μd= μ2-μ1

xd±tα2Sdn= 1.65±2.0933.2020=0.15, 3.15

This interval doesn’t include 0. Therefore, μ2-μ1>0 and μ2>μ1. We


conclude that the patients show an increase in mean
homophone confusion errors over time.

For this procedure to be valid, we assume that the population


of differences has a distribution that is approximately normal
and that the sample differences are randomly selected from the
population differences (the sample n is small).

The histograms for both samples are not mound shaped, the
distribution is skewed, meaning the data is not normally
distributed (however we notice that the histogram on the right
is more mound shaped). Thus, the assumptions are not
satisfied.

9.70 Gambling in Public High Schools.


Ho:p1-p2=0
Ha: p1-p2≠0
p1=468421484=0.218 and q1=0.782
p2=531323199=0.23 and q2=0.77
σp1-p2= p1q1n1+p2q2n2=0.2180.78221484+0.230.7723199=0.00395
Z=p1-p2σp1-p2=0.218-0.230.00395=-3.04
-3.04>0.004
Z>Z0.01 thus Z>Z0.005

Thus, we reject the null hypothesis H₀. We conclude that the


percentages of ninth grade boys who gambled weekly or daily
on any game in 1992 and 1998 are significantly different since
p1-p2>0 and p1>p2 and -3.039>0.004 provides strong evidence
against it.

b. p-value=2×PZ>-3.04=0.0024

p-value is smaller than α=0.01. Therefore, we agree with the


professor.

9.114 Hull Failure Oil Tankers

Data:

Diagram 1 – Spillage by collisions and fires

FIRE:
Mean 69.25
Std Dev 61.363485
Std Err Mean 17.714112
Upper 95% Mean 108.2385
Lower 95% Mean 30.261502
N 12
COLLISIONS:
Mean 76.6
Std Dev 70.362869
Std Err Mean 22.250693
Upper 95% Mean 126.93456
Lower 95% Mean 26.265436
N 10

HULLFAIL:
Mean 58.583333
Std Dev 57.544383
Std Err Mean 16.611633
Upper 95% Mean 95.14529
Lower 95% Mean 22.021376
N 12

GROUNDING:
Mean 47.785714
Std Dev 28.466387
Std Err Mean 7.6079618
Upper 95% Mean 64.221717
Lower 95% Mean 31.349712
N 14

a. Spooled2=0.1161.362+ 970.36220=4298

We consider the variances equal since S1S2>0.75


x1- x2±tα2Sp21n1+1n2
69.25-76.6±1.7254298112+110
-55.77, 41.07

Since the 90% confidence interval includes 0, we can conclude


that there is no evidence of a difference between the mean
spillage amount in accidents caused by collision and the mean
spillage amount in accidents caused by fires or explosions.

b. This case is a small sample with different variances and


different samples.
Ho:μ1-μ2=0
Ha: μ1-μ2≠0
T=x1-x2S12n1+S22n2= 58.58-47.7957.54212+28.47214=0.59
ϑ=S12n1+S22n22S12n12n1-1+S22n22(n2-1)=15.523≈16
0.59<2.120 thus 0.59<t0.025;16

Therefore, we don’t reject the null hypothesis H₀. We can then


conclude that there is no evidence of a difference between the
two means and that μ1-μ2=0.

c. Assumptions made in part a and part bare the following: the


two samples selected from the two populations are
independent (samples are also less than 30) and the two
populations have normal distributions.

For the first diagram, the data is not perfectly symmetrical (the
mean is not identical to the median but it is relatively close). As
for the second diagram, the mean is equal to the median so we
can conclude that the distribution of the data is mound shaped
and is in fact a normal distribution. We can assume that the
assumptions are reasonably satisfied.

9.134 Students Attitudes towards Parents


n=Zα22σD2SE2= 1.6452120.32=30.066≈31

Thus, the number of male required to obtain a 90% confidence


interval with a sampling error of 0.3 is 31.

You might also like